{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9999676343981616,
  "eval_steps": 5,
  "global_step": 7724,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8840553760528564,
      "eval_runtime": 13.5126,
      "eval_samples_per_second": 3.256,
      "eval_steps_per_second": 0.222,
      "step": 0
    },
    {
      "epoch": 0.00012946240735346474,
      "grad_norm": 4.048585840442336,
      "learning_rate": 1.6170763260025876e-09,
      "loss": 7.5117,
      "step": 1
    },
    {
      "epoch": 0.00025892481470692947,
      "grad_norm": 1.4619548166976115,
      "learning_rate": 3.234152652005175e-09,
      "loss": 3.356,
      "step": 2
    },
    {
      "epoch": 0.00038838722206039424,
      "grad_norm": 1.9882128124669458,
      "learning_rate": 4.851228978007762e-09,
      "loss": 4.2322,
      "step": 3
    },
    {
      "epoch": 0.0005178496294138589,
      "grad_norm": 2.1398653569970243,
      "learning_rate": 6.46830530401035e-09,
      "loss": 3.7495,
      "step": 4
    },
    {
      "epoch": 0.0006473120367673237,
      "grad_norm": 2.558235727161999,
      "learning_rate": 8.085381630012938e-09,
      "loss": 4.8491,
      "step": 5
    },
    {
      "epoch": 0.0006473120367673237,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8840553760528564,
      "eval_runtime": 17.0556,
      "eval_samples_per_second": 2.58,
      "eval_steps_per_second": 0.176,
      "step": 5
    },
    {
      "epoch": 0.0007767744441207885,
      "grad_norm": 2.428011730931814,
      "learning_rate": 9.702457956015525e-09,
      "loss": 4.7065,
      "step": 6
    },
    {
      "epoch": 0.0009062368514742531,
      "grad_norm": 2.78377608662843,
      "learning_rate": 1.1319534282018111e-08,
      "loss": 5.9541,
      "step": 7
    },
    {
      "epoch": 0.0010356992588277179,
      "grad_norm": 1.7838311394350632,
      "learning_rate": 1.29366106080207e-08,
      "loss": 4.5327,
      "step": 8
    },
    {
      "epoch": 0.0011651616661811825,
      "grad_norm": 3.5127677525157206,
      "learning_rate": 1.4553686934023287e-08,
      "loss": 6.1851,
      "step": 9
    },
    {
      "epoch": 0.0012946240735346474,
      "grad_norm": 1.743780699692862,
      "learning_rate": 1.6170763260025877e-08,
      "loss": 3.6406,
      "step": 10
    },
    {
      "epoch": 0.0012946240735346474,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8851207494735718,
      "eval_runtime": 15.1683,
      "eval_samples_per_second": 2.901,
      "eval_steps_per_second": 0.198,
      "step": 10
    },
    {
      "epoch": 0.001424086480888112,
      "grad_norm": 2.059856514430833,
      "learning_rate": 1.778783958602846e-08,
      "loss": 4.5706,
      "step": 11
    },
    {
      "epoch": 0.001553548888241577,
      "grad_norm": 2.33399271055671,
      "learning_rate": 1.940491591203105e-08,
      "loss": 4.4399,
      "step": 12
    },
    {
      "epoch": 0.0016830112955950416,
      "grad_norm": 2.3905341429677143,
      "learning_rate": 2.1021992238033634e-08,
      "loss": 5.3672,
      "step": 13
    },
    {
      "epoch": 0.0018124737029485063,
      "grad_norm": 2.1359444235175746,
      "learning_rate": 2.2639068564036222e-08,
      "loss": 4.4424,
      "step": 14
    },
    {
      "epoch": 0.0019419361103019711,
      "grad_norm": 1.9099410879445384,
      "learning_rate": 2.425614489003881e-08,
      "loss": 4.1826,
      "step": 15
    },
    {
      "epoch": 0.0019419361103019711,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8869850635528564,
      "eval_runtime": 14.9328,
      "eval_samples_per_second": 2.947,
      "eval_steps_per_second": 0.201,
      "step": 15
    },
    {
      "epoch": 0.0020713985176554358,
      "grad_norm": 3.5697954412824697,
      "learning_rate": 2.58732212160414e-08,
      "loss": 6.3228,
      "step": 16
    },
    {
      "epoch": 0.0022008609250089004,
      "grad_norm": 2.709006358306772,
      "learning_rate": 2.749029754204399e-08,
      "loss": 5.1042,
      "step": 17
    },
    {
      "epoch": 0.002330323332362365,
      "grad_norm": 2.296208822101843,
      "learning_rate": 2.9107373868046574e-08,
      "loss": 4.438,
      "step": 18
    },
    {
      "epoch": 0.00245978573971583,
      "grad_norm": 2.1548811171938995,
      "learning_rate": 3.072445019404916e-08,
      "loss": 5.1499,
      "step": 19
    },
    {
      "epoch": 0.002589248147069295,
      "grad_norm": 2.268973635895758,
      "learning_rate": 3.2341526520051754e-08,
      "loss": 4.3889,
      "step": 20
    },
    {
      "epoch": 0.002589248147069295,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8849431872367859,
      "eval_runtime": 16.3184,
      "eval_samples_per_second": 2.696,
      "eval_steps_per_second": 0.184,
      "step": 20
    },
    {
      "epoch": 0.0027187105544227595,
      "grad_norm": 1.635931764205022,
      "learning_rate": 3.395860284605434e-08,
      "loss": 3.5371,
      "step": 21
    },
    {
      "epoch": 0.002848172961776224,
      "grad_norm": 1.6615743997623915,
      "learning_rate": 3.557567917205692e-08,
      "loss": 4.1558,
      "step": 22
    },
    {
      "epoch": 0.002977635369129689,
      "grad_norm": 2.056005996616881,
      "learning_rate": 3.7192755498059514e-08,
      "loss": 4.4595,
      "step": 23
    },
    {
      "epoch": 0.003107097776483154,
      "grad_norm": 3.753040242280733,
      "learning_rate": 3.88098318240621e-08,
      "loss": 5.7734,
      "step": 24
    },
    {
      "epoch": 0.0032365601838366185,
      "grad_norm": 2.366113463130707,
      "learning_rate": 4.042690815006469e-08,
      "loss": 3.8374,
      "step": 25
    },
    {
      "epoch": 0.0032365601838366185,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8881391882896423,
      "eval_runtime": 24.3124,
      "eval_samples_per_second": 1.81,
      "eval_steps_per_second": 0.123,
      "step": 25
    },
    {
      "epoch": 0.003366022591190083,
      "grad_norm": 2.3798748862140813,
      "learning_rate": 4.204398447606727e-08,
      "loss": 4.8389,
      "step": 26
    },
    {
      "epoch": 0.003495484998543548,
      "grad_norm": 1.9670885219929928,
      "learning_rate": 4.366106080206986e-08,
      "loss": 4.0776,
      "step": 27
    },
    {
      "epoch": 0.0036249474058970125,
      "grad_norm": 1.7473095176758717,
      "learning_rate": 4.5278137128072445e-08,
      "loss": 3.7896,
      "step": 28
    },
    {
      "epoch": 0.0037544098132504776,
      "grad_norm": 1.6502477708403551,
      "learning_rate": 4.6895213454075036e-08,
      "loss": 4.0874,
      "step": 29
    },
    {
      "epoch": 0.0038838722206039423,
      "grad_norm": 1.7502633563958625,
      "learning_rate": 4.851228978007762e-08,
      "loss": 4.7852,
      "step": 30
    },
    {
      "epoch": 0.0038838722206039423,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8831676244735718,
      "eval_runtime": 16.859,
      "eval_samples_per_second": 2.61,
      "eval_steps_per_second": 0.178,
      "step": 30
    },
    {
      "epoch": 0.004013334627957407,
      "grad_norm": 2.587615402257072,
      "learning_rate": 5.012936610608021e-08,
      "loss": 4.9688,
      "step": 31
    },
    {
      "epoch": 0.0041427970353108716,
      "grad_norm": 2.4101266831742256,
      "learning_rate": 5.17464424320828e-08,
      "loss": 4.6631,
      "step": 32
    },
    {
      "epoch": 0.004272259442664336,
      "grad_norm": 3.093327561925277,
      "learning_rate": 5.336351875808539e-08,
      "loss": 5.7959,
      "step": 33
    },
    {
      "epoch": 0.004401721850017801,
      "grad_norm": 3.3657109917022705,
      "learning_rate": 5.498059508408798e-08,
      "loss": 7.2173,
      "step": 34
    },
    {
      "epoch": 0.0045311842573712655,
      "grad_norm": 1.92332727613145,
      "learning_rate": 5.659767141009056e-08,
      "loss": 3.8652,
      "step": 35
    },
    {
      "epoch": 0.0045311842573712655,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8849431872367859,
      "eval_runtime": 15.8823,
      "eval_samples_per_second": 2.77,
      "eval_steps_per_second": 0.189,
      "step": 35
    },
    {
      "epoch": 0.00466064666472473,
      "grad_norm": 2.490260539009604,
      "learning_rate": 5.821474773609315e-08,
      "loss": 4.8418,
      "step": 36
    },
    {
      "epoch": 0.004790109072078196,
      "grad_norm": 3.273589100056769,
      "learning_rate": 5.983182406209573e-08,
      "loss": 6.3818,
      "step": 37
    },
    {
      "epoch": 0.00491957147943166,
      "grad_norm": 2.6670528339103154,
      "learning_rate": 6.144890038809832e-08,
      "loss": 6.3396,
      "step": 38
    },
    {
      "epoch": 0.005049033886785125,
      "grad_norm": 1.348631600436967,
      "learning_rate": 6.306597671410092e-08,
      "loss": 3.6206,
      "step": 39
    },
    {
      "epoch": 0.00517849629413859,
      "grad_norm": 2.2949071960950675,
      "learning_rate": 6.468305304010351e-08,
      "loss": 4.8369,
      "step": 40
    },
    {
      "epoch": 0.00517849629413859,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8851207494735718,
      "eval_runtime": 15.0616,
      "eval_samples_per_second": 2.921,
      "eval_steps_per_second": 0.199,
      "step": 40
    },
    {
      "epoch": 0.005307958701492054,
      "grad_norm": 2.5142130259666353,
      "learning_rate": 6.630012936610609e-08,
      "loss": 5.5762,
      "step": 41
    },
    {
      "epoch": 0.005437421108845519,
      "grad_norm": 2.5446434752413825,
      "learning_rate": 6.791720569210868e-08,
      "loss": 5.4185,
      "step": 42
    },
    {
      "epoch": 0.005566883516198984,
      "grad_norm": 2.821907647616337,
      "learning_rate": 6.953428201811125e-08,
      "loss": 5.5137,
      "step": 43
    },
    {
      "epoch": 0.005696345923552448,
      "grad_norm": 2.370354951113661,
      "learning_rate": 7.115135834411385e-08,
      "loss": 4.8589,
      "step": 44
    },
    {
      "epoch": 0.005825808330905913,
      "grad_norm": 2.829677972793875,
      "learning_rate": 7.276843467011644e-08,
      "loss": 5.0503,
      "step": 45
    },
    {
      "epoch": 0.005825808330905913,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8833451867103577,
      "eval_runtime": 14.8834,
      "eval_samples_per_second": 2.956,
      "eval_steps_per_second": 0.202,
      "step": 45
    },
    {
      "epoch": 0.005955270738259378,
      "grad_norm": 1.7827160663194768,
      "learning_rate": 7.438551099611903e-08,
      "loss": 4.282,
      "step": 46
    },
    {
      "epoch": 0.006084733145612842,
      "grad_norm": 2.5018263160341982,
      "learning_rate": 7.600258732212161e-08,
      "loss": 4.6943,
      "step": 47
    },
    {
      "epoch": 0.006214195552966308,
      "grad_norm": 2.718135384255652,
      "learning_rate": 7.76196636481242e-08,
      "loss": 5.6211,
      "step": 48
    },
    {
      "epoch": 0.006343657960319772,
      "grad_norm": 4.22821388187826,
      "learning_rate": 7.923673997412678e-08,
      "loss": 8.2871,
      "step": 49
    },
    {
      "epoch": 0.006473120367673237,
      "grad_norm": 1.6934454407957527,
      "learning_rate": 8.085381630012938e-08,
      "loss": 3.7959,
      "step": 50
    },
    {
      "epoch": 0.006473120367673237,
      "eval_PRM Accuracy": 0.4056603773584906,
      "eval_PRM F1": 0.4424778761061947,
      "eval_PRM F1 AUC": 0.5419067574646412,
      "eval_PRM F1 Neg": 0.36363636363636365,
      "eval_PRM NPV": 0.23684210526315788,
      "eval_PRM Precision": 0.8333333333333334,
      "eval_PRM Recall": 0.30120481927710846,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8854758739471436,
      "eval_runtime": 15.1663,
      "eval_samples_per_second": 2.901,
      "eval_steps_per_second": 0.198,
      "step": 50
    },
    {
      "epoch": 0.006602582775026702,
      "grad_norm": 3.120989704267357,
      "learning_rate": 8.247089262613197e-08,
      "loss": 4.354,
      "step": 51
    },
    {
      "epoch": 0.006732045182380166,
      "grad_norm": 2.213437136965206,
      "learning_rate": 8.408796895213454e-08,
      "loss": 4.2153,
      "step": 52
    },
    {
      "epoch": 0.006861507589733631,
      "grad_norm": 3.187478732232664,
      "learning_rate": 8.570504527813713e-08,
      "loss": 6.2373,
      "step": 53
    },
    {
      "epoch": 0.006990969997087096,
      "grad_norm": 2.1448244699623036,
      "learning_rate": 8.732212160413972e-08,
      "loss": 5.0845,
      "step": 54
    },
    {
      "epoch": 0.00712043240444056,
      "grad_norm": 2.993483048913676,
      "learning_rate": 8.893919793014231e-08,
      "loss": 4.8887,
      "step": 55
    },
    {
      "epoch": 0.00712043240444056,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8858309388160706,
      "eval_runtime": 15.8335,
      "eval_samples_per_second": 2.779,
      "eval_steps_per_second": 0.189,
      "step": 55
    },
    {
      "epoch": 0.007249894811794025,
      "grad_norm": 3.5666108788118236,
      "learning_rate": 9.055627425614489e-08,
      "loss": 5.6782,
      "step": 56
    },
    {
      "epoch": 0.00737935721914749,
      "grad_norm": 2.2891689346612236,
      "learning_rate": 9.217335058214748e-08,
      "loss": 4.1877,
      "step": 57
    },
    {
      "epoch": 0.007508819626500955,
      "grad_norm": 2.5428379612016823,
      "learning_rate": 9.379042690815007e-08,
      "loss": 5.2524,
      "step": 58
    },
    {
      "epoch": 0.00763828203385442,
      "grad_norm": 3.8061609228898834,
      "learning_rate": 9.540750323415266e-08,
      "loss": 6.4536,
      "step": 59
    },
    {
      "epoch": 0.0077677444412078845,
      "grad_norm": 2.2419630022062567,
      "learning_rate": 9.702457956015524e-08,
      "loss": 4.6221,
      "step": 60
    },
    {
      "epoch": 0.0077677444412078845,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8837890625,
      "eval_runtime": 15.4702,
      "eval_samples_per_second": 2.844,
      "eval_steps_per_second": 0.194,
      "step": 60
    },
    {
      "epoch": 0.007897206848561348,
      "grad_norm": 1.8782241280158536,
      "learning_rate": 9.864165588615783e-08,
      "loss": 4.4683,
      "step": 61
    },
    {
      "epoch": 0.008026669255914814,
      "grad_norm": 3.570226272161863,
      "learning_rate": 1.0025873221216042e-07,
      "loss": 7.3521,
      "step": 62
    },
    {
      "epoch": 0.00815613166326828,
      "grad_norm": 2.51045834876466,
      "learning_rate": 1.0187580853816302e-07,
      "loss": 5.2793,
      "step": 63
    },
    {
      "epoch": 0.008285594070621743,
      "grad_norm": 1.5575108426738207,
      "learning_rate": 1.034928848641656e-07,
      "loss": 3.665,
      "step": 64
    },
    {
      "epoch": 0.008415056477975209,
      "grad_norm": 1.3322399221426193,
      "learning_rate": 1.0510996119016818e-07,
      "loss": 3.3931,
      "step": 65
    },
    {
      "epoch": 0.008415056477975209,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8874289989471436,
      "eval_runtime": 15.9089,
      "eval_samples_per_second": 2.766,
      "eval_steps_per_second": 0.189,
      "step": 65
    },
    {
      "epoch": 0.008544518885328672,
      "grad_norm": 2.492364990454832,
      "learning_rate": 1.0672703751617078e-07,
      "loss": 5.5532,
      "step": 66
    },
    {
      "epoch": 0.008673981292682138,
      "grad_norm": 3.7390863760396558,
      "learning_rate": 1.0834411384217337e-07,
      "loss": 6.2378,
      "step": 67
    },
    {
      "epoch": 0.008803443700035602,
      "grad_norm": 1.8473156925064862,
      "learning_rate": 1.0996119016817596e-07,
      "loss": 4.2271,
      "step": 68
    },
    {
      "epoch": 0.008932906107389067,
      "grad_norm": 1.8988855304261645,
      "learning_rate": 1.1157826649417852e-07,
      "loss": 4.6084,
      "step": 69
    },
    {
      "epoch": 0.009062368514742531,
      "grad_norm": 1.7803128991594623,
      "learning_rate": 1.1319534282018111e-07,
      "loss": 4.1704,
      "step": 70
    },
    {
      "epoch": 0.009062368514742531,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8866299986839294,
      "eval_runtime": 17.0252,
      "eval_samples_per_second": 2.584,
      "eval_steps_per_second": 0.176,
      "step": 70
    },
    {
      "epoch": 0.009191830922095997,
      "grad_norm": 3.5763120139801003,
      "learning_rate": 1.148124191461837e-07,
      "loss": 6.332,
      "step": 71
    },
    {
      "epoch": 0.00932129332944946,
      "grad_norm": 2.2141728938352876,
      "learning_rate": 1.164294954721863e-07,
      "loss": 4.2615,
      "step": 72
    },
    {
      "epoch": 0.009450755736802926,
      "grad_norm": 3.2073335886129595,
      "learning_rate": 1.1804657179818888e-07,
      "loss": 6.3149,
      "step": 73
    },
    {
      "epoch": 0.009580218144156391,
      "grad_norm": 1.3526098832419702,
      "learning_rate": 1.1966364812419147e-07,
      "loss": 3.5325,
      "step": 74
    },
    {
      "epoch": 0.009709680551509855,
      "grad_norm": 1.6799971849764905,
      "learning_rate": 1.2128072445019406e-07,
      "loss": 3.2056,
      "step": 75
    },
    {
      "epoch": 0.009709680551509855,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8876953125,
      "eval_runtime": 17.0093,
      "eval_samples_per_second": 2.587,
      "eval_steps_per_second": 0.176,
      "step": 75
    },
    {
      "epoch": 0.00983914295886332,
      "grad_norm": 2.924083764881815,
      "learning_rate": 1.2289780077619665e-07,
      "loss": 4.8882,
      "step": 76
    },
    {
      "epoch": 0.009968605366216785,
      "grad_norm": 1.5295160943757464,
      "learning_rate": 1.2451487710219921e-07,
      "loss": 3.3582,
      "step": 77
    },
    {
      "epoch": 0.01009806777357025,
      "grad_norm": 3.3154666333701295,
      "learning_rate": 1.2613195342820183e-07,
      "loss": 6.1145,
      "step": 78
    },
    {
      "epoch": 0.010227530180923714,
      "grad_norm": 3.8710027337937976,
      "learning_rate": 1.2774902975420442e-07,
      "loss": 7.8452,
      "step": 79
    },
    {
      "epoch": 0.01035699258827718,
      "grad_norm": 2.766212572050214,
      "learning_rate": 1.2936610608020701e-07,
      "loss": 5.6997,
      "step": 80
    },
    {
      "epoch": 0.01035699258827718,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.884765625,
      "eval_runtime": 15.54,
      "eval_samples_per_second": 2.831,
      "eval_steps_per_second": 0.193,
      "step": 80
    },
    {
      "epoch": 0.010486454995630643,
      "grad_norm": 3.1439318411520234,
      "learning_rate": 1.3098318240620958e-07,
      "loss": 5.9404,
      "step": 81
    },
    {
      "epoch": 0.010615917402984109,
      "grad_norm": 2.632291719511737,
      "learning_rate": 1.3260025873221217e-07,
      "loss": 4.7563,
      "step": 82
    },
    {
      "epoch": 0.010745379810337572,
      "grad_norm": 2.1378632862198614,
      "learning_rate": 1.3421733505821476e-07,
      "loss": 4.5176,
      "step": 83
    },
    {
      "epoch": 0.010874842217691038,
      "grad_norm": 3.0020895997866677,
      "learning_rate": 1.3583441138421735e-07,
      "loss": 5.7744,
      "step": 84
    },
    {
      "epoch": 0.011004304625044503,
      "grad_norm": 2.300999714939973,
      "learning_rate": 1.3745148771021992e-07,
      "loss": 4.9636,
      "step": 85
    },
    {
      "epoch": 0.011004304625044503,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8877840638160706,
      "eval_runtime": 15.4949,
      "eval_samples_per_second": 2.84,
      "eval_steps_per_second": 0.194,
      "step": 85
    },
    {
      "epoch": 0.011133767032397967,
      "grad_norm": 2.3778393912517926,
      "learning_rate": 1.390685640362225e-07,
      "loss": 5.46,
      "step": 86
    },
    {
      "epoch": 0.011263229439751433,
      "grad_norm": 2.265376177284017,
      "learning_rate": 1.406856403622251e-07,
      "loss": 4.312,
      "step": 87
    },
    {
      "epoch": 0.011392691847104897,
      "grad_norm": 1.975378983729391,
      "learning_rate": 1.423027166882277e-07,
      "loss": 3.9331,
      "step": 88
    },
    {
      "epoch": 0.011522154254458362,
      "grad_norm": 3.2796357543801578,
      "learning_rate": 1.4391979301423028e-07,
      "loss": 7.1753,
      "step": 89
    },
    {
      "epoch": 0.011651616661811826,
      "grad_norm": 2.537483324528456,
      "learning_rate": 1.4553686934023288e-07,
      "loss": 5.0693,
      "step": 90
    },
    {
      "epoch": 0.011651616661811826,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8862748742103577,
      "eval_runtime": 15.0123,
      "eval_samples_per_second": 2.931,
      "eval_steps_per_second": 0.2,
      "step": 90
    },
    {
      "epoch": 0.011781079069165291,
      "grad_norm": 1.5533230028334497,
      "learning_rate": 1.4715394566623547e-07,
      "loss": 4.3076,
      "step": 91
    },
    {
      "epoch": 0.011910541476518755,
      "grad_norm": 2.3953687079845194,
      "learning_rate": 1.4877102199223806e-07,
      "loss": 4.5845,
      "step": 92
    },
    {
      "epoch": 0.01204000388387222,
      "grad_norm": 1.6248421691564126,
      "learning_rate": 1.5038809831824065e-07,
      "loss": 3.7153,
      "step": 93
    },
    {
      "epoch": 0.012169466291225685,
      "grad_norm": 3.1074035832168847,
      "learning_rate": 1.5200517464424321e-07,
      "loss": 5.8594,
      "step": 94
    },
    {
      "epoch": 0.01229892869857915,
      "grad_norm": 1.6482149797758445,
      "learning_rate": 1.536222509702458e-07,
      "loss": 3.4629,
      "step": 95
    },
    {
      "epoch": 0.01229892869857915,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8864524364471436,
      "eval_runtime": 15.5412,
      "eval_samples_per_second": 2.831,
      "eval_steps_per_second": 0.193,
      "step": 95
    },
    {
      "epoch": 0.012428391105932616,
      "grad_norm": 3.6188297855501768,
      "learning_rate": 1.552393272962484e-07,
      "loss": 6.8918,
      "step": 96
    },
    {
      "epoch": 0.01255785351328608,
      "grad_norm": 2.7987644398773637,
      "learning_rate": 1.56856403622251e-07,
      "loss": 6.1279,
      "step": 97
    },
    {
      "epoch": 0.012687315920639545,
      "grad_norm": 2.3520379888616576,
      "learning_rate": 1.5847347994825355e-07,
      "loss": 4.769,
      "step": 98
    },
    {
      "epoch": 0.012816778327993009,
      "grad_norm": 3.246292126351779,
      "learning_rate": 1.6009055627425617e-07,
      "loss": 6.2231,
      "step": 99
    },
    {
      "epoch": 0.012946240735346474,
      "grad_norm": 2.6083710560500837,
      "learning_rate": 1.6170763260025876e-07,
      "loss": 5.2629,
      "step": 100
    },
    {
      "epoch": 0.012946240735346474,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8890269994735718,
      "eval_runtime": 14.9289,
      "eval_samples_per_second": 2.947,
      "eval_steps_per_second": 0.201,
      "step": 100
    },
    {
      "epoch": 0.013075703142699938,
      "grad_norm": 2.6541142347399296,
      "learning_rate": 1.6332470892626133e-07,
      "loss": 5.1086,
      "step": 101
    },
    {
      "epoch": 0.013205165550053403,
      "grad_norm": 1.6494228492552965,
      "learning_rate": 1.6494178525226394e-07,
      "loss": 3.917,
      "step": 102
    },
    {
      "epoch": 0.013334627957406867,
      "grad_norm": 2.921868684935309,
      "learning_rate": 1.665588615782665e-07,
      "loss": 6.1169,
      "step": 103
    },
    {
      "epoch": 0.013464090364760333,
      "grad_norm": 2.13241188421152,
      "learning_rate": 1.6817593790426907e-07,
      "loss": 4.29,
      "step": 104
    },
    {
      "epoch": 0.013593552772113798,
      "grad_norm": 1.1749121931211275,
      "learning_rate": 1.697930142302717e-07,
      "loss": 3.2705,
      "step": 105
    },
    {
      "epoch": 0.013593552772113798,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8862748742103577,
      "eval_runtime": 16.0015,
      "eval_samples_per_second": 2.75,
      "eval_steps_per_second": 0.187,
      "step": 105
    },
    {
      "epoch": 0.013723015179467262,
      "grad_norm": 1.575721321415035,
      "learning_rate": 1.7141009055627426e-07,
      "loss": 3.2231,
      "step": 106
    },
    {
      "epoch": 0.013852477586820728,
      "grad_norm": 3.1740074105498346,
      "learning_rate": 1.7302716688227687e-07,
      "loss": 6.1416,
      "step": 107
    },
    {
      "epoch": 0.013981939994174191,
      "grad_norm": 2.8235795914302306,
      "learning_rate": 1.7464424320827944e-07,
      "loss": 6.5957,
      "step": 108
    },
    {
      "epoch": 0.014111402401527657,
      "grad_norm": 2.7800697021429106,
      "learning_rate": 1.7626131953428203e-07,
      "loss": 5.5332,
      "step": 109
    },
    {
      "epoch": 0.01424086480888112,
      "grad_norm": 1.9166248786656255,
      "learning_rate": 1.7787839586028462e-07,
      "loss": 4.1089,
      "step": 110
    },
    {
      "epoch": 0.01424086480888112,
      "eval_PRM Accuracy": 0.4056603773584906,
      "eval_PRM F1": 0.4424778761061947,
      "eval_PRM F1 AUC": 0.5419067574646412,
      "eval_PRM F1 Neg": 0.36363636363636365,
      "eval_PRM NPV": 0.23684210526315788,
      "eval_PRM Precision": 0.8333333333333334,
      "eval_PRM Recall": 0.30120481927710846,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8856534361839294,
      "eval_runtime": 14.6177,
      "eval_samples_per_second": 3.01,
      "eval_steps_per_second": 0.205,
      "step": 110
    },
    {
      "epoch": 0.014370327216234586,
      "grad_norm": 2.7418573899602325,
      "learning_rate": 1.7949547218628721e-07,
      "loss": 5.448,
      "step": 111
    },
    {
      "epoch": 0.01449978962358805,
      "grad_norm": 1.791130152898085,
      "learning_rate": 1.8111254851228978e-07,
      "loss": 3.7705,
      "step": 112
    },
    {
      "epoch": 0.014629252030941516,
      "grad_norm": 1.7712886879433232,
      "learning_rate": 1.827296248382924e-07,
      "loss": 4.0708,
      "step": 113
    },
    {
      "epoch": 0.01475871443829498,
      "grad_norm": 1.4011926721258692,
      "learning_rate": 1.8434670116429496e-07,
      "loss": 3.3662,
      "step": 114
    },
    {
      "epoch": 0.014888176845648445,
      "grad_norm": 2.5362570649868417,
      "learning_rate": 1.8596377749029758e-07,
      "loss": 4.9705,
      "step": 115
    },
    {
      "epoch": 0.014888176845648445,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8846768736839294,
      "eval_runtime": 16.422,
      "eval_samples_per_second": 2.679,
      "eval_steps_per_second": 0.183,
      "step": 115
    },
    {
      "epoch": 0.01501763925300191,
      "grad_norm": 2.0416731238959525,
      "learning_rate": 1.8758085381630014e-07,
      "loss": 3.9224,
      "step": 116
    },
    {
      "epoch": 0.015147101660355374,
      "grad_norm": 2.4303010252138733,
      "learning_rate": 1.891979301423027e-07,
      "loss": 5.5867,
      "step": 117
    },
    {
      "epoch": 0.01527656406770884,
      "grad_norm": 2.913544758821116,
      "learning_rate": 1.9081500646830533e-07,
      "loss": 5.1328,
      "step": 118
    },
    {
      "epoch": 0.015406026475062303,
      "grad_norm": 2.0386331121658685,
      "learning_rate": 1.924320827943079e-07,
      "loss": 4.9512,
      "step": 119
    },
    {
      "epoch": 0.015535488882415769,
      "grad_norm": 2.6467398937578515,
      "learning_rate": 1.9404915912031048e-07,
      "loss": 5.2549,
      "step": 120
    },
    {
      "epoch": 0.015535488882415769,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8829900622367859,
      "eval_runtime": 15.9024,
      "eval_samples_per_second": 2.767,
      "eval_steps_per_second": 0.189,
      "step": 120
    },
    {
      "epoch": 0.015664951289769233,
      "grad_norm": 2.4626541900261922,
      "learning_rate": 1.9566623544631307e-07,
      "loss": 4.1807,
      "step": 121
    },
    {
      "epoch": 0.015794413697122697,
      "grad_norm": 1.8878277717350793,
      "learning_rate": 1.9728331177231567e-07,
      "loss": 3.5337,
      "step": 122
    },
    {
      "epoch": 0.015923876104476164,
      "grad_norm": 2.8574871996520304,
      "learning_rate": 1.9890038809831828e-07,
      "loss": 5.0161,
      "step": 123
    },
    {
      "epoch": 0.016053338511829628,
      "grad_norm": 1.9082654564956048,
      "learning_rate": 2.0051746442432085e-07,
      "loss": 4.3105,
      "step": 124
    },
    {
      "epoch": 0.01618280091918309,
      "grad_norm": 2.210541489132059,
      "learning_rate": 2.021345407503234e-07,
      "loss": 4.6948,
      "step": 125
    },
    {
      "epoch": 0.01618280091918309,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8853870630264282,
      "eval_runtime": 15.0559,
      "eval_samples_per_second": 2.922,
      "eval_steps_per_second": 0.199,
      "step": 125
    },
    {
      "epoch": 0.01631226332653656,
      "grad_norm": 2.7336759960768435,
      "learning_rate": 2.0375161707632603e-07,
      "loss": 5.4995,
      "step": 126
    },
    {
      "epoch": 0.016441725733890022,
      "grad_norm": 2.244046155368536,
      "learning_rate": 2.053686934023286e-07,
      "loss": 4.2437,
      "step": 127
    },
    {
      "epoch": 0.016571188141243486,
      "grad_norm": 2.45262536092704,
      "learning_rate": 2.069857697283312e-07,
      "loss": 4.8989,
      "step": 128
    },
    {
      "epoch": 0.01670065054859695,
      "grad_norm": 3.3040287729974747,
      "learning_rate": 2.0860284605433378e-07,
      "loss": 6.6123,
      "step": 129
    },
    {
      "epoch": 0.016830112955950417,
      "grad_norm": 2.9261239742898564,
      "learning_rate": 2.1021992238033637e-07,
      "loss": 4.8569,
      "step": 130
    },
    {
      "epoch": 0.016830112955950417,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8856534361839294,
      "eval_runtime": 15.5434,
      "eval_samples_per_second": 2.831,
      "eval_steps_per_second": 0.193,
      "step": 130
    },
    {
      "epoch": 0.01695957536330388,
      "grad_norm": 1.8669972552610263,
      "learning_rate": 2.1183699870633896e-07,
      "loss": 4.4214,
      "step": 131
    },
    {
      "epoch": 0.017089037770657345,
      "grad_norm": 2.8355338022322005,
      "learning_rate": 2.1345407503234155e-07,
      "loss": 5.6006,
      "step": 132
    },
    {
      "epoch": 0.01721850017801081,
      "grad_norm": 3.8775566026344146,
      "learning_rate": 2.1507115135834412e-07,
      "loss": 6.9272,
      "step": 133
    },
    {
      "epoch": 0.017347962585364276,
      "grad_norm": 2.1906529658414975,
      "learning_rate": 2.1668822768434673e-07,
      "loss": 4.6504,
      "step": 134
    },
    {
      "epoch": 0.01747742499271774,
      "grad_norm": 1.7570863207219325,
      "learning_rate": 2.183053040103493e-07,
      "loss": 4.4053,
      "step": 135
    },
    {
      "epoch": 0.01747742499271774,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8849431872367859,
      "eval_runtime": 15.3918,
      "eval_samples_per_second": 2.859,
      "eval_steps_per_second": 0.195,
      "step": 135
    },
    {
      "epoch": 0.017606887400071203,
      "grad_norm": 1.3841515491996987,
      "learning_rate": 2.1992238033635192e-07,
      "loss": 3.7124,
      "step": 136
    },
    {
      "epoch": 0.01773634980742467,
      "grad_norm": 1.178526476409248,
      "learning_rate": 2.2153945666235448e-07,
      "loss": 3.4648,
      "step": 137
    },
    {
      "epoch": 0.017865812214778135,
      "grad_norm": 3.1071071039731044,
      "learning_rate": 2.2315653298835705e-07,
      "loss": 6.1064,
      "step": 138
    },
    {
      "epoch": 0.0179952746221316,
      "grad_norm": 2.5958202424020924,
      "learning_rate": 2.2477360931435966e-07,
      "loss": 5.0176,
      "step": 139
    },
    {
      "epoch": 0.018124737029485062,
      "grad_norm": 3.5329237791889043,
      "learning_rate": 2.2639068564036223e-07,
      "loss": 7.2939,
      "step": 140
    },
    {
      "epoch": 0.018124737029485062,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8851207494735718,
      "eval_runtime": 15.4656,
      "eval_samples_per_second": 2.845,
      "eval_steps_per_second": 0.194,
      "step": 140
    },
    {
      "epoch": 0.01825419943683853,
      "grad_norm": 1.698099498401497,
      "learning_rate": 2.2800776196636482e-07,
      "loss": 3.9492,
      "step": 141
    },
    {
      "epoch": 0.018383661844191993,
      "grad_norm": 2.0444575058001493,
      "learning_rate": 2.296248382923674e-07,
      "loss": 4.1685,
      "step": 142
    },
    {
      "epoch": 0.018513124251545457,
      "grad_norm": 2.420495782296046,
      "learning_rate": 2.3124191461837e-07,
      "loss": 5.0999,
      "step": 143
    },
    {
      "epoch": 0.01864258665889892,
      "grad_norm": 2.2734740240960436,
      "learning_rate": 2.328589909443726e-07,
      "loss": 4.312,
      "step": 144
    },
    {
      "epoch": 0.018772049066252388,
      "grad_norm": 2.5306624787049286,
      "learning_rate": 2.3447606727037519e-07,
      "loss": 4.9043,
      "step": 145
    },
    {
      "epoch": 0.018772049066252388,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8851207494735718,
      "eval_runtime": 16.3471,
      "eval_samples_per_second": 2.692,
      "eval_steps_per_second": 0.184,
      "step": 145
    },
    {
      "epoch": 0.018901511473605852,
      "grad_norm": 1.4008106167888073,
      "learning_rate": 2.3609314359637775e-07,
      "loss": 3.7585,
      "step": 146
    },
    {
      "epoch": 0.019030973880959316,
      "grad_norm": 3.7456861553981318,
      "learning_rate": 2.3771021992238037e-07,
      "loss": 7.1523,
      "step": 147
    },
    {
      "epoch": 0.019160436288312783,
      "grad_norm": 2.641442840623338,
      "learning_rate": 2.3932729624838293e-07,
      "loss": 5.2881,
      "step": 148
    },
    {
      "epoch": 0.019289898695666247,
      "grad_norm": 1.9569795328315067,
      "learning_rate": 2.409443725743855e-07,
      "loss": 4.395,
      "step": 149
    },
    {
      "epoch": 0.01941936110301971,
      "grad_norm": 2.3216166306121555,
      "learning_rate": 2.425614489003881e-07,
      "loss": 5.334,
      "step": 150
    },
    {
      "epoch": 0.01941936110301971,
      "eval_PRM Accuracy": 0.4056603773584906,
      "eval_PRM F1": 0.4424778761061947,
      "eval_PRM F1 AUC": 0.5419067574646412,
      "eval_PRM F1 Neg": 0.36363636363636365,
      "eval_PRM NPV": 0.23684210526315788,
      "eval_PRM Precision": 0.8333333333333334,
      "eval_PRM Recall": 0.30120481927710846,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8857421875,
      "eval_runtime": 15.7514,
      "eval_samples_per_second": 2.793,
      "eval_steps_per_second": 0.19,
      "step": 150
    },
    {
      "epoch": 0.019548823510373174,
      "grad_norm": 2.1971180558555403,
      "learning_rate": 2.441785252263907e-07,
      "loss": 4.4736,
      "step": 151
    },
    {
      "epoch": 0.01967828591772664,
      "grad_norm": 3.150682034268968,
      "learning_rate": 2.457956015523933e-07,
      "loss": 6.6304,
      "step": 152
    },
    {
      "epoch": 0.019807748325080105,
      "grad_norm": 1.7781467557657258,
      "learning_rate": 2.4741267787839586e-07,
      "loss": 4.0156,
      "step": 153
    },
    {
      "epoch": 0.01993721073243357,
      "grad_norm": 1.9218504498599223,
      "learning_rate": 2.4902975420439843e-07,
      "loss": 4.6631,
      "step": 154
    },
    {
      "epoch": 0.020066673139787033,
      "grad_norm": 3.0912369033963416,
      "learning_rate": 2.5064683053040105e-07,
      "loss": 5.2312,
      "step": 155
    },
    {
      "epoch": 0.020066673139787033,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8844993114471436,
      "eval_runtime": 14.5707,
      "eval_samples_per_second": 3.02,
      "eval_steps_per_second": 0.206,
      "step": 155
    },
    {
      "epoch": 0.0201961355471405,
      "grad_norm": 3.374085982719268,
      "learning_rate": 2.5226390685640366e-07,
      "loss": 4.8442,
      "step": 156
    },
    {
      "epoch": 0.020325597954493964,
      "grad_norm": 2.6023678787454925,
      "learning_rate": 2.5388098318240623e-07,
      "loss": 5.2119,
      "step": 157
    },
    {
      "epoch": 0.020455060361847428,
      "grad_norm": 3.106650050358539,
      "learning_rate": 2.5549805950840885e-07,
      "loss": 5.978,
      "step": 158
    },
    {
      "epoch": 0.020584522769200895,
      "grad_norm": 2.240948031320514,
      "learning_rate": 2.571151358344114e-07,
      "loss": 4.7949,
      "step": 159
    },
    {
      "epoch": 0.02071398517655436,
      "grad_norm": 1.7837401909792663,
      "learning_rate": 2.5873221216041403e-07,
      "loss": 4.1484,
      "step": 160
    },
    {
      "epoch": 0.02071398517655436,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8870738744735718,
      "eval_runtime": 15.8666,
      "eval_samples_per_second": 2.773,
      "eval_steps_per_second": 0.189,
      "step": 160
    },
    {
      "epoch": 0.020843447583907822,
      "grad_norm": 2.772672736542775,
      "learning_rate": 2.603492884864166e-07,
      "loss": 4.7573,
      "step": 161
    },
    {
      "epoch": 0.020972909991261286,
      "grad_norm": 1.3714409296270307,
      "learning_rate": 2.6196636481241916e-07,
      "loss": 3.5115,
      "step": 162
    },
    {
      "epoch": 0.021102372398614754,
      "grad_norm": 2.031963096215658,
      "learning_rate": 2.635834411384218e-07,
      "loss": 4.1006,
      "step": 163
    },
    {
      "epoch": 0.021231834805968217,
      "grad_norm": 2.1993468910076066,
      "learning_rate": 2.6520051746442434e-07,
      "loss": 4.301,
      "step": 164
    },
    {
      "epoch": 0.02136129721332168,
      "grad_norm": 2.259418281777508,
      "learning_rate": 2.6681759379042696e-07,
      "loss": 4.8267,
      "step": 165
    },
    {
      "epoch": 0.02136129721332168,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8838778138160706,
      "eval_runtime": 15.4824,
      "eval_samples_per_second": 2.842,
      "eval_steps_per_second": 0.194,
      "step": 165
    },
    {
      "epoch": 0.021490759620675145,
      "grad_norm": 3.73088511507556,
      "learning_rate": 2.684346701164295e-07,
      "loss": 6.0649,
      "step": 166
    },
    {
      "epoch": 0.021620222028028612,
      "grad_norm": 2.6557645389732873,
      "learning_rate": 2.700517464424321e-07,
      "loss": 4.6919,
      "step": 167
    },
    {
      "epoch": 0.021749684435382076,
      "grad_norm": 1.997316671729813,
      "learning_rate": 2.716688227684347e-07,
      "loss": 4.7817,
      "step": 168
    },
    {
      "epoch": 0.02187914684273554,
      "grad_norm": 1.9657324745031033,
      "learning_rate": 2.7328589909443727e-07,
      "loss": 4.4263,
      "step": 169
    },
    {
      "epoch": 0.022008609250089007,
      "grad_norm": 2.4627709291744835,
      "learning_rate": 2.7490297542043984e-07,
      "loss": 5.0391,
      "step": 170
    },
    {
      "epoch": 0.022008609250089007,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8838778138160706,
      "eval_runtime": 14.1935,
      "eval_samples_per_second": 3.1,
      "eval_steps_per_second": 0.211,
      "step": 170
    },
    {
      "epoch": 0.02213807165744247,
      "grad_norm": 3.052788131039653,
      "learning_rate": 2.7652005174644245e-07,
      "loss": 4.7004,
      "step": 171
    },
    {
      "epoch": 0.022267534064795935,
      "grad_norm": 2.6186961622702554,
      "learning_rate": 2.78137128072445e-07,
      "loss": 4.7441,
      "step": 172
    },
    {
      "epoch": 0.0223969964721494,
      "grad_norm": 3.72505733347434,
      "learning_rate": 2.7975420439844764e-07,
      "loss": 7.8696,
      "step": 173
    },
    {
      "epoch": 0.022526458879502866,
      "grad_norm": 1.9464865151658428,
      "learning_rate": 2.813712807244502e-07,
      "loss": 3.6841,
      "step": 174
    },
    {
      "epoch": 0.02265592128685633,
      "grad_norm": 3.0311616691881302,
      "learning_rate": 2.8298835705045277e-07,
      "loss": 5.6128,
      "step": 175
    },
    {
      "epoch": 0.02265592128685633,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8851207494735718,
      "eval_runtime": 14.8654,
      "eval_samples_per_second": 2.96,
      "eval_steps_per_second": 0.202,
      "step": 175
    },
    {
      "epoch": 0.022785383694209793,
      "grad_norm": 3.2153799248538815,
      "learning_rate": 2.846054333764554e-07,
      "loss": 5.5913,
      "step": 176
    },
    {
      "epoch": 0.022914846101563257,
      "grad_norm": 3.1273926216154497,
      "learning_rate": 2.86222509702458e-07,
      "loss": 5.6323,
      "step": 177
    },
    {
      "epoch": 0.023044308508916724,
      "grad_norm": 2.723639596237429,
      "learning_rate": 2.8783958602846057e-07,
      "loss": 5.3838,
      "step": 178
    },
    {
      "epoch": 0.023173770916270188,
      "grad_norm": 2.7628886706248044,
      "learning_rate": 2.894566623544632e-07,
      "loss": 5.126,
      "step": 179
    },
    {
      "epoch": 0.023303233323623652,
      "grad_norm": 2.3460860198951368,
      "learning_rate": 2.9107373868046575e-07,
      "loss": 4.2046,
      "step": 180
    },
    {
      "epoch": 0.023303233323623652,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8838778138160706,
      "eval_runtime": 14.4704,
      "eval_samples_per_second": 3.041,
      "eval_steps_per_second": 0.207,
      "step": 180
    },
    {
      "epoch": 0.02343269573097712,
      "grad_norm": 1.8977281350603303,
      "learning_rate": 2.9269081500646837e-07,
      "loss": 3.7952,
      "step": 181
    },
    {
      "epoch": 0.023562158138330583,
      "grad_norm": 4.633373815126988,
      "learning_rate": 2.9430789133247093e-07,
      "loss": 8.4346,
      "step": 182
    },
    {
      "epoch": 0.023691620545684047,
      "grad_norm": 1.3213131050692126,
      "learning_rate": 2.959249676584735e-07,
      "loss": 3.7607,
      "step": 183
    },
    {
      "epoch": 0.02382108295303751,
      "grad_norm": 1.990277012223307,
      "learning_rate": 2.975420439844761e-07,
      "loss": 4.9785,
      "step": 184
    },
    {
      "epoch": 0.023950545360390978,
      "grad_norm": 2.239661015493053,
      "learning_rate": 2.991591203104787e-07,
      "loss": 5.7068,
      "step": 185
    },
    {
      "epoch": 0.023950545360390978,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8870738744735718,
      "eval_runtime": 14.5341,
      "eval_samples_per_second": 3.027,
      "eval_steps_per_second": 0.206,
      "step": 185
    },
    {
      "epoch": 0.02408000776774444,
      "grad_norm": 1.7206395570915416,
      "learning_rate": 3.007761966364813e-07,
      "loss": 4.1523,
      "step": 186
    },
    {
      "epoch": 0.024209470175097905,
      "grad_norm": 2.703519646145678,
      "learning_rate": 3.0239327296248386e-07,
      "loss": 5.522,
      "step": 187
    },
    {
      "epoch": 0.02433893258245137,
      "grad_norm": 1.8351375522927664,
      "learning_rate": 3.0401034928848643e-07,
      "loss": 4.0796,
      "step": 188
    },
    {
      "epoch": 0.024468394989804836,
      "grad_norm": 3.7534253262775192,
      "learning_rate": 3.0562742561448905e-07,
      "loss": 7.3333,
      "step": 189
    },
    {
      "epoch": 0.0245978573971583,
      "grad_norm": 3.6605853981413317,
      "learning_rate": 3.072445019404916e-07,
      "loss": 4.0422,
      "step": 190
    },
    {
      "epoch": 0.0245978573971583,
      "eval_PRM Accuracy": 0.4056603773584906,
      "eval_PRM F1": 0.4424778761061947,
      "eval_PRM F1 AUC": 0.5419067574646412,
      "eval_PRM F1 Neg": 0.36363636363636365,
      "eval_PRM NPV": 0.23684210526315788,
      "eval_PRM Precision": 0.8333333333333334,
      "eval_PRM Recall": 0.30120481927710846,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8857421875,
      "eval_runtime": 15.5331,
      "eval_samples_per_second": 2.833,
      "eval_steps_per_second": 0.193,
      "step": 190
    },
    {
      "epoch": 0.024727319804511764,
      "grad_norm": 2.199660582380855,
      "learning_rate": 3.088615782664942e-07,
      "loss": 5.0649,
      "step": 191
    },
    {
      "epoch": 0.02485678221186523,
      "grad_norm": 2.023418667320133,
      "learning_rate": 3.104786545924968e-07,
      "loss": 3.8857,
      "step": 192
    },
    {
      "epoch": 0.024986244619218695,
      "grad_norm": 1.560125306308657,
      "learning_rate": 3.1209573091849936e-07,
      "loss": 3.4775,
      "step": 193
    },
    {
      "epoch": 0.02511570702657216,
      "grad_norm": 2.1055469781306724,
      "learning_rate": 3.13712807244502e-07,
      "loss": 4.6436,
      "step": 194
    },
    {
      "epoch": 0.025245169433925622,
      "grad_norm": 2.1623865960455886,
      "learning_rate": 3.1532988357050454e-07,
      "loss": 4.2874,
      "step": 195
    },
    {
      "epoch": 0.025245169433925622,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8839666247367859,
      "eval_runtime": 14.4616,
      "eval_samples_per_second": 3.043,
      "eval_steps_per_second": 0.207,
      "step": 195
    },
    {
      "epoch": 0.02537463184127909,
      "grad_norm": 2.0618961739343296,
      "learning_rate": 3.169469598965071e-07,
      "loss": 4.7808,
      "step": 196
    },
    {
      "epoch": 0.025504094248632554,
      "grad_norm": 2.5836308429390726,
      "learning_rate": 3.185640362225097e-07,
      "loss": 4.6035,
      "step": 197
    },
    {
      "epoch": 0.025633556655986017,
      "grad_norm": 1.9751794727575602,
      "learning_rate": 3.2018111254851234e-07,
      "loss": 3.7432,
      "step": 198
    },
    {
      "epoch": 0.02576301906333948,
      "grad_norm": 1.4354870173720242,
      "learning_rate": 3.217981888745149e-07,
      "loss": 3.8357,
      "step": 199
    },
    {
      "epoch": 0.02589248147069295,
      "grad_norm": 3.6734925154622338,
      "learning_rate": 3.234152652005175e-07,
      "loss": 6.9282,
      "step": 200
    },
    {
      "epoch": 0.02589248147069295,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8876065611839294,
      "eval_runtime": 14.5467,
      "eval_samples_per_second": 3.025,
      "eval_steps_per_second": 0.206,
      "step": 200
    },
    {
      "epoch": 0.026021943878046412,
      "grad_norm": 2.473074742494256,
      "learning_rate": 3.250323415265201e-07,
      "loss": 5.3984,
      "step": 201
    },
    {
      "epoch": 0.026151406285399876,
      "grad_norm": 2.5287299002672285,
      "learning_rate": 3.2664941785252265e-07,
      "loss": 5.5527,
      "step": 202
    },
    {
      "epoch": 0.026280868692753343,
      "grad_norm": 2.456362736987199,
      "learning_rate": 3.2826649417852527e-07,
      "loss": 4.6409,
      "step": 203
    },
    {
      "epoch": 0.026410331100106807,
      "grad_norm": 1.9215058395998181,
      "learning_rate": 3.298835705045279e-07,
      "loss": 4.1421,
      "step": 204
    },
    {
      "epoch": 0.02653979350746027,
      "grad_norm": 3.9408260148033922,
      "learning_rate": 3.315006468305304e-07,
      "loss": 7.333,
      "step": 205
    },
    {
      "epoch": 0.02653979350746027,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.884765625,
      "eval_runtime": 13.8341,
      "eval_samples_per_second": 3.181,
      "eval_steps_per_second": 0.217,
      "step": 205
    },
    {
      "epoch": 0.026669255914813735,
      "grad_norm": 3.5973768278716283,
      "learning_rate": 3.33117723156533e-07,
      "loss": 6.6558,
      "step": 206
    },
    {
      "epoch": 0.026798718322167202,
      "grad_norm": 2.7321593067095344,
      "learning_rate": 3.3473479948253564e-07,
      "loss": 4.6677,
      "step": 207
    },
    {
      "epoch": 0.026928180729520666,
      "grad_norm": 2.745785213443762,
      "learning_rate": 3.3635187580853815e-07,
      "loss": 4.5298,
      "step": 208
    },
    {
      "epoch": 0.02705764313687413,
      "grad_norm": 2.362963418655679,
      "learning_rate": 3.3796895213454077e-07,
      "loss": 4.4668,
      "step": 209
    },
    {
      "epoch": 0.027187105544227597,
      "grad_norm": 2.170885869291569,
      "learning_rate": 3.395860284605434e-07,
      "loss": 4.5439,
      "step": 210
    },
    {
      "epoch": 0.027187105544227597,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8869850635528564,
      "eval_runtime": 16.0254,
      "eval_samples_per_second": 2.746,
      "eval_steps_per_second": 0.187,
      "step": 210
    },
    {
      "epoch": 0.02731656795158106,
      "grad_norm": 2.8036859242240992,
      "learning_rate": 3.412031047865459e-07,
      "loss": 5.0664,
      "step": 211
    },
    {
      "epoch": 0.027446030358934524,
      "grad_norm": 1.9738146255572215,
      "learning_rate": 3.428201811125485e-07,
      "loss": 4.4531,
      "step": 212
    },
    {
      "epoch": 0.027575492766287988,
      "grad_norm": 3.7762534844913676,
      "learning_rate": 3.4443725743855113e-07,
      "loss": 6.1953,
      "step": 213
    },
    {
      "epoch": 0.027704955173641455,
      "grad_norm": 1.3922138159878588,
      "learning_rate": 3.4605433376455375e-07,
      "loss": 3.5786,
      "step": 214
    },
    {
      "epoch": 0.02783441758099492,
      "grad_norm": 1.1881268842261485,
      "learning_rate": 3.4767141009055626e-07,
      "loss": 2.936,
      "step": 215
    },
    {
      "epoch": 0.02783441758099492,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8840553760528564,
      "eval_runtime": 14.9814,
      "eval_samples_per_second": 2.937,
      "eval_steps_per_second": 0.2,
      "step": 215
    },
    {
      "epoch": 0.027963879988348383,
      "grad_norm": 2.518946321063278,
      "learning_rate": 3.492884864165589e-07,
      "loss": 5.0696,
      "step": 216
    },
    {
      "epoch": 0.028093342395701847,
      "grad_norm": 2.152032693026809,
      "learning_rate": 3.509055627425615e-07,
      "loss": 5.2261,
      "step": 217
    },
    {
      "epoch": 0.028222804803055314,
      "grad_norm": 2.9739441597144705,
      "learning_rate": 3.5252263906856406e-07,
      "loss": 6.2876,
      "step": 218
    },
    {
      "epoch": 0.028352267210408778,
      "grad_norm": 1.9749725553254946,
      "learning_rate": 3.541397153945666e-07,
      "loss": 4.2852,
      "step": 219
    },
    {
      "epoch": 0.02848172961776224,
      "grad_norm": 2.504483731096505,
      "learning_rate": 3.5575679172056924e-07,
      "loss": 4.562,
      "step": 220
    },
    {
      "epoch": 0.02848172961776224,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.884765625,
      "eval_runtime": 14.3789,
      "eval_samples_per_second": 3.06,
      "eval_steps_per_second": 0.209,
      "step": 220
    },
    {
      "epoch": 0.02861119202511571,
      "grad_norm": 3.237303782437478,
      "learning_rate": 3.573738680465718e-07,
      "loss": 6.4438,
      "step": 221
    },
    {
      "epoch": 0.028740654432469172,
      "grad_norm": 2.9563482106484487,
      "learning_rate": 3.5899094437257443e-07,
      "loss": 5.6177,
      "step": 222
    },
    {
      "epoch": 0.028870116839822636,
      "grad_norm": 2.169875290961729,
      "learning_rate": 3.6060802069857704e-07,
      "loss": 4.7012,
      "step": 223
    },
    {
      "epoch": 0.0289995792471761,
      "grad_norm": 3.097390281971058,
      "learning_rate": 3.6222509702457956e-07,
      "loss": 4.9597,
      "step": 224
    },
    {
      "epoch": 0.029129041654529567,
      "grad_norm": 1.1916093538628139,
      "learning_rate": 3.638421733505822e-07,
      "loss": 3.1731,
      "step": 225
    },
    {
      "epoch": 0.029129041654529567,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8854758739471436,
      "eval_runtime": 15.49,
      "eval_samples_per_second": 2.841,
      "eval_steps_per_second": 0.194,
      "step": 225
    },
    {
      "epoch": 0.02925850406188303,
      "grad_norm": 2.5146257457979346,
      "learning_rate": 3.654592496765848e-07,
      "loss": 4.9551,
      "step": 226
    },
    {
      "epoch": 0.029387966469236495,
      "grad_norm": 2.6612886213539113,
      "learning_rate": 3.670763260025874e-07,
      "loss": 4.5801,
      "step": 227
    },
    {
      "epoch": 0.02951742887658996,
      "grad_norm": 2.350864586600711,
      "learning_rate": 3.686934023285899e-07,
      "loss": 4.4111,
      "step": 228
    },
    {
      "epoch": 0.029646891283943426,
      "grad_norm": 2.9354493578714655,
      "learning_rate": 3.7031047865459254e-07,
      "loss": 4.6353,
      "step": 229
    },
    {
      "epoch": 0.02977635369129689,
      "grad_norm": 2.10405268997446,
      "learning_rate": 3.7192755498059516e-07,
      "loss": 3.8438,
      "step": 230
    },
    {
      "epoch": 0.02977635369129689,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8859197497367859,
      "eval_runtime": 14.6915,
      "eval_samples_per_second": 2.995,
      "eval_steps_per_second": 0.204,
      "step": 230
    },
    {
      "epoch": 0.029905816098650354,
      "grad_norm": 2.796776889785926,
      "learning_rate": 3.7354463130659767e-07,
      "loss": 5.1143,
      "step": 231
    },
    {
      "epoch": 0.03003527850600382,
      "grad_norm": 2.540196823573686,
      "learning_rate": 3.751617076326003e-07,
      "loss": 4.4585,
      "step": 232
    },
    {
      "epoch": 0.030164740913357285,
      "grad_norm": 1.9734948597315027,
      "learning_rate": 3.767787839586029e-07,
      "loss": 4.624,
      "step": 233
    },
    {
      "epoch": 0.03029420332071075,
      "grad_norm": 1.413122722618321,
      "learning_rate": 3.783958602846054e-07,
      "loss": 3.5017,
      "step": 234
    },
    {
      "epoch": 0.030423665728064212,
      "grad_norm": 1.6196061176180687,
      "learning_rate": 3.8001293661060803e-07,
      "loss": 3.5574,
      "step": 235
    },
    {
      "epoch": 0.030423665728064212,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8846768736839294,
      "eval_runtime": 14.4343,
      "eval_samples_per_second": 3.048,
      "eval_steps_per_second": 0.208,
      "step": 235
    },
    {
      "epoch": 0.03055312813541768,
      "grad_norm": 1.9251633527239032,
      "learning_rate": 3.8163001293661065e-07,
      "loss": 4.2485,
      "step": 236
    },
    {
      "epoch": 0.030682590542771143,
      "grad_norm": 2.86788432030648,
      "learning_rate": 3.832470892626132e-07,
      "loss": 5.1436,
      "step": 237
    },
    {
      "epoch": 0.030812052950124607,
      "grad_norm": 3.4081952293967217,
      "learning_rate": 3.848641655886158e-07,
      "loss": 5.4128,
      "step": 238
    },
    {
      "epoch": 0.03094151535747807,
      "grad_norm": 1.0069334503784966,
      "learning_rate": 3.864812419146184e-07,
      "loss": 2.7949,
      "step": 239
    },
    {
      "epoch": 0.031070977764831538,
      "grad_norm": 2.54531275256124,
      "learning_rate": 3.8809831824062096e-07,
      "loss": 5.3416,
      "step": 240
    },
    {
      "epoch": 0.031070977764831538,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8854758739471436,
      "eval_runtime": 15.2595,
      "eval_samples_per_second": 2.883,
      "eval_steps_per_second": 0.197,
      "step": 240
    },
    {
      "epoch": 0.031200440172185002,
      "grad_norm": 1.8770741090775689,
      "learning_rate": 3.897153945666236e-07,
      "loss": 3.8792,
      "step": 241
    },
    {
      "epoch": 0.031329902579538466,
      "grad_norm": 2.9652913430837136,
      "learning_rate": 3.9133247089262615e-07,
      "loss": 6.0073,
      "step": 242
    },
    {
      "epoch": 0.03145936498689193,
      "grad_norm": 4.106740705941837,
      "learning_rate": 3.9294954721862877e-07,
      "loss": 7.5737,
      "step": 243
    },
    {
      "epoch": 0.03158882739424539,
      "grad_norm": 2.0455585729583854,
      "learning_rate": 3.9456662354463133e-07,
      "loss": 4.6748,
      "step": 244
    },
    {
      "epoch": 0.03171828980159886,
      "grad_norm": 1.8645159203862933,
      "learning_rate": 3.9618369987063395e-07,
      "loss": 3.7271,
      "step": 245
    },
    {
      "epoch": 0.03171828980159886,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8874289989471436,
      "eval_runtime": 15.0412,
      "eval_samples_per_second": 2.925,
      "eval_steps_per_second": 0.199,
      "step": 245
    },
    {
      "epoch": 0.03184775220895233,
      "grad_norm": 2.0926513605013555,
      "learning_rate": 3.9780077619663657e-07,
      "loss": 4.0507,
      "step": 246
    },
    {
      "epoch": 0.03197721461630579,
      "grad_norm": 2.7296823858108135,
      "learning_rate": 3.994178525226391e-07,
      "loss": 5.2227,
      "step": 247
    },
    {
      "epoch": 0.032106677023659255,
      "grad_norm": 3.878143159601418,
      "learning_rate": 4.010349288486417e-07,
      "loss": 6.4014,
      "step": 248
    },
    {
      "epoch": 0.03223613943101272,
      "grad_norm": 2.309639696039142,
      "learning_rate": 4.026520051746443e-07,
      "loss": 5.0869,
      "step": 249
    },
    {
      "epoch": 0.03236560183836618,
      "grad_norm": 1.4112146486784303,
      "learning_rate": 4.042690815006468e-07,
      "loss": 3.1626,
      "step": 250
    },
    {
      "epoch": 0.03236560183836618,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8870738744735718,
      "eval_runtime": 14.9055,
      "eval_samples_per_second": 2.952,
      "eval_steps_per_second": 0.201,
      "step": 250
    },
    {
      "epoch": 0.03249506424571965,
      "grad_norm": 1.8751606104067502,
      "learning_rate": 4.0588615782664944e-07,
      "loss": 4.4971,
      "step": 251
    },
    {
      "epoch": 0.03262452665307312,
      "grad_norm": 2.579439703784705,
      "learning_rate": 4.0750323415265206e-07,
      "loss": 4.5474,
      "step": 252
    },
    {
      "epoch": 0.03275398906042658,
      "grad_norm": 1.9307201135550025,
      "learning_rate": 4.0912031047865457e-07,
      "loss": 3.6321,
      "step": 253
    },
    {
      "epoch": 0.032883451467780045,
      "grad_norm": 2.8048169313003037,
      "learning_rate": 4.107373868046572e-07,
      "loss": 6.1855,
      "step": 254
    },
    {
      "epoch": 0.033012913875133505,
      "grad_norm": 1.9406061976658442,
      "learning_rate": 4.123544631306598e-07,
      "loss": 4.2798,
      "step": 255
    },
    {
      "epoch": 0.033012913875133505,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8869850635528564,
      "eval_runtime": 14.5647,
      "eval_samples_per_second": 3.021,
      "eval_steps_per_second": 0.206,
      "step": 255
    },
    {
      "epoch": 0.03314237628248697,
      "grad_norm": 2.8163829206102613,
      "learning_rate": 4.139715394566624e-07,
      "loss": 4.9077,
      "step": 256
    },
    {
      "epoch": 0.03327183868984044,
      "grad_norm": 1.6884950389012772,
      "learning_rate": 4.1558861578266494e-07,
      "loss": 3.3872,
      "step": 257
    },
    {
      "epoch": 0.0334013010971939,
      "grad_norm": 1.4713171619014873,
      "learning_rate": 4.1720569210866756e-07,
      "loss": 3.6279,
      "step": 258
    },
    {
      "epoch": 0.03353076350454737,
      "grad_norm": 2.398750440671604,
      "learning_rate": 4.1882276843467017e-07,
      "loss": 4.675,
      "step": 259
    },
    {
      "epoch": 0.033660225911900835,
      "grad_norm": 2.325987124008959,
      "learning_rate": 4.2043984476067274e-07,
      "loss": 4.5254,
      "step": 260
    },
    {
      "epoch": 0.033660225911900835,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8865411877632141,
      "eval_runtime": 14.997,
      "eval_samples_per_second": 2.934,
      "eval_steps_per_second": 0.2,
      "step": 260
    },
    {
      "epoch": 0.033789688319254295,
      "grad_norm": 2.8100993157322263,
      "learning_rate": 4.220569210866753e-07,
      "loss": 4.8752,
      "step": 261
    },
    {
      "epoch": 0.03391915072660776,
      "grad_norm": 3.648392493445176,
      "learning_rate": 4.236739974126779e-07,
      "loss": 6.2065,
      "step": 262
    },
    {
      "epoch": 0.03404861313396123,
      "grad_norm": 2.6805319521377213,
      "learning_rate": 4.252910737386805e-07,
      "loss": 4.9004,
      "step": 263
    },
    {
      "epoch": 0.03417807554131469,
      "grad_norm": 2.3796110213603514,
      "learning_rate": 4.269081500646831e-07,
      "loss": 4.3872,
      "step": 264
    },
    {
      "epoch": 0.03430753794866816,
      "grad_norm": 2.40515345650876,
      "learning_rate": 4.2852522639068567e-07,
      "loss": 5.4297,
      "step": 265
    },
    {
      "epoch": 0.03430753794866816,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8852983117103577,
      "eval_runtime": 15.096,
      "eval_samples_per_second": 2.915,
      "eval_steps_per_second": 0.199,
      "step": 265
    },
    {
      "epoch": 0.03443700035602162,
      "grad_norm": 2.9258342379152245,
      "learning_rate": 4.3014230271668823e-07,
      "loss": 5.5503,
      "step": 266
    },
    {
      "epoch": 0.034566462763375085,
      "grad_norm": 3.61105782982967,
      "learning_rate": 4.3175937904269085e-07,
      "loss": 6.4419,
      "step": 267
    },
    {
      "epoch": 0.03469592517072855,
      "grad_norm": 2.78146236444209,
      "learning_rate": 4.3337645536869347e-07,
      "loss": 4.7144,
      "step": 268
    },
    {
      "epoch": 0.03482538757808201,
      "grad_norm": 2.9687116894361303,
      "learning_rate": 4.34993531694696e-07,
      "loss": 4.6682,
      "step": 269
    },
    {
      "epoch": 0.03495484998543548,
      "grad_norm": 2.4493788817536717,
      "learning_rate": 4.366106080206986e-07,
      "loss": 5.0649,
      "step": 270
    },
    {
      "epoch": 0.03495484998543548,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.884765625,
      "eval_runtime": 15.2162,
      "eval_samples_per_second": 2.892,
      "eval_steps_per_second": 0.197,
      "step": 270
    },
    {
      "epoch": 0.03508431239278895,
      "grad_norm": 3.6729121975196892,
      "learning_rate": 4.382276843467012e-07,
      "loss": 5.1895,
      "step": 271
    },
    {
      "epoch": 0.03521377480014241,
      "grad_norm": 2.2009377117963362,
      "learning_rate": 4.3984476067270383e-07,
      "loss": 4.1357,
      "step": 272
    },
    {
      "epoch": 0.035343237207495874,
      "grad_norm": 2.8980251483316244,
      "learning_rate": 4.4146183699870635e-07,
      "loss": 4.7271,
      "step": 273
    },
    {
      "epoch": 0.03547269961484934,
      "grad_norm": 2.649286538793547,
      "learning_rate": 4.4307891332470896e-07,
      "loss": 5.3291,
      "step": 274
    },
    {
      "epoch": 0.0356021620222028,
      "grad_norm": 2.407608486644509,
      "learning_rate": 4.446959896507116e-07,
      "loss": 4.6665,
      "step": 275
    },
    {
      "epoch": 0.0356021620222028,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8879616260528564,
      "eval_runtime": 15.3019,
      "eval_samples_per_second": 2.875,
      "eval_steps_per_second": 0.196,
      "step": 275
    },
    {
      "epoch": 0.03573162442955627,
      "grad_norm": 1.8412114012486007,
      "learning_rate": 4.463130659767141e-07,
      "loss": 3.5312,
      "step": 276
    },
    {
      "epoch": 0.03586108683690973,
      "grad_norm": 1.8829031229272333,
      "learning_rate": 4.479301423027167e-07,
      "loss": 4.104,
      "step": 277
    },
    {
      "epoch": 0.0359905492442632,
      "grad_norm": 3.318217480470199,
      "learning_rate": 4.4954721862871933e-07,
      "loss": 6.1147,
      "step": 278
    },
    {
      "epoch": 0.036120011651616664,
      "grad_norm": 3.349631812469815,
      "learning_rate": 4.511642949547219e-07,
      "loss": 5.5811,
      "step": 279
    },
    {
      "epoch": 0.036249474058970124,
      "grad_norm": 2.434116348743928,
      "learning_rate": 4.5278137128072446e-07,
      "loss": 4.1511,
      "step": 280
    },
    {
      "epoch": 0.036249474058970124,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.888671875,
      "eval_runtime": 13.3428,
      "eval_samples_per_second": 3.298,
      "eval_steps_per_second": 0.225,
      "step": 280
    },
    {
      "epoch": 0.03637893646632359,
      "grad_norm": 1.516175337782875,
      "learning_rate": 4.543984476067271e-07,
      "loss": 3.6128,
      "step": 281
    },
    {
      "epoch": 0.03650839887367706,
      "grad_norm": 2.622182149721442,
      "learning_rate": 4.5601552393272964e-07,
      "loss": 5.5483,
      "step": 282
    },
    {
      "epoch": 0.03663786128103052,
      "grad_norm": 2.0649261496070586,
      "learning_rate": 4.5763260025873226e-07,
      "loss": 4.5525,
      "step": 283
    },
    {
      "epoch": 0.036767323688383986,
      "grad_norm": 2.468057106361623,
      "learning_rate": 4.592496765847348e-07,
      "loss": 4.4458,
      "step": 284
    },
    {
      "epoch": 0.036896786095737454,
      "grad_norm": 2.4041084976851974,
      "learning_rate": 4.6086675291073744e-07,
      "loss": 4.571,
      "step": 285
    },
    {
      "epoch": 0.036896786095737454,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8911576867103577,
      "eval_runtime": 15.0977,
      "eval_samples_per_second": 2.914,
      "eval_steps_per_second": 0.199,
      "step": 285
    },
    {
      "epoch": 0.037026248503090914,
      "grad_norm": 2.421453764910423,
      "learning_rate": 4.6248382923674e-07,
      "loss": 4.3494,
      "step": 286
    },
    {
      "epoch": 0.03715571091044438,
      "grad_norm": 2.938594637147776,
      "learning_rate": 4.641009055627426e-07,
      "loss": 5.8699,
      "step": 287
    },
    {
      "epoch": 0.03728517331779784,
      "grad_norm": 3.3245404572673185,
      "learning_rate": 4.657179818887452e-07,
      "loss": 5.748,
      "step": 288
    },
    {
      "epoch": 0.03741463572515131,
      "grad_norm": 1.660595946323085,
      "learning_rate": 4.6733505821474775e-07,
      "loss": 3.3091,
      "step": 289
    },
    {
      "epoch": 0.037544098132504776,
      "grad_norm": 3.419572001471463,
      "learning_rate": 4.6895213454075037e-07,
      "loss": 6.2827,
      "step": 290
    },
    {
      "epoch": 0.037544098132504776,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8868963122367859,
      "eval_runtime": 14.2251,
      "eval_samples_per_second": 3.093,
      "eval_steps_per_second": 0.211,
      "step": 290
    },
    {
      "epoch": 0.037673560539858236,
      "grad_norm": 3.1209620215384737,
      "learning_rate": 4.70569210866753e-07,
      "loss": 5.6821,
      "step": 291
    },
    {
      "epoch": 0.037803022947211704,
      "grad_norm": 2.728390871733252,
      "learning_rate": 4.721862871927555e-07,
      "loss": 5.4678,
      "step": 292
    },
    {
      "epoch": 0.03793248535456517,
      "grad_norm": 1.1858990026523804,
      "learning_rate": 4.738033635187581e-07,
      "loss": 3.1748,
      "step": 293
    },
    {
      "epoch": 0.03806194776191863,
      "grad_norm": 1.0343287607700118,
      "learning_rate": 4.7542043984476074e-07,
      "loss": 2.7825,
      "step": 294
    },
    {
      "epoch": 0.0381914101692721,
      "grad_norm": 2.2100956196876402,
      "learning_rate": 4.770375161707633e-07,
      "loss": 4.1787,
      "step": 295
    },
    {
      "epoch": 0.0381914101692721,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8861860632896423,
      "eval_runtime": 14.5204,
      "eval_samples_per_second": 3.03,
      "eval_steps_per_second": 0.207,
      "step": 295
    },
    {
      "epoch": 0.038320872576625566,
      "grad_norm": 1.7395521864032368,
      "learning_rate": 4.786545924967659e-07,
      "loss": 3.9448,
      "step": 296
    },
    {
      "epoch": 0.038450334983979026,
      "grad_norm": 2.116504616762591,
      "learning_rate": 4.802716688227685e-07,
      "loss": 3.8701,
      "step": 297
    },
    {
      "epoch": 0.03857979739133249,
      "grad_norm": 1.980522519321597,
      "learning_rate": 4.81888745148771e-07,
      "loss": 4.3459,
      "step": 298
    },
    {
      "epoch": 0.038709259798685954,
      "grad_norm": 2.306067251236745,
      "learning_rate": 4.835058214747737e-07,
      "loss": 3.9229,
      "step": 299
    },
    {
      "epoch": 0.03883872220603942,
      "grad_norm": 2.52751944911705,
      "learning_rate": 4.851228978007762e-07,
      "loss": 5.0142,
      "step": 300
    },
    {
      "epoch": 0.03883872220603942,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8876065611839294,
      "eval_runtime": 14.5251,
      "eval_samples_per_second": 3.029,
      "eval_steps_per_second": 0.207,
      "step": 300
    },
    {
      "epoch": 0.03896818461339289,
      "grad_norm": 1.1084112549934337,
      "learning_rate": 4.867399741267789e-07,
      "loss": 3.595,
      "step": 301
    },
    {
      "epoch": 0.03909764702074635,
      "grad_norm": 2.555051046122174,
      "learning_rate": 4.883570504527814e-07,
      "loss": 5.9507,
      "step": 302
    },
    {
      "epoch": 0.039227109428099816,
      "grad_norm": 2.3812197403864244,
      "learning_rate": 4.89974126778784e-07,
      "loss": 4.7889,
      "step": 303
    },
    {
      "epoch": 0.03935657183545328,
      "grad_norm": 2.3063834650836927,
      "learning_rate": 4.915912031047866e-07,
      "loss": 4.5542,
      "step": 304
    },
    {
      "epoch": 0.03948603424280674,
      "grad_norm": 1.5953157628698709,
      "learning_rate": 4.932082794307892e-07,
      "loss": 3.4177,
      "step": 305
    },
    {
      "epoch": 0.03948603424280674,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8868963122367859,
      "eval_runtime": 15.031,
      "eval_samples_per_second": 2.927,
      "eval_steps_per_second": 0.2,
      "step": 305
    },
    {
      "epoch": 0.03961549665016021,
      "grad_norm": 2.52890823339106,
      "learning_rate": 4.948253557567917e-07,
      "loss": 4.3569,
      "step": 306
    },
    {
      "epoch": 0.03974495905751368,
      "grad_norm": 0.7325735914213573,
      "learning_rate": 4.964424320827944e-07,
      "loss": 2.3125,
      "step": 307
    },
    {
      "epoch": 0.03987442146486714,
      "grad_norm": 2.310293392821235,
      "learning_rate": 4.980595084087969e-07,
      "loss": 4.1499,
      "step": 308
    },
    {
      "epoch": 0.040003883872220605,
      "grad_norm": 1.402104868982115,
      "learning_rate": 4.996765847347995e-07,
      "loss": 3.123,
      "step": 309
    },
    {
      "epoch": 0.040133346279574066,
      "grad_norm": 1.894080574611163,
      "learning_rate": 5.012936610608021e-07,
      "loss": 4.0978,
      "step": 310
    },
    {
      "epoch": 0.040133346279574066,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8870738744735718,
      "eval_runtime": 14.8262,
      "eval_samples_per_second": 2.968,
      "eval_steps_per_second": 0.202,
      "step": 310
    },
    {
      "epoch": 0.04026280868692753,
      "grad_norm": 1.357102970182327,
      "learning_rate": 5.029107373868047e-07,
      "loss": 3.1089,
      "step": 311
    },
    {
      "epoch": 0.040392271094281,
      "grad_norm": 3.497980890858202,
      "learning_rate": 5.045278137128073e-07,
      "loss": 5.3418,
      "step": 312
    },
    {
      "epoch": 0.04052173350163446,
      "grad_norm": 1.0639089337277525,
      "learning_rate": 5.061448900388099e-07,
      "loss": 2.8896,
      "step": 313
    },
    {
      "epoch": 0.04065119590898793,
      "grad_norm": 3.300938656190228,
      "learning_rate": 5.077619663648125e-07,
      "loss": 5.9189,
      "step": 314
    },
    {
      "epoch": 0.040780658316341395,
      "grad_norm": 1.2413862644441418,
      "learning_rate": 5.09379042690815e-07,
      "loss": 3.4585,
      "step": 315
    },
    {
      "epoch": 0.040780658316341395,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8891157507896423,
      "eval_runtime": 15.4693,
      "eval_samples_per_second": 2.844,
      "eval_steps_per_second": 0.194,
      "step": 315
    },
    {
      "epoch": 0.040910120723694855,
      "grad_norm": 4.332081774178866,
      "learning_rate": 5.109961190168177e-07,
      "loss": 7.1943,
      "step": 316
    },
    {
      "epoch": 0.04103958313104832,
      "grad_norm": 3.573961749380484,
      "learning_rate": 5.126131953428203e-07,
      "loss": 5.5,
      "step": 317
    },
    {
      "epoch": 0.04116904553840179,
      "grad_norm": 3.266568775666242,
      "learning_rate": 5.142302716688228e-07,
      "loss": 5.0718,
      "step": 318
    },
    {
      "epoch": 0.04129850794575525,
      "grad_norm": 2.150736768179415,
      "learning_rate": 5.158473479948254e-07,
      "loss": 3.7144,
      "step": 319
    },
    {
      "epoch": 0.04142797035310872,
      "grad_norm": 2.9888130898725063,
      "learning_rate": 5.174644243208281e-07,
      "loss": 4.749,
      "step": 320
    },
    {
      "epoch": 0.04142797035310872,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8865411877632141,
      "eval_runtime": 16.5528,
      "eval_samples_per_second": 2.658,
      "eval_steps_per_second": 0.181,
      "step": 320
    },
    {
      "epoch": 0.04155743276046218,
      "grad_norm": 3.4246676283314756,
      "learning_rate": 5.190815006468305e-07,
      "loss": 5.1213,
      "step": 321
    },
    {
      "epoch": 0.041686895167815645,
      "grad_norm": 2.2618818130722693,
      "learning_rate": 5.206985769728332e-07,
      "loss": 4.0854,
      "step": 322
    },
    {
      "epoch": 0.04181635757516911,
      "grad_norm": 4.625362737261295,
      "learning_rate": 5.223156532988358e-07,
      "loss": 7.9058,
      "step": 323
    },
    {
      "epoch": 0.04194581998252257,
      "grad_norm": 1.9023520509429879,
      "learning_rate": 5.239327296248383e-07,
      "loss": 3.9753,
      "step": 324
    },
    {
      "epoch": 0.04207528238987604,
      "grad_norm": 2.43325767545436,
      "learning_rate": 5.255498059508409e-07,
      "loss": 4.7739,
      "step": 325
    },
    {
      "epoch": 0.04207528238987604,
      "eval_PRM Accuracy": 0.4056603773584906,
      "eval_PRM F1": 0.4424778761061947,
      "eval_PRM F1 AUC": 0.5419067574646412,
      "eval_PRM F1 Neg": 0.36363636363636365,
      "eval_PRM NPV": 0.23684210526315788,
      "eval_PRM Precision": 0.8333333333333334,
      "eval_PRM Recall": 0.30120481927710846,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8883167505264282,
      "eval_runtime": 15.0397,
      "eval_samples_per_second": 2.926,
      "eval_steps_per_second": 0.199,
      "step": 325
    },
    {
      "epoch": 0.04220474479722951,
      "grad_norm": 2.207312548983323,
      "learning_rate": 5.271668822768436e-07,
      "loss": 4.5195,
      "step": 326
    },
    {
      "epoch": 0.04233420720458297,
      "grad_norm": 2.4928797281193606,
      "learning_rate": 5.28783958602846e-07,
      "loss": 4.6545,
      "step": 327
    },
    {
      "epoch": 0.042463669611936435,
      "grad_norm": 2.489344579365318,
      "learning_rate": 5.304010349288487e-07,
      "loss": 4.6489,
      "step": 328
    },
    {
      "epoch": 0.0425931320192899,
      "grad_norm": 4.57429433288173,
      "learning_rate": 5.320181112548512e-07,
      "loss": 7.3164,
      "step": 329
    },
    {
      "epoch": 0.04272259442664336,
      "grad_norm": 3.255168861943587,
      "learning_rate": 5.336351875808539e-07,
      "loss": 6.5586,
      "step": 330
    },
    {
      "epoch": 0.04272259442664336,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8872514367103577,
      "eval_runtime": 14.6803,
      "eval_samples_per_second": 2.997,
      "eval_steps_per_second": 0.204,
      "step": 330
    },
    {
      "epoch": 0.04285205683399683,
      "grad_norm": 3.9597125898952092,
      "learning_rate": 5.352522639068565e-07,
      "loss": 7.1958,
      "step": 331
    },
    {
      "epoch": 0.04298151924135029,
      "grad_norm": 2.5225364315179766,
      "learning_rate": 5.36869340232859e-07,
      "loss": 4.2236,
      "step": 332
    },
    {
      "epoch": 0.04311098164870376,
      "grad_norm": 2.459820107417454,
      "learning_rate": 5.384864165588616e-07,
      "loss": 4.5474,
      "step": 333
    },
    {
      "epoch": 0.043240444056057224,
      "grad_norm": 3.2767650864133366,
      "learning_rate": 5.401034928848642e-07,
      "loss": 5.583,
      "step": 334
    },
    {
      "epoch": 0.043369906463410685,
      "grad_norm": 2.7676992302367696,
      "learning_rate": 5.417205692108668e-07,
      "loss": 4.689,
      "step": 335
    },
    {
      "epoch": 0.043369906463410685,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.88671875,
      "eval_runtime": 14.3717,
      "eval_samples_per_second": 3.062,
      "eval_steps_per_second": 0.209,
      "step": 335
    },
    {
      "epoch": 0.04349936887076415,
      "grad_norm": 2.051096966942122,
      "learning_rate": 5.433376455368694e-07,
      "loss": 3.8599,
      "step": 336
    },
    {
      "epoch": 0.04362883127811762,
      "grad_norm": 3.844604070905217,
      "learning_rate": 5.44954721862872e-07,
      "loss": 6.0503,
      "step": 337
    },
    {
      "epoch": 0.04375829368547108,
      "grad_norm": 4.316281408572996,
      "learning_rate": 5.465717981888745e-07,
      "loss": 8.3369,
      "step": 338
    },
    {
      "epoch": 0.04388775609282455,
      "grad_norm": 3.821369370757319,
      "learning_rate": 5.481888745148772e-07,
      "loss": 5.4961,
      "step": 339
    },
    {
      "epoch": 0.044017218500178014,
      "grad_norm": 4.54905571010069,
      "learning_rate": 5.498059508408797e-07,
      "loss": 6.9441,
      "step": 340
    },
    {
      "epoch": 0.044017218500178014,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8871626257896423,
      "eval_runtime": 14.1195,
      "eval_samples_per_second": 3.116,
      "eval_steps_per_second": 0.212,
      "step": 340
    },
    {
      "epoch": 0.044146680907531474,
      "grad_norm": 4.952659576810068,
      "learning_rate": 5.514230271668823e-07,
      "loss": 8.1987,
      "step": 341
    },
    {
      "epoch": 0.04427614331488494,
      "grad_norm": 3.302548186599011,
      "learning_rate": 5.530401034928849e-07,
      "loss": 5.1035,
      "step": 342
    },
    {
      "epoch": 0.0444056057222384,
      "grad_norm": 4.000622890748476,
      "learning_rate": 5.546571798188876e-07,
      "loss": 6.5605,
      "step": 343
    },
    {
      "epoch": 0.04453506812959187,
      "grad_norm": 5.127080244551329,
      "learning_rate": 5.5627425614489e-07,
      "loss": 8.0862,
      "step": 344
    },
    {
      "epoch": 0.044664530536945336,
      "grad_norm": 1.7447122779201634,
      "learning_rate": 5.578913324708927e-07,
      "loss": 4.0793,
      "step": 345
    },
    {
      "epoch": 0.044664530536945336,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8879616260528564,
      "eval_runtime": 14.9009,
      "eval_samples_per_second": 2.953,
      "eval_steps_per_second": 0.201,
      "step": 345
    },
    {
      "epoch": 0.0447939929442988,
      "grad_norm": 4.038935042030998,
      "learning_rate": 5.595084087968953e-07,
      "loss": 6.3291,
      "step": 346
    },
    {
      "epoch": 0.044923455351652264,
      "grad_norm": 3.2818578613931506,
      "learning_rate": 5.611254851228978e-07,
      "loss": 5.6594,
      "step": 347
    },
    {
      "epoch": 0.04505291775900573,
      "grad_norm": 2.8882454187139217,
      "learning_rate": 5.627425614489004e-07,
      "loss": 4.877,
      "step": 348
    },
    {
      "epoch": 0.04518238016635919,
      "grad_norm": 2.0416649463880256,
      "learning_rate": 5.643596377749031e-07,
      "loss": 3.7373,
      "step": 349
    },
    {
      "epoch": 0.04531184257371266,
      "grad_norm": 1.6936810512434821,
      "learning_rate": 5.659767141009055e-07,
      "loss": 3.5723,
      "step": 350
    },
    {
      "epoch": 0.04531184257371266,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8869850635528564,
      "eval_runtime": 14.1331,
      "eval_samples_per_second": 3.113,
      "eval_steps_per_second": 0.212,
      "step": 350
    },
    {
      "epoch": 0.045441304981066126,
      "grad_norm": 2.0200771091713503,
      "learning_rate": 5.675937904269082e-07,
      "loss": 4.0376,
      "step": 351
    },
    {
      "epoch": 0.045570767388419586,
      "grad_norm": 2.1854125513454976,
      "learning_rate": 5.692108667529108e-07,
      "loss": 4.7383,
      "step": 352
    },
    {
      "epoch": 0.045700229795773054,
      "grad_norm": 3.1451764651800267,
      "learning_rate": 5.708279430789133e-07,
      "loss": 5.8213,
      "step": 353
    },
    {
      "epoch": 0.045829692203126514,
      "grad_norm": 2.1449333513952196,
      "learning_rate": 5.72445019404916e-07,
      "loss": 3.7407,
      "step": 354
    },
    {
      "epoch": 0.04595915461047998,
      "grad_norm": 3.9835954047834705,
      "learning_rate": 5.740620957309186e-07,
      "loss": 6.1538,
      "step": 355
    },
    {
      "epoch": 0.04595915461047998,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8890269994735718,
      "eval_runtime": 14.3354,
      "eval_samples_per_second": 3.069,
      "eval_steps_per_second": 0.209,
      "step": 355
    },
    {
      "epoch": 0.04608861701783345,
      "grad_norm": 2.8708628349479204,
      "learning_rate": 5.756791720569211e-07,
      "loss": 5.3345,
      "step": 356
    },
    {
      "epoch": 0.04621807942518691,
      "grad_norm": 4.249639661394875,
      "learning_rate": 5.772962483829237e-07,
      "loss": 6.1455,
      "step": 357
    },
    {
      "epoch": 0.046347541832540376,
      "grad_norm": 3.325582864238537,
      "learning_rate": 5.789133247089264e-07,
      "loss": 4.8584,
      "step": 358
    },
    {
      "epoch": 0.04647700423989384,
      "grad_norm": 4.466988684148305,
      "learning_rate": 5.805304010349289e-07,
      "loss": 7.502,
      "step": 359
    },
    {
      "epoch": 0.046606466647247304,
      "grad_norm": 2.4849567677341424,
      "learning_rate": 5.821474773609315e-07,
      "loss": 5.0854,
      "step": 360
    },
    {
      "epoch": 0.046606466647247304,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8874289989471436,
      "eval_runtime": 14.55,
      "eval_samples_per_second": 3.024,
      "eval_steps_per_second": 0.206,
      "step": 360
    },
    {
      "epoch": 0.04673592905460077,
      "grad_norm": 3.7224181988715532,
      "learning_rate": 5.837645536869341e-07,
      "loss": 6.3174,
      "step": 361
    },
    {
      "epoch": 0.04686539146195424,
      "grad_norm": 3.234172791958926,
      "learning_rate": 5.853816300129367e-07,
      "loss": 6.0649,
      "step": 362
    },
    {
      "epoch": 0.0469948538693077,
      "grad_norm": 2.588006383504166,
      "learning_rate": 5.869987063389392e-07,
      "loss": 4.3237,
      "step": 363
    },
    {
      "epoch": 0.047124316276661166,
      "grad_norm": 4.537494485701655,
      "learning_rate": 5.886157826649419e-07,
      "loss": 6.4592,
      "step": 364
    },
    {
      "epoch": 0.047253778684014626,
      "grad_norm": 1.8112966664258494,
      "learning_rate": 5.902328589909444e-07,
      "loss": 3.6387,
      "step": 365
    },
    {
      "epoch": 0.047253778684014626,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8860085010528564,
      "eval_runtime": 15.407,
      "eval_samples_per_second": 2.856,
      "eval_steps_per_second": 0.195,
      "step": 365
    },
    {
      "epoch": 0.04738324109136809,
      "grad_norm": 3.5146563433221965,
      "learning_rate": 5.91849935316947e-07,
      "loss": 4.7334,
      "step": 366
    },
    {
      "epoch": 0.04751270349872156,
      "grad_norm": 1.4379114753778197,
      "learning_rate": 5.934670116429496e-07,
      "loss": 3.2285,
      "step": 367
    },
    {
      "epoch": 0.04764216590607502,
      "grad_norm": 1.908983344446691,
      "learning_rate": 5.950840879689522e-07,
      "loss": 4.0811,
      "step": 368
    },
    {
      "epoch": 0.04777162831342849,
      "grad_norm": 2.169763911252574,
      "learning_rate": 5.967011642949547e-07,
      "loss": 4.6023,
      "step": 369
    },
    {
      "epoch": 0.047901090720781955,
      "grad_norm": 2.1575595312092926,
      "learning_rate": 5.983182406209574e-07,
      "loss": 3.9585,
      "step": 370
    },
    {
      "epoch": 0.047901090720781955,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8864524364471436,
      "eval_runtime": 15.3038,
      "eval_samples_per_second": 2.875,
      "eval_steps_per_second": 0.196,
      "step": 370
    },
    {
      "epoch": 0.048030553128135416,
      "grad_norm": 5.318161033192671,
      "learning_rate": 5.999353169469599e-07,
      "loss": 7.8088,
      "step": 371
    },
    {
      "epoch": 0.04816001553548888,
      "grad_norm": 1.9980890310244137,
      "learning_rate": 6.015523932729626e-07,
      "loss": 3.5575,
      "step": 372
    },
    {
      "epoch": 0.04828947794284235,
      "grad_norm": 3.949506127219697,
      "learning_rate": 6.031694695989651e-07,
      "loss": 6.4028,
      "step": 373
    },
    {
      "epoch": 0.04841894035019581,
      "grad_norm": 1.7781458866902489,
      "learning_rate": 6.047865459249677e-07,
      "loss": 3.4214,
      "step": 374
    },
    {
      "epoch": 0.04854840275754928,
      "grad_norm": 2.4135744708988938,
      "learning_rate": 6.064036222509703e-07,
      "loss": 4.8521,
      "step": 375
    },
    {
      "epoch": 0.04854840275754928,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8888494372367859,
      "eval_runtime": 16.1404,
      "eval_samples_per_second": 2.726,
      "eval_steps_per_second": 0.186,
      "step": 375
    },
    {
      "epoch": 0.04867786516490274,
      "grad_norm": 2.427114114429261,
      "learning_rate": 6.080206985769729e-07,
      "loss": 4.2141,
      "step": 376
    },
    {
      "epoch": 0.048807327572256205,
      "grad_norm": 2.206430166153848,
      "learning_rate": 6.096377749029755e-07,
      "loss": 3.7795,
      "step": 377
    },
    {
      "epoch": 0.04893678997960967,
      "grad_norm": 3.3553556761191574,
      "learning_rate": 6.112548512289781e-07,
      "loss": 4.9688,
      "step": 378
    },
    {
      "epoch": 0.04906625238696313,
      "grad_norm": 0.9161142075888727,
      "learning_rate": 6.128719275549807e-07,
      "loss": 3.0227,
      "step": 379
    },
    {
      "epoch": 0.0491957147943166,
      "grad_norm": 3.471482846055819,
      "learning_rate": 6.144890038809832e-07,
      "loss": 5.7178,
      "step": 380
    },
    {
      "epoch": 0.0491957147943166,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8895596861839294,
      "eval_runtime": 14.2623,
      "eval_samples_per_second": 3.085,
      "eval_steps_per_second": 0.21,
      "step": 380
    },
    {
      "epoch": 0.04932517720167007,
      "grad_norm": 1.9819082834705617,
      "learning_rate": 6.161060802069859e-07,
      "loss": 3.8608,
      "step": 381
    },
    {
      "epoch": 0.04945463960902353,
      "grad_norm": 2.903210725442833,
      "learning_rate": 6.177231565329884e-07,
      "loss": 5.5146,
      "step": 382
    },
    {
      "epoch": 0.049584102016376995,
      "grad_norm": 2.6708054607635168,
      "learning_rate": 6.19340232858991e-07,
      "loss": 4.4766,
      "step": 383
    },
    {
      "epoch": 0.04971356442373046,
      "grad_norm": 3.1183717764760837,
      "learning_rate": 6.209573091849936e-07,
      "loss": 5.4492,
      "step": 384
    },
    {
      "epoch": 0.04984302683108392,
      "grad_norm": 3.7400801881411736,
      "learning_rate": 6.225743855109962e-07,
      "loss": 6.1411,
      "step": 385
    },
    {
      "epoch": 0.04984302683108392,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8873401880264282,
      "eval_runtime": 16.4864,
      "eval_samples_per_second": 2.669,
      "eval_steps_per_second": 0.182,
      "step": 385
    },
    {
      "epoch": 0.04997248923843739,
      "grad_norm": 2.571588858217802,
      "learning_rate": 6.241914618369987e-07,
      "loss": 4.6377,
      "step": 386
    },
    {
      "epoch": 0.05010195164579085,
      "grad_norm": 1.905362147390013,
      "learning_rate": 6.258085381630014e-07,
      "loss": 3.9556,
      "step": 387
    },
    {
      "epoch": 0.05023141405314432,
      "grad_norm": 2.703486235188581,
      "learning_rate": 6.27425614489004e-07,
      "loss": 4.6113,
      "step": 388
    },
    {
      "epoch": 0.050360876460497785,
      "grad_norm": 2.0525631550788725,
      "learning_rate": 6.290426908150066e-07,
      "loss": 4.4429,
      "step": 389
    },
    {
      "epoch": 0.050490338867851245,
      "grad_norm": 3.215100360275304,
      "learning_rate": 6.306597671410091e-07,
      "loss": 5.5059,
      "step": 390
    },
    {
      "epoch": 0.050490338867851245,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8899147510528564,
      "eval_runtime": 18.5744,
      "eval_samples_per_second": 2.369,
      "eval_steps_per_second": 0.162,
      "step": 390
    },
    {
      "epoch": 0.05061980127520471,
      "grad_norm": 2.7883839598086984,
      "learning_rate": 6.322768434670116e-07,
      "loss": 5.1589,
      "step": 391
    },
    {
      "epoch": 0.05074926368255818,
      "grad_norm": 2.39681815806041,
      "learning_rate": 6.338939197930142e-07,
      "loss": 3.928,
      "step": 392
    },
    {
      "epoch": 0.05087872608991164,
      "grad_norm": 2.7225392239613098,
      "learning_rate": 6.355109961190169e-07,
      "loss": 4.7026,
      "step": 393
    },
    {
      "epoch": 0.05100818849726511,
      "grad_norm": 4.518284177510381,
      "learning_rate": 6.371280724450194e-07,
      "loss": 6.3408,
      "step": 394
    },
    {
      "epoch": 0.051137650904618574,
      "grad_norm": 2.5612781804229034,
      "learning_rate": 6.387451487710221e-07,
      "loss": 4.2725,
      "step": 395
    },
    {
      "epoch": 0.051137650904618574,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8902698755264282,
      "eval_runtime": 14.162,
      "eval_samples_per_second": 3.107,
      "eval_steps_per_second": 0.212,
      "step": 395
    },
    {
      "epoch": 0.051267113311972035,
      "grad_norm": 2.280496654996542,
      "learning_rate": 6.403622250970247e-07,
      "loss": 4.4756,
      "step": 396
    },
    {
      "epoch": 0.0513965757193255,
      "grad_norm": 3.021073527285268,
      "learning_rate": 6.419793014230271e-07,
      "loss": 5.7964,
      "step": 397
    },
    {
      "epoch": 0.05152603812667896,
      "grad_norm": 2.6373988360828675,
      "learning_rate": 6.435963777490298e-07,
      "loss": 5.1425,
      "step": 398
    },
    {
      "epoch": 0.05165550053403243,
      "grad_norm": 1.84905961264132,
      "learning_rate": 6.452134540750324e-07,
      "loss": 3.5947,
      "step": 399
    },
    {
      "epoch": 0.0517849629413859,
      "grad_norm": 2.277941994831721,
      "learning_rate": 6.46830530401035e-07,
      "loss": 4.4268,
      "step": 400
    },
    {
      "epoch": 0.0517849629413859,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8890269994735718,
      "eval_runtime": 14.6147,
      "eval_samples_per_second": 3.011,
      "eval_steps_per_second": 0.205,
      "step": 400
    },
    {
      "epoch": 0.05191442534873936,
      "grad_norm": 3.3556177810260066,
      "learning_rate": 6.484476067270376e-07,
      "loss": 5.3779,
      "step": 401
    },
    {
      "epoch": 0.052043887756092824,
      "grad_norm": 2.67169301328284,
      "learning_rate": 6.500646830530402e-07,
      "loss": 4.3022,
      "step": 402
    },
    {
      "epoch": 0.05217335016344629,
      "grad_norm": 2.7219829214054916,
      "learning_rate": 6.516817593790426e-07,
      "loss": 4.4153,
      "step": 403
    },
    {
      "epoch": 0.05230281257079975,
      "grad_norm": 2.5490216138266737,
      "learning_rate": 6.532988357050453e-07,
      "loss": 3.7578,
      "step": 404
    },
    {
      "epoch": 0.05243227497815322,
      "grad_norm": 1.788090552474802,
      "learning_rate": 6.549159120310479e-07,
      "loss": 3.8025,
      "step": 405
    },
    {
      "epoch": 0.05243227497815322,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8891157507896423,
      "eval_runtime": 14.8253,
      "eval_samples_per_second": 2.968,
      "eval_steps_per_second": 0.202,
      "step": 405
    },
    {
      "epoch": 0.052561737385506686,
      "grad_norm": 3.703355461870673,
      "learning_rate": 6.565329883570505e-07,
      "loss": 6.5244,
      "step": 406
    },
    {
      "epoch": 0.05269119979286015,
      "grad_norm": 3.380054763866634,
      "learning_rate": 6.581500646830531e-07,
      "loss": 5.3091,
      "step": 407
    },
    {
      "epoch": 0.052820662200213614,
      "grad_norm": 3.84788336684906,
      "learning_rate": 6.597671410090558e-07,
      "loss": 6.6704,
      "step": 408
    },
    {
      "epoch": 0.05295012460756708,
      "grad_norm": 2.375997572908149,
      "learning_rate": 6.613842173350583e-07,
      "loss": 3.1401,
      "step": 409
    },
    {
      "epoch": 0.05307958701492054,
      "grad_norm": 2.2528166969422854,
      "learning_rate": 6.630012936610608e-07,
      "loss": 3.981,
      "step": 410
    },
    {
      "epoch": 0.05307958701492054,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8892045617103577,
      "eval_runtime": 14.5997,
      "eval_samples_per_second": 3.014,
      "eval_steps_per_second": 0.205,
      "step": 410
    },
    {
      "epoch": 0.05320904942227401,
      "grad_norm": 2.7506030359883282,
      "learning_rate": 6.646183699870634e-07,
      "loss": 3.6577,
      "step": 411
    },
    {
      "epoch": 0.05333851182962747,
      "grad_norm": 2.7556404206790663,
      "learning_rate": 6.66235446313066e-07,
      "loss": 4.3684,
      "step": 412
    },
    {
      "epoch": 0.053467974236980936,
      "grad_norm": 2.6646633731112406,
      "learning_rate": 6.678525226390686e-07,
      "loss": 3.9849,
      "step": 413
    },
    {
      "epoch": 0.053597436644334404,
      "grad_norm": 3.212966981020102,
      "learning_rate": 6.694695989650713e-07,
      "loss": 6.1631,
      "step": 414
    },
    {
      "epoch": 0.053726899051687864,
      "grad_norm": 1.6691294126989227,
      "learning_rate": 6.710866752910738e-07,
      "loss": 4.3237,
      "step": 415
    },
    {
      "epoch": 0.053726899051687864,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8884943127632141,
      "eval_runtime": 14.796,
      "eval_samples_per_second": 2.974,
      "eval_steps_per_second": 0.203,
      "step": 415
    },
    {
      "epoch": 0.05385636145904133,
      "grad_norm": 2.3037675071234434,
      "learning_rate": 6.727037516170763e-07,
      "loss": 5.248,
      "step": 416
    },
    {
      "epoch": 0.0539858238663948,
      "grad_norm": 2.7338076925970465,
      "learning_rate": 6.74320827943079e-07,
      "loss": 5.0679,
      "step": 417
    },
    {
      "epoch": 0.05411528627374826,
      "grad_norm": 3.3411195413599852,
      "learning_rate": 6.759379042690815e-07,
      "loss": 5.0293,
      "step": 418
    },
    {
      "epoch": 0.054244748681101726,
      "grad_norm": 2.8988184361633214,
      "learning_rate": 6.775549805950841e-07,
      "loss": 4.6348,
      "step": 419
    },
    {
      "epoch": 0.05437421108845519,
      "grad_norm": 3.258555741636187,
      "learning_rate": 6.791720569210868e-07,
      "loss": 5.2798,
      "step": 420
    },
    {
      "epoch": 0.05437421108845519,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8904474377632141,
      "eval_runtime": 13.7417,
      "eval_samples_per_second": 3.202,
      "eval_steps_per_second": 0.218,
      "step": 420
    },
    {
      "epoch": 0.054503673495808654,
      "grad_norm": 2.845290843273376,
      "learning_rate": 6.807891332470893e-07,
      "loss": 5.0366,
      "step": 421
    },
    {
      "epoch": 0.05463313590316212,
      "grad_norm": 3.229948321320305,
      "learning_rate": 6.824062095730918e-07,
      "loss": 5.7407,
      "step": 422
    },
    {
      "epoch": 0.05476259831051558,
      "grad_norm": 3.986785020505741,
      "learning_rate": 6.840232858990945e-07,
      "loss": 6.0254,
      "step": 423
    },
    {
      "epoch": 0.05489206071786905,
      "grad_norm": 2.4422246084783574,
      "learning_rate": 6.85640362225097e-07,
      "loss": 4.2644,
      "step": 424
    },
    {
      "epoch": 0.055021523125222516,
      "grad_norm": 1.2793166510673597,
      "learning_rate": 6.872574385510997e-07,
      "loss": 2.6863,
      "step": 425
    },
    {
      "epoch": 0.055021523125222516,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8899147510528564,
      "eval_runtime": 15.7681,
      "eval_samples_per_second": 2.79,
      "eval_steps_per_second": 0.19,
      "step": 425
    },
    {
      "epoch": 0.055150985532575976,
      "grad_norm": 2.97927622169673,
      "learning_rate": 6.888745148771023e-07,
      "loss": 5.5723,
      "step": 426
    },
    {
      "epoch": 0.05528044793992944,
      "grad_norm": 3.6203735487325255,
      "learning_rate": 6.904915912031049e-07,
      "loss": 5.0869,
      "step": 427
    },
    {
      "epoch": 0.05540991034728291,
      "grad_norm": 1.9163894983398424,
      "learning_rate": 6.921086675291075e-07,
      "loss": 4.2043,
      "step": 428
    },
    {
      "epoch": 0.05553937275463637,
      "grad_norm": 3.4733572365302283,
      "learning_rate": 6.9372574385511e-07,
      "loss": 5.252,
      "step": 429
    },
    {
      "epoch": 0.05566883516198984,
      "grad_norm": 3.566197558885763,
      "learning_rate": 6.953428201811125e-07,
      "loss": 5.3286,
      "step": 430
    },
    {
      "epoch": 0.05566883516198984,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8918678760528564,
      "eval_runtime": 13.5644,
      "eval_samples_per_second": 3.244,
      "eval_steps_per_second": 0.221,
      "step": 430
    },
    {
      "epoch": 0.055798297569343305,
      "grad_norm": 2.035928802771165,
      "learning_rate": 6.969598965071152e-07,
      "loss": 4.3062,
      "step": 431
    },
    {
      "epoch": 0.055927759976696766,
      "grad_norm": 3.3999902557564505,
      "learning_rate": 6.985769728331178e-07,
      "loss": 5.4189,
      "step": 432
    },
    {
      "epoch": 0.05605722238405023,
      "grad_norm": 2.4809527375688374,
      "learning_rate": 7.001940491591204e-07,
      "loss": 4.335,
      "step": 433
    },
    {
      "epoch": 0.05618668479140369,
      "grad_norm": 1.5447436872047051,
      "learning_rate": 7.01811125485123e-07,
      "loss": 3.5166,
      "step": 434
    },
    {
      "epoch": 0.05631614719875716,
      "grad_norm": 2.1227145286062705,
      "learning_rate": 7.034282018111255e-07,
      "loss": 3.7466,
      "step": 435
    },
    {
      "epoch": 0.05631614719875716,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8898259997367859,
      "eval_runtime": 13.4876,
      "eval_samples_per_second": 3.262,
      "eval_steps_per_second": 0.222,
      "step": 435
    },
    {
      "epoch": 0.05644560960611063,
      "grad_norm": 2.8037983416166057,
      "learning_rate": 7.050452781371281e-07,
      "loss": 4.9492,
      "step": 436
    },
    {
      "epoch": 0.05657507201346409,
      "grad_norm": 2.618688269998642,
      "learning_rate": 7.066623544631307e-07,
      "loss": 4.6328,
      "step": 437
    },
    {
      "epoch": 0.056704534420817555,
      "grad_norm": 2.36505143201812,
      "learning_rate": 7.082794307891333e-07,
      "loss": 4.5103,
      "step": 438
    },
    {
      "epoch": 0.05683399682817102,
      "grad_norm": 2.525390782647627,
      "learning_rate": 7.098965071151359e-07,
      "loss": 5.0796,
      "step": 439
    },
    {
      "epoch": 0.05696345923552448,
      "grad_norm": 2.371315772725317,
      "learning_rate": 7.115135834411385e-07,
      "loss": 3.8354,
      "step": 440
    },
    {
      "epoch": 0.05696345923552448,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8881391882896423,
      "eval_runtime": 14.4353,
      "eval_samples_per_second": 3.048,
      "eval_steps_per_second": 0.208,
      "step": 440
    },
    {
      "epoch": 0.05709292164287795,
      "grad_norm": 2.389207964725016,
      "learning_rate": 7.131306597671412e-07,
      "loss": 4.4639,
      "step": 441
    },
    {
      "epoch": 0.05722238405023142,
      "grad_norm": 2.431164961294971,
      "learning_rate": 7.147477360931436e-07,
      "loss": 4.1638,
      "step": 442
    },
    {
      "epoch": 0.05735184645758488,
      "grad_norm": 1.8199313844536043,
      "learning_rate": 7.163648124191462e-07,
      "loss": 3.2183,
      "step": 443
    },
    {
      "epoch": 0.057481308864938345,
      "grad_norm": 2.624071289645593,
      "learning_rate": 7.179818887451489e-07,
      "loss": 5.0615,
      "step": 444
    },
    {
      "epoch": 0.057610771272291805,
      "grad_norm": 2.921808231280524,
      "learning_rate": 7.195989650711514e-07,
      "loss": 5.3948,
      "step": 445
    },
    {
      "epoch": 0.057610771272291805,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8890269994735718,
      "eval_runtime": 14.2128,
      "eval_samples_per_second": 3.096,
      "eval_steps_per_second": 0.211,
      "step": 445
    },
    {
      "epoch": 0.05774023367964527,
      "grad_norm": 2.4882204607921357,
      "learning_rate": 7.212160413971541e-07,
      "loss": 5.0454,
      "step": 446
    },
    {
      "epoch": 0.05786969608699874,
      "grad_norm": 5.292947670925015,
      "learning_rate": 7.228331177231567e-07,
      "loss": 7.4385,
      "step": 447
    },
    {
      "epoch": 0.0579991584943522,
      "grad_norm": 3.5914756621173463,
      "learning_rate": 7.244501940491591e-07,
      "loss": 5.6191,
      "step": 448
    },
    {
      "epoch": 0.05812862090170567,
      "grad_norm": 2.447688584948203,
      "learning_rate": 7.260672703751617e-07,
      "loss": 4.2495,
      "step": 449
    },
    {
      "epoch": 0.058258083309059135,
      "grad_norm": 2.1663943903015634,
      "learning_rate": 7.276843467011643e-07,
      "loss": 3.2456,
      "step": 450
    },
    {
      "epoch": 0.058258083309059135,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8911576867103577,
      "eval_runtime": 14.5949,
      "eval_samples_per_second": 3.015,
      "eval_steps_per_second": 0.206,
      "step": 450
    },
    {
      "epoch": 0.058387545716412595,
      "grad_norm": 2.8159884845122183,
      "learning_rate": 7.293014230271669e-07,
      "loss": 4.5798,
      "step": 451
    },
    {
      "epoch": 0.05851700812376606,
      "grad_norm": 3.1860772839931486,
      "learning_rate": 7.309184993531696e-07,
      "loss": 5.5576,
      "step": 452
    },
    {
      "epoch": 0.05864647053111953,
      "grad_norm": 3.1715276442448603,
      "learning_rate": 7.325355756791721e-07,
      "loss": 5.3296,
      "step": 453
    },
    {
      "epoch": 0.05877593293847299,
      "grad_norm": 1.8783448074542377,
      "learning_rate": 7.341526520051748e-07,
      "loss": 4.0239,
      "step": 454
    },
    {
      "epoch": 0.05890539534582646,
      "grad_norm": 1.6432187545216004,
      "learning_rate": 7.357697283311773e-07,
      "loss": 3.6694,
      "step": 455
    },
    {
      "epoch": 0.05890539534582646,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8937322497367859,
      "eval_runtime": 15.0472,
      "eval_samples_per_second": 2.924,
      "eval_steps_per_second": 0.199,
      "step": 455
    },
    {
      "epoch": 0.05903485775317992,
      "grad_norm": 3.9878701303834188,
      "learning_rate": 7.373868046571798e-07,
      "loss": 5.7021,
      "step": 456
    },
    {
      "epoch": 0.059164320160533385,
      "grad_norm": 2.795985084096421,
      "learning_rate": 7.390038809831824e-07,
      "loss": 4.4189,
      "step": 457
    },
    {
      "epoch": 0.05929378256788685,
      "grad_norm": 3.6794018234948163,
      "learning_rate": 7.406209573091851e-07,
      "loss": 4.9004,
      "step": 458
    },
    {
      "epoch": 0.05942324497524031,
      "grad_norm": 3.1117472837893887,
      "learning_rate": 7.422380336351876e-07,
      "loss": 5.7109,
      "step": 459
    },
    {
      "epoch": 0.05955270738259378,
      "grad_norm": 3.6666275759650735,
      "learning_rate": 7.438551099611903e-07,
      "loss": 5.7236,
      "step": 460
    },
    {
      "epoch": 0.05955270738259378,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8926668763160706,
      "eval_runtime": 14.9814,
      "eval_samples_per_second": 2.937,
      "eval_steps_per_second": 0.2,
      "step": 460
    },
    {
      "epoch": 0.05968216978994725,
      "grad_norm": 4.109836747246771,
      "learning_rate": 7.454721862871928e-07,
      "loss": 5.9663,
      "step": 461
    },
    {
      "epoch": 0.05981163219730071,
      "grad_norm": 2.9539081126626514,
      "learning_rate": 7.470892626131953e-07,
      "loss": 5.04,
      "step": 462
    },
    {
      "epoch": 0.059941094604654174,
      "grad_norm": 2.108616044894569,
      "learning_rate": 7.48706338939198e-07,
      "loss": 4.1074,
      "step": 463
    },
    {
      "epoch": 0.06007055701200764,
      "grad_norm": 2.3062645651100695,
      "learning_rate": 7.503234152652006e-07,
      "loss": 4.207,
      "step": 464
    },
    {
      "epoch": 0.0602000194193611,
      "grad_norm": 1.970504808382239,
      "learning_rate": 7.519404915912032e-07,
      "loss": 3.3633,
      "step": 465
    },
    {
      "epoch": 0.0602000194193611,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8894708752632141,
      "eval_runtime": 14.5718,
      "eval_samples_per_second": 3.02,
      "eval_steps_per_second": 0.206,
      "step": 465
    },
    {
      "epoch": 0.06032948182671457,
      "grad_norm": 3.7110334716004316,
      "learning_rate": 7.535575679172058e-07,
      "loss": 5.3782,
      "step": 466
    },
    {
      "epoch": 0.06045894423406803,
      "grad_norm": 3.2511972407968432,
      "learning_rate": 7.551746442432084e-07,
      "loss": 5.1436,
      "step": 467
    },
    {
      "epoch": 0.0605884066414215,
      "grad_norm": 4.045181351030494,
      "learning_rate": 7.567917205692108e-07,
      "loss": 5.3154,
      "step": 468
    },
    {
      "epoch": 0.060717869048774964,
      "grad_norm": 2.5219104109115245,
      "learning_rate": 7.584087968952135e-07,
      "loss": 4.4136,
      "step": 469
    },
    {
      "epoch": 0.060847331456128424,
      "grad_norm": 2.655383749644346,
      "learning_rate": 7.600258732212161e-07,
      "loss": 4.3418,
      "step": 470
    },
    {
      "epoch": 0.060847331456128424,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8932883739471436,
      "eval_runtime": 14.3407,
      "eval_samples_per_second": 3.068,
      "eval_steps_per_second": 0.209,
      "step": 470
    },
    {
      "epoch": 0.06097679386348189,
      "grad_norm": 3.162879797053235,
      "learning_rate": 7.616429495472187e-07,
      "loss": 5.5503,
      "step": 471
    },
    {
      "epoch": 0.06110625627083536,
      "grad_norm": 2.4947505686132256,
      "learning_rate": 7.632600258732213e-07,
      "loss": 4.1929,
      "step": 472
    },
    {
      "epoch": 0.06123571867818882,
      "grad_norm": 2.4022537273906552,
      "learning_rate": 7.64877102199224e-07,
      "loss": 4.6172,
      "step": 473
    },
    {
      "epoch": 0.061365181085542286,
      "grad_norm": 2.786311565932309,
      "learning_rate": 7.664941785252264e-07,
      "loss": 4.5874,
      "step": 474
    },
    {
      "epoch": 0.061494643492895754,
      "grad_norm": 2.7403569099140204,
      "learning_rate": 7.68111254851229e-07,
      "loss": 4.144,
      "step": 475
    },
    {
      "epoch": 0.061494643492895754,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8924005627632141,
      "eval_runtime": 15.4681,
      "eval_samples_per_second": 2.845,
      "eval_steps_per_second": 0.194,
      "step": 475
    },
    {
      "epoch": 0.061624105900249214,
      "grad_norm": 4.056895151035421,
      "learning_rate": 7.697283311772316e-07,
      "loss": 6.5996,
      "step": 476
    },
    {
      "epoch": 0.06175356830760268,
      "grad_norm": 3.019615547745032,
      "learning_rate": 7.713454075032342e-07,
      "loss": 4.4626,
      "step": 477
    },
    {
      "epoch": 0.06188303071495614,
      "grad_norm": 1.7076858671943458,
      "learning_rate": 7.729624838292368e-07,
      "loss": 3.6279,
      "step": 478
    },
    {
      "epoch": 0.06201249312230961,
      "grad_norm": 4.584219984959799,
      "learning_rate": 7.745795601552395e-07,
      "loss": 5.9634,
      "step": 479
    },
    {
      "epoch": 0.062141955529663076,
      "grad_norm": 2.588902587312457,
      "learning_rate": 7.761966364812419e-07,
      "loss": 4.2678,
      "step": 480
    },
    {
      "epoch": 0.062141955529663076,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8933771252632141,
      "eval_runtime": 15.4397,
      "eval_samples_per_second": 2.85,
      "eval_steps_per_second": 0.194,
      "step": 480
    },
    {
      "epoch": 0.062271417937016536,
      "grad_norm": 3.2694234740804493,
      "learning_rate": 7.778137128072445e-07,
      "loss": 4.5474,
      "step": 481
    },
    {
      "epoch": 0.062400880344370004,
      "grad_norm": 2.744101141601191,
      "learning_rate": 7.794307891332472e-07,
      "loss": 4.4507,
      "step": 482
    },
    {
      "epoch": 0.06253034275172346,
      "grad_norm": 2.6973642665625928,
      "learning_rate": 7.810478654592497e-07,
      "loss": 4.2295,
      "step": 483
    },
    {
      "epoch": 0.06265980515907693,
      "grad_norm": 2.8343522695658883,
      "learning_rate": 7.826649417852523e-07,
      "loss": 4.8569,
      "step": 484
    },
    {
      "epoch": 0.0627892675664304,
      "grad_norm": 1.5759981758527295,
      "learning_rate": 7.84282018111255e-07,
      "loss": 3.3433,
      "step": 485
    },
    {
      "epoch": 0.0627892675664304,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8951526880264282,
      "eval_runtime": 14.3716,
      "eval_samples_per_second": 3.062,
      "eval_steps_per_second": 0.209,
      "step": 485
    },
    {
      "epoch": 0.06291872997378387,
      "grad_norm": 3.065308612184375,
      "learning_rate": 7.858990944372575e-07,
      "loss": 4.9492,
      "step": 486
    },
    {
      "epoch": 0.06304819238113733,
      "grad_norm": 3.6197399144473685,
      "learning_rate": 7.8751617076326e-07,
      "loss": 5.8706,
      "step": 487
    },
    {
      "epoch": 0.06317765478849079,
      "grad_norm": 4.4138003218402035,
      "learning_rate": 7.891332470892627e-07,
      "loss": 5.9075,
      "step": 488
    },
    {
      "epoch": 0.06330711719584425,
      "grad_norm": 3.647047343731257,
      "learning_rate": 7.907503234152652e-07,
      "loss": 5.3857,
      "step": 489
    },
    {
      "epoch": 0.06343657960319772,
      "grad_norm": 2.087150238196096,
      "learning_rate": 7.923673997412679e-07,
      "loss": 3.5122,
      "step": 490
    },
    {
      "epoch": 0.06343657960319772,
      "eval_PRM Accuracy": 0.39622641509433965,
      "eval_PRM F1": 0.42857142857142855,
      "eval_PRM F1 AUC": 0.5358826610790991,
      "eval_PRM F1 Neg": 0.36,
      "eval_PRM NPV": 0.23376623376623376,
      "eval_PRM Precision": 0.8275862068965517,
      "eval_PRM Recall": 0.2891566265060241,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8950639367103577,
      "eval_runtime": 14.6726,
      "eval_samples_per_second": 2.999,
      "eval_steps_per_second": 0.204,
      "step": 490
    },
    {
      "epoch": 0.06356604201055119,
      "grad_norm": 2.915236854067916,
      "learning_rate": 7.939844760672705e-07,
      "loss": 4.5864,
      "step": 491
    },
    {
      "epoch": 0.06369550441790466,
      "grad_norm": 3.8377116990770133,
      "learning_rate": 7.956015523932731e-07,
      "loss": 5.5801,
      "step": 492
    },
    {
      "epoch": 0.06382496682525812,
      "grad_norm": 1.4409468802970715,
      "learning_rate": 7.972186287192756e-07,
      "loss": 3.3596,
      "step": 493
    },
    {
      "epoch": 0.06395442923261158,
      "grad_norm": 2.1640560193412863,
      "learning_rate": 7.988357050452782e-07,
      "loss": 4.2856,
      "step": 494
    },
    {
      "epoch": 0.06408389163996504,
      "grad_norm": 3.1133391337473446,
      "learning_rate": 8.004527813712807e-07,
      "loss": 4.3745,
      "step": 495
    },
    {
      "epoch": 0.06408389163996504,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8969282507896423,
      "eval_runtime": 13.4116,
      "eval_samples_per_second": 3.281,
      "eval_steps_per_second": 0.224,
      "step": 495
    },
    {
      "epoch": 0.06421335404731851,
      "grad_norm": 3.1589057385124737,
      "learning_rate": 8.020698576972834e-07,
      "loss": 4.1719,
      "step": 496
    },
    {
      "epoch": 0.06434281645467198,
      "grad_norm": 2.5352609093998444,
      "learning_rate": 8.03686934023286e-07,
      "loss": 4.3894,
      "step": 497
    },
    {
      "epoch": 0.06447227886202545,
      "grad_norm": 2.52113941438638,
      "learning_rate": 8.053040103492886e-07,
      "loss": 4.124,
      "step": 498
    },
    {
      "epoch": 0.0646017412693789,
      "grad_norm": 2.444114794921976,
      "learning_rate": 8.069210866752912e-07,
      "loss": 4.2605,
      "step": 499
    },
    {
      "epoch": 0.06473120367673237,
      "grad_norm": 1.8003741061422533,
      "learning_rate": 8.085381630012937e-07,
      "loss": 3.2925,
      "step": 500
    },
    {
      "epoch": 0.06473120367673237,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8959516882896423,
      "eval_runtime": 15.069,
      "eval_samples_per_second": 2.92,
      "eval_steps_per_second": 0.199,
      "step": 500
    },
    {
      "epoch": 0.06486066608408583,
      "grad_norm": 3.0540591692146033,
      "learning_rate": 8.101552393272963e-07,
      "loss": 5.0771,
      "step": 501
    },
    {
      "epoch": 0.0649901284914393,
      "grad_norm": 4.033475081295743,
      "learning_rate": 8.117723156532989e-07,
      "loss": 7.1152,
      "step": 502
    },
    {
      "epoch": 0.06511959089879277,
      "grad_norm": 3.3108240258330173,
      "learning_rate": 8.133893919793015e-07,
      "loss": 4.3901,
      "step": 503
    },
    {
      "epoch": 0.06524905330614623,
      "grad_norm": 1.8525932833865588,
      "learning_rate": 8.150064683053041e-07,
      "loss": 4.0447,
      "step": 504
    },
    {
      "epoch": 0.06537851571349969,
      "grad_norm": 3.107982421650894,
      "learning_rate": 8.166235446313067e-07,
      "loss": 5.0181,
      "step": 505
    },
    {
      "epoch": 0.06537851571349969,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8978160619735718,
      "eval_runtime": 14.3945,
      "eval_samples_per_second": 3.057,
      "eval_steps_per_second": 0.208,
      "step": 505
    },
    {
      "epoch": 0.06550797812085316,
      "grad_norm": 2.937715221227242,
      "learning_rate": 8.182406209573091e-07,
      "loss": 5.1091,
      "step": 506
    },
    {
      "epoch": 0.06563744052820662,
      "grad_norm": 4.507810999525033,
      "learning_rate": 8.198576972833118e-07,
      "loss": 5.6196,
      "step": 507
    },
    {
      "epoch": 0.06576690293556009,
      "grad_norm": 2.451316981139926,
      "learning_rate": 8.214747736093144e-07,
      "loss": 3.7998,
      "step": 508
    },
    {
      "epoch": 0.06589636534291356,
      "grad_norm": 3.636112888850396,
      "learning_rate": 8.23091849935317e-07,
      "loss": 5.957,
      "step": 509
    },
    {
      "epoch": 0.06602582775026701,
      "grad_norm": 3.0092481611553707,
      "learning_rate": 8.247089262613196e-07,
      "loss": 4.2854,
      "step": 510
    },
    {
      "epoch": 0.06602582775026701,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8939098119735718,
      "eval_runtime": 14.3544,
      "eval_samples_per_second": 3.065,
      "eval_steps_per_second": 0.209,
      "step": 510
    },
    {
      "epoch": 0.06615529015762048,
      "grad_norm": 3.0288191538893163,
      "learning_rate": 8.263260025873223e-07,
      "loss": 4.3193,
      "step": 511
    },
    {
      "epoch": 0.06628475256497394,
      "grad_norm": 4.736715033864446,
      "learning_rate": 8.279430789133249e-07,
      "loss": 6.5122,
      "step": 512
    },
    {
      "epoch": 0.06641421497232741,
      "grad_norm": 4.116504028763779,
      "learning_rate": 8.295601552393273e-07,
      "loss": 5.3965,
      "step": 513
    },
    {
      "epoch": 0.06654367737968088,
      "grad_norm": 1.9387286014118366,
      "learning_rate": 8.311772315653299e-07,
      "loss": 3.7812,
      "step": 514
    },
    {
      "epoch": 0.06667313978703435,
      "grad_norm": 2.155192815384066,
      "learning_rate": 8.327943078913325e-07,
      "loss": 4.3325,
      "step": 515
    },
    {
      "epoch": 0.06667313978703435,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8963068127632141,
      "eval_runtime": 14.2892,
      "eval_samples_per_second": 3.079,
      "eval_steps_per_second": 0.21,
      "step": 515
    },
    {
      "epoch": 0.0668026021943878,
      "grad_norm": 4.196319636548519,
      "learning_rate": 8.344113842173351e-07,
      "loss": 5.3833,
      "step": 516
    },
    {
      "epoch": 0.06693206460174127,
      "grad_norm": 2.3249406032097872,
      "learning_rate": 8.360284605433378e-07,
      "loss": 4.3223,
      "step": 517
    },
    {
      "epoch": 0.06706152700909473,
      "grad_norm": 3.6040502213305112,
      "learning_rate": 8.376455368693403e-07,
      "loss": 4.2324,
      "step": 518
    },
    {
      "epoch": 0.0671909894164482,
      "grad_norm": 2.9002581708497632,
      "learning_rate": 8.392626131953428e-07,
      "loss": 5.1431,
      "step": 519
    },
    {
      "epoch": 0.06732045182380167,
      "grad_norm": 2.440567957752614,
      "learning_rate": 8.408796895213455e-07,
      "loss": 4.0225,
      "step": 520
    },
    {
      "epoch": 0.06732045182380167,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.896484375,
      "eval_runtime": 15.323,
      "eval_samples_per_second": 2.872,
      "eval_steps_per_second": 0.196,
      "step": 520
    },
    {
      "epoch": 0.06744991423115512,
      "grad_norm": 3.3779231054013796,
      "learning_rate": 8.42496765847348e-07,
      "loss": 5.6074,
      "step": 521
    },
    {
      "epoch": 0.06757937663850859,
      "grad_norm": 1.7600910136409669,
      "learning_rate": 8.441138421733506e-07,
      "loss": 3.5586,
      "step": 522
    },
    {
      "epoch": 0.06770883904586206,
      "grad_norm": 2.0663228752988054,
      "learning_rate": 8.457309184993533e-07,
      "loss": 3.3735,
      "step": 523
    },
    {
      "epoch": 0.06783830145321552,
      "grad_norm": 4.4520688492695815,
      "learning_rate": 8.473479948253558e-07,
      "loss": 6.2312,
      "step": 524
    },
    {
      "epoch": 0.06796776386056899,
      "grad_norm": 3.3994368259801386,
      "learning_rate": 8.489650711513583e-07,
      "loss": 4.7607,
      "step": 525
    },
    {
      "epoch": 0.06796776386056899,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8990589380264282,
      "eval_runtime": 14.6892,
      "eval_samples_per_second": 2.995,
      "eval_steps_per_second": 0.204,
      "step": 525
    },
    {
      "epoch": 0.06809722626792246,
      "grad_norm": 2.657575768338923,
      "learning_rate": 8.50582147477361e-07,
      "loss": 4.5498,
      "step": 526
    },
    {
      "epoch": 0.06822668867527591,
      "grad_norm": 2.791066485016929,
      "learning_rate": 8.521992238033635e-07,
      "loss": 4.8328,
      "step": 527
    },
    {
      "epoch": 0.06835615108262938,
      "grad_norm": 3.7134360308747096,
      "learning_rate": 8.538163001293662e-07,
      "loss": 5.7866,
      "step": 528
    },
    {
      "epoch": 0.06848561348998285,
      "grad_norm": 4.907739684354129,
      "learning_rate": 8.554333764553688e-07,
      "loss": 6.2227,
      "step": 529
    },
    {
      "epoch": 0.06861507589733631,
      "grad_norm": 2.052334601212319,
      "learning_rate": 8.570504527813713e-07,
      "loss": 3.3701,
      "step": 530
    },
    {
      "epoch": 0.06861507589733631,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.8987926244735718,
      "eval_runtime": 13.717,
      "eval_samples_per_second": 3.208,
      "eval_steps_per_second": 0.219,
      "step": 530
    },
    {
      "epoch": 0.06874453830468978,
      "grad_norm": 3.8891433361105894,
      "learning_rate": 8.58667529107374e-07,
      "loss": 4.9106,
      "step": 531
    },
    {
      "epoch": 0.06887400071204323,
      "grad_norm": 2.2893042313461036,
      "learning_rate": 8.602846054333765e-07,
      "loss": 4.3201,
      "step": 532
    },
    {
      "epoch": 0.0690034631193967,
      "grad_norm": 3.4573813278641876,
      "learning_rate": 8.61901681759379e-07,
      "loss": 5.292,
      "step": 533
    },
    {
      "epoch": 0.06913292552675017,
      "grad_norm": 2.4345783037899076,
      "learning_rate": 8.635187580853817e-07,
      "loss": 4.0806,
      "step": 534
    },
    {
      "epoch": 0.06926238793410364,
      "grad_norm": 3.1589647066428896,
      "learning_rate": 8.651358344113843e-07,
      "loss": 5.3013,
      "step": 535
    },
    {
      "epoch": 0.06926238793410364,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.900390625,
      "eval_runtime": 14.9793,
      "eval_samples_per_second": 2.937,
      "eval_steps_per_second": 0.2,
      "step": 535
    },
    {
      "epoch": 0.0693918503414571,
      "grad_norm": 4.985296325628345,
      "learning_rate": 8.667529107373869e-07,
      "loss": 6.3066,
      "step": 536
    },
    {
      "epoch": 0.06952131274881057,
      "grad_norm": 3.398097220565681,
      "learning_rate": 8.683699870633895e-07,
      "loss": 4.2542,
      "step": 537
    },
    {
      "epoch": 0.06965077515616402,
      "grad_norm": 1.7443168575181354,
      "learning_rate": 8.69987063389392e-07,
      "loss": 3.6675,
      "step": 538
    },
    {
      "epoch": 0.06978023756351749,
      "grad_norm": 2.70134571401946,
      "learning_rate": 8.716041397153946e-07,
      "loss": 4.7905,
      "step": 539
    },
    {
      "epoch": 0.06990969997087096,
      "grad_norm": 2.8409211200764792,
      "learning_rate": 8.732212160413972e-07,
      "loss": 4.2915,
      "step": 540
    },
    {
      "epoch": 0.06990969997087096,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9003018736839294,
      "eval_runtime": 14.4287,
      "eval_samples_per_second": 3.049,
      "eval_steps_per_second": 0.208,
      "step": 540
    },
    {
      "epoch": 0.07003916237822443,
      "grad_norm": 3.3697084981128893,
      "learning_rate": 8.748382923673998e-07,
      "loss": 4.6543,
      "step": 541
    },
    {
      "epoch": 0.0701686247855779,
      "grad_norm": 2.120090059916605,
      "learning_rate": 8.764553686934024e-07,
      "loss": 3.554,
      "step": 542
    },
    {
      "epoch": 0.07029808719293135,
      "grad_norm": 4.873555209018028,
      "learning_rate": 8.78072445019405e-07,
      "loss": 7.2856,
      "step": 543
    },
    {
      "epoch": 0.07042754960028481,
      "grad_norm": 4.094476922579132,
      "learning_rate": 8.796895213454077e-07,
      "loss": 6.144,
      "step": 544
    },
    {
      "epoch": 0.07055701200763828,
      "grad_norm": 1.9128453528499718,
      "learning_rate": 8.813065976714101e-07,
      "loss": 3.0903,
      "step": 545
    },
    {
      "epoch": 0.07055701200763828,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9029651880264282,
      "eval_runtime": 14.4127,
      "eval_samples_per_second": 3.053,
      "eval_steps_per_second": 0.208,
      "step": 545
    },
    {
      "epoch": 0.07068647441499175,
      "grad_norm": 3.088880716633396,
      "learning_rate": 8.829236739974127e-07,
      "loss": 5.2263,
      "step": 546
    },
    {
      "epoch": 0.07081593682234522,
      "grad_norm": 4.0050980997038454,
      "learning_rate": 8.845407503234154e-07,
      "loss": 5.7109,
      "step": 547
    },
    {
      "epoch": 0.07094539922969868,
      "grad_norm": 2.525558785824187,
      "learning_rate": 8.861578266494179e-07,
      "loss": 4.5991,
      "step": 548
    },
    {
      "epoch": 0.07107486163705214,
      "grad_norm": 3.5294589158774814,
      "learning_rate": 8.877749029754205e-07,
      "loss": 5.3726,
      "step": 549
    },
    {
      "epoch": 0.0712043240444056,
      "grad_norm": 1.744927169196934,
      "learning_rate": 8.893919793014232e-07,
      "loss": 3.8086,
      "step": 550
    },
    {
      "epoch": 0.0712043240444056,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9018110632896423,
      "eval_runtime": 14.412,
      "eval_samples_per_second": 3.053,
      "eval_steps_per_second": 0.208,
      "step": 550
    },
    {
      "epoch": 0.07133378645175907,
      "grad_norm": 3.1989162668420454,
      "learning_rate": 8.910090556274256e-07,
      "loss": 4.9919,
      "step": 551
    },
    {
      "epoch": 0.07146324885911254,
      "grad_norm": 3.4064242196300887,
      "learning_rate": 8.926261319534282e-07,
      "loss": 4.6013,
      "step": 552
    },
    {
      "epoch": 0.071592711266466,
      "grad_norm": 3.104943466065872,
      "learning_rate": 8.942432082794309e-07,
      "loss": 5.2639,
      "step": 553
    },
    {
      "epoch": 0.07172217367381946,
      "grad_norm": 3.2076788640969904,
      "learning_rate": 8.958602846054334e-07,
      "loss": 4.9648,
      "step": 554
    },
    {
      "epoch": 0.07185163608117293,
      "grad_norm": 4.384616840432897,
      "learning_rate": 8.974773609314361e-07,
      "loss": 5.7134,
      "step": 555
    },
    {
      "epoch": 0.07185163608117293,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9041193127632141,
      "eval_runtime": 14.6811,
      "eval_samples_per_second": 2.997,
      "eval_steps_per_second": 0.204,
      "step": 555
    },
    {
      "epoch": 0.0719810984885264,
      "grad_norm": 2.8606649552219032,
      "learning_rate": 8.990944372574387e-07,
      "loss": 4.3687,
      "step": 556
    },
    {
      "epoch": 0.07211056089587986,
      "grad_norm": 4.6360260728371525,
      "learning_rate": 9.007115135834413e-07,
      "loss": 5.1851,
      "step": 557
    },
    {
      "epoch": 0.07224002330323333,
      "grad_norm": 4.110115682068368,
      "learning_rate": 9.023285899094438e-07,
      "loss": 5.0361,
      "step": 558
    },
    {
      "epoch": 0.0723694857105868,
      "grad_norm": 2.569193297039986,
      "learning_rate": 9.039456662354464e-07,
      "loss": 3.9204,
      "step": 559
    },
    {
      "epoch": 0.07249894811794025,
      "grad_norm": 4.696046617483736,
      "learning_rate": 9.055627425614489e-07,
      "loss": 6.6973,
      "step": 560
    },
    {
      "epoch": 0.07249894811794025,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9022549986839294,
      "eval_runtime": 14.2253,
      "eval_samples_per_second": 3.093,
      "eval_steps_per_second": 0.211,
      "step": 560
    },
    {
      "epoch": 0.07262841052529372,
      "grad_norm": 2.7338776532354547,
      "learning_rate": 9.071798188874516e-07,
      "loss": 4.1255,
      "step": 561
    },
    {
      "epoch": 0.07275787293264718,
      "grad_norm": 1.9844796035867727,
      "learning_rate": 9.087968952134542e-07,
      "loss": 3.7235,
      "step": 562
    },
    {
      "epoch": 0.07288733534000065,
      "grad_norm": 3.447358447957195,
      "learning_rate": 9.104139715394568e-07,
      "loss": 5.0107,
      "step": 563
    },
    {
      "epoch": 0.07301679774735412,
      "grad_norm": 3.1047084367884197,
      "learning_rate": 9.120310478654593e-07,
      "loss": 4.2332,
      "step": 564
    },
    {
      "epoch": 0.07314626015470757,
      "grad_norm": 5.146389209471009,
      "learning_rate": 9.136481241914618e-07,
      "loss": 5.7676,
      "step": 565
    },
    {
      "epoch": 0.07314626015470757,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9055397510528564,
      "eval_runtime": 14.2443,
      "eval_samples_per_second": 3.089,
      "eval_steps_per_second": 0.211,
      "step": 565
    },
    {
      "epoch": 0.07327572256206104,
      "grad_norm": 4.501006598165547,
      "learning_rate": 9.152652005174645e-07,
      "loss": 5.9028,
      "step": 566
    },
    {
      "epoch": 0.0734051849694145,
      "grad_norm": 2.566261197143029,
      "learning_rate": 9.168822768434671e-07,
      "loss": 4.1855,
      "step": 567
    },
    {
      "epoch": 0.07353464737676797,
      "grad_norm": 3.318519305720438,
      "learning_rate": 9.184993531694696e-07,
      "loss": 4.4175,
      "step": 568
    },
    {
      "epoch": 0.07366410978412144,
      "grad_norm": 2.2329243207688627,
      "learning_rate": 9.201164294954723e-07,
      "loss": 3.4346,
      "step": 569
    },
    {
      "epoch": 0.07379357219147491,
      "grad_norm": 3.2854758216697997,
      "learning_rate": 9.217335058214749e-07,
      "loss": 4.0674,
      "step": 570
    },
    {
      "epoch": 0.07379357219147491,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9040305614471436,
      "eval_runtime": 14.7875,
      "eval_samples_per_second": 2.975,
      "eval_steps_per_second": 0.203,
      "step": 570
    },
    {
      "epoch": 0.07392303459882836,
      "grad_norm": 1.6680723161798654,
      "learning_rate": 9.233505821474773e-07,
      "loss": 3.1938,
      "step": 571
    },
    {
      "epoch": 0.07405249700618183,
      "grad_norm": 3.1843966375356128,
      "learning_rate": 9.2496765847348e-07,
      "loss": 5.1934,
      "step": 572
    },
    {
      "epoch": 0.0741819594135353,
      "grad_norm": 4.506517328135267,
      "learning_rate": 9.265847347994826e-07,
      "loss": 6.0032,
      "step": 573
    },
    {
      "epoch": 0.07431142182088876,
      "grad_norm": 4.845730851811751,
      "learning_rate": 9.282018111254852e-07,
      "loss": 6.4753,
      "step": 574
    },
    {
      "epoch": 0.07444088422824223,
      "grad_norm": 3.6721389107272473,
      "learning_rate": 9.298188874514878e-07,
      "loss": 5.3804,
      "step": 575
    },
    {
      "epoch": 0.07444088422824223,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9096235632896423,
      "eval_runtime": 22.1478,
      "eval_samples_per_second": 1.987,
      "eval_steps_per_second": 0.135,
      "step": 575
    },
    {
      "epoch": 0.07457034663559568,
      "grad_norm": 3.5015722139086947,
      "learning_rate": 9.314359637774904e-07,
      "loss": 4.3586,
      "step": 576
    },
    {
      "epoch": 0.07469980904294915,
      "grad_norm": 2.6449535120499448,
      "learning_rate": 9.330530401034929e-07,
      "loss": 3.9878,
      "step": 577
    },
    {
      "epoch": 0.07482927145030262,
      "grad_norm": 1.9122836693706646,
      "learning_rate": 9.346701164294955e-07,
      "loss": 3.5132,
      "step": 578
    },
    {
      "epoch": 0.07495873385765608,
      "grad_norm": 3.460105436402456,
      "learning_rate": 9.362871927554981e-07,
      "loss": 4.9424,
      "step": 579
    },
    {
      "epoch": 0.07508819626500955,
      "grad_norm": 2.3803755045750723,
      "learning_rate": 9.379042690815007e-07,
      "loss": 3.2925,
      "step": 580
    },
    {
      "epoch": 0.07508819626500955,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9056285619735718,
      "eval_runtime": 15.1217,
      "eval_samples_per_second": 2.91,
      "eval_steps_per_second": 0.198,
      "step": 580
    },
    {
      "epoch": 0.07521765867236302,
      "grad_norm": 2.5524823901163307,
      "learning_rate": 9.395213454075033e-07,
      "loss": 4.2061,
      "step": 581
    },
    {
      "epoch": 0.07534712107971647,
      "grad_norm": 3.213826628952772,
      "learning_rate": 9.41138421733506e-07,
      "loss": 5.3286,
      "step": 582
    },
    {
      "epoch": 0.07547658348706994,
      "grad_norm": 5.185873619621562,
      "learning_rate": 9.427554980595084e-07,
      "loss": 6.583,
      "step": 583
    },
    {
      "epoch": 0.07560604589442341,
      "grad_norm": 3.4682000183730515,
      "learning_rate": 9.44372574385511e-07,
      "loss": 5.1123,
      "step": 584
    },
    {
      "epoch": 0.07573550830177687,
      "grad_norm": 2.542333651184107,
      "learning_rate": 9.459896507115137e-07,
      "loss": 4.5977,
      "step": 585
    },
    {
      "epoch": 0.07573550830177687,
      "eval_PRM Accuracy": 0.37735849056603776,
      "eval_PRM F1": 0.4,
      "eval_PRM F1 AUC": 0.5238344683080147,
      "eval_PRM F1 Neg": 0.35294117647058826,
      "eval_PRM NPV": 0.22784810126582278,
      "eval_PRM Precision": 0.8148148148148148,
      "eval_PRM Recall": 0.26506024096385544,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9098011255264282,
      "eval_runtime": 15.9048,
      "eval_samples_per_second": 2.766,
      "eval_steps_per_second": 0.189,
      "step": 585
    },
    {
      "epoch": 0.07586497070913034,
      "grad_norm": 4.495054423452107,
      "learning_rate": 9.476067270375162e-07,
      "loss": 5.8911,
      "step": 586
    },
    {
      "epoch": 0.0759944331164838,
      "grad_norm": 4.286291676009975,
      "learning_rate": 9.492238033635188e-07,
      "loss": 5.875,
      "step": 587
    },
    {
      "epoch": 0.07612389552383726,
      "grad_norm": 2.2731867866477113,
      "learning_rate": 9.508408796895215e-07,
      "loss": 4.1794,
      "step": 588
    },
    {
      "epoch": 0.07625335793119073,
      "grad_norm": 4.967621782770409,
      "learning_rate": 9.52457956015524e-07,
      "loss": 5.6804,
      "step": 589
    },
    {
      "epoch": 0.0763828203385442,
      "grad_norm": 5.421279083915253,
      "learning_rate": 9.540750323415266e-07,
      "loss": 6.9507,
      "step": 590
    },
    {
      "epoch": 0.0763828203385442,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9108664989471436,
      "eval_runtime": 14.3094,
      "eval_samples_per_second": 3.075,
      "eval_steps_per_second": 0.21,
      "step": 590
    },
    {
      "epoch": 0.07651228274589766,
      "grad_norm": 3.4462083989481127,
      "learning_rate": 9.55692108667529e-07,
      "loss": 5.4868,
      "step": 591
    },
    {
      "epoch": 0.07664174515325113,
      "grad_norm": 3.8268458683344617,
      "learning_rate": 9.573091849935317e-07,
      "loss": 4.8584,
      "step": 592
    },
    {
      "epoch": 0.07677120756060458,
      "grad_norm": 1.8740977019432772,
      "learning_rate": 9.589262613195344e-07,
      "loss": 3.6318,
      "step": 593
    },
    {
      "epoch": 0.07690066996795805,
      "grad_norm": 4.548373022254427,
      "learning_rate": 9.60543337645537e-07,
      "loss": 6.3418,
      "step": 594
    },
    {
      "epoch": 0.07703013237531152,
      "grad_norm": 2.886714885850772,
      "learning_rate": 9.621604139715395e-07,
      "loss": 4.147,
      "step": 595
    },
    {
      "epoch": 0.07703013237531152,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9083806872367859,
      "eval_runtime": 13.9448,
      "eval_samples_per_second": 3.155,
      "eval_steps_per_second": 0.215,
      "step": 595
    },
    {
      "epoch": 0.07715959478266499,
      "grad_norm": 3.7486317271092116,
      "learning_rate": 9.63777490297542e-07,
      "loss": 5.3306,
      "step": 596
    },
    {
      "epoch": 0.07728905719001845,
      "grad_norm": 3.8631528248660496,
      "learning_rate": 9.653945666235447e-07,
      "loss": 4.8525,
      "step": 597
    },
    {
      "epoch": 0.07741851959737191,
      "grad_norm": 1.6416582456791071,
      "learning_rate": 9.670116429495473e-07,
      "loss": 3.3135,
      "step": 598
    },
    {
      "epoch": 0.07754798200472537,
      "grad_norm": 2.9644891715232733,
      "learning_rate": 9.686287192755498e-07,
      "loss": 3.7283,
      "step": 599
    },
    {
      "epoch": 0.07767744441207884,
      "grad_norm": 2.5950098481380723,
      "learning_rate": 9.702457956015525e-07,
      "loss": 3.395,
      "step": 600
    },
    {
      "epoch": 0.07767744441207884,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9129083752632141,
      "eval_runtime": 13.4266,
      "eval_samples_per_second": 3.277,
      "eval_steps_per_second": 0.223,
      "step": 600
    },
    {
      "epoch": 0.07780690681943231,
      "grad_norm": 2.493656861678879,
      "learning_rate": 9.718628719275551e-07,
      "loss": 3.5049,
      "step": 601
    },
    {
      "epoch": 0.07793636922678578,
      "grad_norm": 5.086132842962849,
      "learning_rate": 9.734799482535578e-07,
      "loss": 6.459,
      "step": 602
    },
    {
      "epoch": 0.07806583163413924,
      "grad_norm": 3.9699953853182386,
      "learning_rate": 9.750970245795603e-07,
      "loss": 5.4473,
      "step": 603
    },
    {
      "epoch": 0.0781952940414927,
      "grad_norm": 3.01379362488202,
      "learning_rate": 9.767141009055627e-07,
      "loss": 4.4377,
      "step": 604
    },
    {
      "epoch": 0.07832475644884616,
      "grad_norm": 2.976840641852765,
      "learning_rate": 9.783311772315654e-07,
      "loss": 3.8774,
      "step": 605
    },
    {
      "epoch": 0.07832475644884616,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9095348119735718,
      "eval_runtime": 15.5441,
      "eval_samples_per_second": 2.831,
      "eval_steps_per_second": 0.193,
      "step": 605
    },
    {
      "epoch": 0.07845421885619963,
      "grad_norm": 4.206832861987359,
      "learning_rate": 9.79948253557568e-07,
      "loss": 5.4097,
      "step": 606
    },
    {
      "epoch": 0.0785836812635531,
      "grad_norm": 5.446122871682926,
      "learning_rate": 9.815653298835705e-07,
      "loss": 6.4014,
      "step": 607
    },
    {
      "epoch": 0.07871314367090657,
      "grad_norm": 2.63538982982655,
      "learning_rate": 9.831824062095732e-07,
      "loss": 3.2788,
      "step": 608
    },
    {
      "epoch": 0.07884260607826002,
      "grad_norm": 5.477882117028058,
      "learning_rate": 9.847994825355757e-07,
      "loss": 7.3257,
      "step": 609
    },
    {
      "epoch": 0.07897206848561349,
      "grad_norm": 5.467805438149073,
      "learning_rate": 9.864165588615783e-07,
      "loss": 6.6133,
      "step": 610
    },
    {
      "epoch": 0.07897206848561349,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9138849377632141,
      "eval_runtime": 14.4609,
      "eval_samples_per_second": 3.043,
      "eval_steps_per_second": 0.207,
      "step": 610
    },
    {
      "epoch": 0.07910153089296695,
      "grad_norm": 2.902241539133023,
      "learning_rate": 9.88033635187581e-07,
      "loss": 3.7368,
      "step": 611
    },
    {
      "epoch": 0.07923099330032042,
      "grad_norm": 2.948470323457141,
      "learning_rate": 9.896507115135835e-07,
      "loss": 4.4819,
      "step": 612
    },
    {
      "epoch": 0.07936045570767389,
      "grad_norm": 3.6804164605289014,
      "learning_rate": 9.912677878395861e-07,
      "loss": 5.4275,
      "step": 613
    },
    {
      "epoch": 0.07948991811502736,
      "grad_norm": 3.308617908453641,
      "learning_rate": 9.928848641655888e-07,
      "loss": 3.865,
      "step": 614
    },
    {
      "epoch": 0.07961938052238081,
      "grad_norm": 3.0972564895323926,
      "learning_rate": 9.945019404915913e-07,
      "loss": 4.1226,
      "step": 615
    },
    {
      "epoch": 0.07961938052238081,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9142400622367859,
      "eval_runtime": 15.1752,
      "eval_samples_per_second": 2.899,
      "eval_steps_per_second": 0.198,
      "step": 615
    },
    {
      "epoch": 0.07974884292973428,
      "grad_norm": 4.431881520626385,
      "learning_rate": 9.961190168175937e-07,
      "loss": 4.9678,
      "step": 616
    },
    {
      "epoch": 0.07987830533708774,
      "grad_norm": 2.0140038228607686,
      "learning_rate": 9.977360931435964e-07,
      "loss": 3.4004,
      "step": 617
    },
    {
      "epoch": 0.08000776774444121,
      "grad_norm": 2.4860917383878918,
      "learning_rate": 9.99353169469599e-07,
      "loss": 3.9187,
      "step": 618
    },
    {
      "epoch": 0.08013723015179468,
      "grad_norm": 3.3119505720959994,
      "learning_rate": 1.0009702457956017e-06,
      "loss": 3.8672,
      "step": 619
    },
    {
      "epoch": 0.08026669255914813,
      "grad_norm": 3.718516242590131,
      "learning_rate": 1.0025873221216042e-06,
      "loss": 5.3223,
      "step": 620
    },
    {
      "epoch": 0.08026669255914813,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9167258739471436,
      "eval_runtime": 15.7604,
      "eval_samples_per_second": 2.792,
      "eval_steps_per_second": 0.19,
      "step": 620
    },
    {
      "epoch": 0.0803961549665016,
      "grad_norm": 1.6958635648600953,
      "learning_rate": 1.0042043984476069e-06,
      "loss": 3.0889,
      "step": 621
    },
    {
      "epoch": 0.08052561737385507,
      "grad_norm": 2.5357550243833566,
      "learning_rate": 1.0058214747736093e-06,
      "loss": 3.9617,
      "step": 622
    },
    {
      "epoch": 0.08065507978120853,
      "grad_norm": 2.5902109060460003,
      "learning_rate": 1.007438551099612e-06,
      "loss": 3.4467,
      "step": 623
    },
    {
      "epoch": 0.080784542188562,
      "grad_norm": 3.421798539664192,
      "learning_rate": 1.0090556274256147e-06,
      "loss": 4.6733,
      "step": 624
    },
    {
      "epoch": 0.08091400459591547,
      "grad_norm": 3.814242406216809,
      "learning_rate": 1.0106727037516171e-06,
      "loss": 4.7397,
      "step": 625
    },
    {
      "epoch": 0.08091400459591547,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9137073755264282,
      "eval_runtime": 15.0257,
      "eval_samples_per_second": 2.928,
      "eval_steps_per_second": 0.2,
      "step": 625
    },
    {
      "epoch": 0.08104346700326892,
      "grad_norm": 3.9540614587429306,
      "learning_rate": 1.0122897800776198e-06,
      "loss": 6.1907,
      "step": 626
    },
    {
      "epoch": 0.08117292941062239,
      "grad_norm": 3.946503648694806,
      "learning_rate": 1.0139068564036225e-06,
      "loss": 5.1562,
      "step": 627
    },
    {
      "epoch": 0.08130239181797586,
      "grad_norm": 2.556311376618824,
      "learning_rate": 1.015523932729625e-06,
      "loss": 3.7217,
      "step": 628
    },
    {
      "epoch": 0.08143185422532932,
      "grad_norm": 4.098755670921853,
      "learning_rate": 1.0171410090556274e-06,
      "loss": 5.8599,
      "step": 629
    },
    {
      "epoch": 0.08156131663268279,
      "grad_norm": 2.988489969751208,
      "learning_rate": 1.01875808538163e-06,
      "loss": 4.0645,
      "step": 630
    },
    {
      "epoch": 0.08156131663268279,
      "eval_PRM Accuracy": 0.37735849056603776,
      "eval_PRM F1": 0.4,
      "eval_PRM F1 AUC": 0.5238344683080147,
      "eval_PRM F1 Neg": 0.35294117647058826,
      "eval_PRM NPV": 0.22784810126582278,
      "eval_PRM Precision": 0.8148148148148148,
      "eval_PRM Recall": 0.26506024096385544,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9172585010528564,
      "eval_runtime": 14.6966,
      "eval_samples_per_second": 2.994,
      "eval_steps_per_second": 0.204,
      "step": 630
    },
    {
      "epoch": 0.08169077904003624,
      "grad_norm": 3.311031159862222,
      "learning_rate": 1.0203751617076327e-06,
      "loss": 5.4185,
      "step": 631
    },
    {
      "epoch": 0.08182024144738971,
      "grad_norm": 3.7274143308730396,
      "learning_rate": 1.0219922380336354e-06,
      "loss": 4.9485,
      "step": 632
    },
    {
      "epoch": 0.08194970385474318,
      "grad_norm": 1.8728233992325662,
      "learning_rate": 1.0236093143596378e-06,
      "loss": 3.0996,
      "step": 633
    },
    {
      "epoch": 0.08207916626209665,
      "grad_norm": 3.930862207779473,
      "learning_rate": 1.0252263906856405e-06,
      "loss": 5.2373,
      "step": 634
    },
    {
      "epoch": 0.08220862866945011,
      "grad_norm": 3.3166952982896842,
      "learning_rate": 1.026843467011643e-06,
      "loss": 4.1846,
      "step": 635
    },
    {
      "epoch": 0.08220862866945011,
      "eval_PRM Accuracy": 0.37735849056603776,
      "eval_PRM F1": 0.4,
      "eval_PRM F1 AUC": 0.5238344683080147,
      "eval_PRM F1 Neg": 0.35294117647058826,
      "eval_PRM NPV": 0.22784810126582278,
      "eval_PRM Precision": 0.8148148148148148,
      "eval_PRM Recall": 0.26506024096385544,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9192116260528564,
      "eval_runtime": 14.9267,
      "eval_samples_per_second": 2.948,
      "eval_steps_per_second": 0.201,
      "step": 635
    },
    {
      "epoch": 0.08233809107680358,
      "grad_norm": 2.1132919104041235,
      "learning_rate": 1.0284605433376456e-06,
      "loss": 3.5691,
      "step": 636
    },
    {
      "epoch": 0.08246755348415703,
      "grad_norm": 4.264581045355218,
      "learning_rate": 1.030077619663648e-06,
      "loss": 4.8135,
      "step": 637
    },
    {
      "epoch": 0.0825970158915105,
      "grad_norm": 3.601980431909174,
      "learning_rate": 1.0316946959896508e-06,
      "loss": 4.5483,
      "step": 638
    },
    {
      "epoch": 0.08272647829886397,
      "grad_norm": 2.4651396349143675,
      "learning_rate": 1.0333117723156534e-06,
      "loss": 3.4932,
      "step": 639
    },
    {
      "epoch": 0.08285594070621743,
      "grad_norm": 3.1828995893611456,
      "learning_rate": 1.0349288486416561e-06,
      "loss": 4.1899,
      "step": 640
    },
    {
      "epoch": 0.08285594070621743,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9197443127632141,
      "eval_runtime": 15.2649,
      "eval_samples_per_second": 2.882,
      "eval_steps_per_second": 0.197,
      "step": 640
    },
    {
      "epoch": 0.0829854031135709,
      "grad_norm": 5.193147904721469,
      "learning_rate": 1.0365459249676586e-06,
      "loss": 6.2974,
      "step": 641
    },
    {
      "epoch": 0.08311486552092436,
      "grad_norm": 4.983833227627457,
      "learning_rate": 1.038163001293661e-06,
      "loss": 5.7546,
      "step": 642
    },
    {
      "epoch": 0.08324432792827782,
      "grad_norm": 3.6240315909900556,
      "learning_rate": 1.0397800776196637e-06,
      "loss": 5.4492,
      "step": 643
    },
    {
      "epoch": 0.08337379033563129,
      "grad_norm": 1.803634512316626,
      "learning_rate": 1.0413971539456664e-06,
      "loss": 3.6833,
      "step": 644
    },
    {
      "epoch": 0.08350325274298476,
      "grad_norm": 4.284586551004286,
      "learning_rate": 1.0430142302716688e-06,
      "loss": 5.7485,
      "step": 645
    },
    {
      "epoch": 0.08350325274298476,
      "eval_PRM Accuracy": 0.37735849056603776,
      "eval_PRM F1": 0.4,
      "eval_PRM F1 AUC": 0.5238344683080147,
      "eval_PRM F1 Neg": 0.35294117647058826,
      "eval_PRM NPV": 0.22784810126582278,
      "eval_PRM Precision": 0.8148148148148148,
      "eval_PRM Recall": 0.26506024096385544,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9215198755264282,
      "eval_runtime": 17.0676,
      "eval_samples_per_second": 2.578,
      "eval_steps_per_second": 0.176,
      "step": 645
    },
    {
      "epoch": 0.08363271515033822,
      "grad_norm": 3.9553847939873887,
      "learning_rate": 1.0446313065976715e-06,
      "loss": 5.0854,
      "step": 646
    },
    {
      "epoch": 0.08376217755769169,
      "grad_norm": 3.8843957975852663,
      "learning_rate": 1.0462483829236742e-06,
      "loss": 5.2982,
      "step": 647
    },
    {
      "epoch": 0.08389163996504514,
      "grad_norm": 2.3196091731352615,
      "learning_rate": 1.0478654592496766e-06,
      "loss": 3.854,
      "step": 648
    },
    {
      "epoch": 0.08402110237239861,
      "grad_norm": 4.835532044194964,
      "learning_rate": 1.0494825355756793e-06,
      "loss": 5.293,
      "step": 649
    },
    {
      "epoch": 0.08415056477975208,
      "grad_norm": 2.9182714707524715,
      "learning_rate": 1.0510996119016818e-06,
      "loss": 4.1074,
      "step": 650
    },
    {
      "epoch": 0.08415056477975208,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9204545617103577,
      "eval_runtime": 14.3587,
      "eval_samples_per_second": 3.064,
      "eval_steps_per_second": 0.209,
      "step": 650
    },
    {
      "epoch": 0.08428002718710555,
      "grad_norm": 3.6171615781565603,
      "learning_rate": 1.0527166882276844e-06,
      "loss": 4.5073,
      "step": 651
    },
    {
      "epoch": 0.08440948959445901,
      "grad_norm": 3.1926590750818096,
      "learning_rate": 1.0543337645536871e-06,
      "loss": 4.1917,
      "step": 652
    },
    {
      "epoch": 0.08453895200181247,
      "grad_norm": 3.3432764264914288,
      "learning_rate": 1.0559508408796896e-06,
      "loss": 4.3716,
      "step": 653
    },
    {
      "epoch": 0.08466841440916593,
      "grad_norm": 3.71751725958284,
      "learning_rate": 1.057567917205692e-06,
      "loss": 4.2324,
      "step": 654
    },
    {
      "epoch": 0.0847978768165194,
      "grad_norm": 1.3069433915514344,
      "learning_rate": 1.0591849935316947e-06,
      "loss": 2.7007,
      "step": 655
    },
    {
      "epoch": 0.0847978768165194,
      "eval_PRM Accuracy": 0.37735849056603776,
      "eval_PRM F1": 0.4,
      "eval_PRM F1 AUC": 0.5238344683080147,
      "eval_PRM F1 Neg": 0.35294117647058826,
      "eval_PRM NPV": 0.22784810126582278,
      "eval_PRM Precision": 0.8148148148148148,
      "eval_PRM Recall": 0.26506024096385544,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9228515625,
      "eval_runtime": 14.0529,
      "eval_samples_per_second": 3.131,
      "eval_steps_per_second": 0.213,
      "step": 655
    },
    {
      "epoch": 0.08492733922387287,
      "grad_norm": 4.182457616046026,
      "learning_rate": 1.0608020698576974e-06,
      "loss": 4.8091,
      "step": 656
    },
    {
      "epoch": 0.08505680163122634,
      "grad_norm": 2.5033354725795647,
      "learning_rate": 1.0624191461837e-06,
      "loss": 3.8281,
      "step": 657
    },
    {
      "epoch": 0.0851862640385798,
      "grad_norm": 2.8090170748188514,
      "learning_rate": 1.0640362225097025e-06,
      "loss": 4.7661,
      "step": 658
    },
    {
      "epoch": 0.08531572644593326,
      "grad_norm": 3.4595552174598208,
      "learning_rate": 1.0656532988357052e-06,
      "loss": 4.7891,
      "step": 659
    },
    {
      "epoch": 0.08544518885328672,
      "grad_norm": 2.667249572922346,
      "learning_rate": 1.0672703751617078e-06,
      "loss": 3.4492,
      "step": 660
    },
    {
      "epoch": 0.08544518885328672,
      "eval_PRM Accuracy": 0.37735849056603776,
      "eval_PRM F1": 0.4,
      "eval_PRM F1 AUC": 0.5238344683080147,
      "eval_PRM F1 Neg": 0.35294117647058826,
      "eval_PRM NPV": 0.22784810126582278,
      "eval_PRM Precision": 0.8148148148148148,
      "eval_PRM Recall": 0.26506024096385544,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9223188757896423,
      "eval_runtime": 14.2847,
      "eval_samples_per_second": 3.08,
      "eval_steps_per_second": 0.21,
      "step": 660
    },
    {
      "epoch": 0.08557465126064019,
      "grad_norm": 3.815069042013388,
      "learning_rate": 1.0688874514877103e-06,
      "loss": 4.5798,
      "step": 661
    },
    {
      "epoch": 0.08570411366799366,
      "grad_norm": 3.777846261367631,
      "learning_rate": 1.070504527813713e-06,
      "loss": 4.9731,
      "step": 662
    },
    {
      "epoch": 0.08583357607534713,
      "grad_norm": 5.213592986227693,
      "learning_rate": 1.0721216041397154e-06,
      "loss": 6.7822,
      "step": 663
    },
    {
      "epoch": 0.08596303848270058,
      "grad_norm": 2.9773730804335647,
      "learning_rate": 1.073738680465718e-06,
      "loss": 3.7539,
      "step": 664
    },
    {
      "epoch": 0.08609250089005405,
      "grad_norm": 4.015777203526543,
      "learning_rate": 1.0753557567917208e-06,
      "loss": 5.2158,
      "step": 665
    },
    {
      "epoch": 0.08609250089005405,
      "eval_PRM Accuracy": 0.36792452830188677,
      "eval_PRM F1": 0.3853211009174312,
      "eval_PRM F1 AUC": 0.5178103719224726,
      "eval_PRM F1 Neg": 0.34951456310679613,
      "eval_PRM NPV": 0.225,
      "eval_PRM Precision": 0.8076923076923077,
      "eval_PRM Recall": 0.25301204819277107,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9247159361839294,
      "eval_runtime": 21.0285,
      "eval_samples_per_second": 2.092,
      "eval_steps_per_second": 0.143,
      "step": 665
    },
    {
      "epoch": 0.08622196329740751,
      "grad_norm": 2.65627042409324,
      "learning_rate": 1.0769728331177232e-06,
      "loss": 4.1765,
      "step": 666
    },
    {
      "epoch": 0.08635142570476098,
      "grad_norm": 5.125522979331907,
      "learning_rate": 1.0785899094437257e-06,
      "loss": 6.1421,
      "step": 667
    },
    {
      "epoch": 0.08648088811211445,
      "grad_norm": 3.910613713004684,
      "learning_rate": 1.0802069857697284e-06,
      "loss": 5.6338,
      "step": 668
    },
    {
      "epoch": 0.08661035051946792,
      "grad_norm": 2.728668919911025,
      "learning_rate": 1.081824062095731e-06,
      "loss": 4.1968,
      "step": 669
    },
    {
      "epoch": 0.08673981292682137,
      "grad_norm": 3.7978918043442955,
      "learning_rate": 1.0834411384217337e-06,
      "loss": 4.626,
      "step": 670
    },
    {
      "epoch": 0.08673981292682137,
      "eval_PRM Accuracy": 0.36792452830188677,
      "eval_PRM F1": 0.3853211009174312,
      "eval_PRM F1 AUC": 0.5178103719224726,
      "eval_PRM F1 Neg": 0.34951456310679613,
      "eval_PRM NPV": 0.225,
      "eval_PRM Precision": 0.8076923076923077,
      "eval_PRM Recall": 0.25301204819277107,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9281782507896423,
      "eval_runtime": 14.6199,
      "eval_samples_per_second": 3.01,
      "eval_steps_per_second": 0.205,
      "step": 670
    },
    {
      "epoch": 0.08686927533417484,
      "grad_norm": 1.7110243603847934,
      "learning_rate": 1.0850582147477362e-06,
      "loss": 3.05,
      "step": 671
    },
    {
      "epoch": 0.0869987377415283,
      "grad_norm": 6.078163216415924,
      "learning_rate": 1.0866752910737388e-06,
      "loss": 6.9224,
      "step": 672
    },
    {
      "epoch": 0.08712820014888177,
      "grad_norm": 2.441968873569535,
      "learning_rate": 1.0882923673997415e-06,
      "loss": 3.5649,
      "step": 673
    },
    {
      "epoch": 0.08725766255623524,
      "grad_norm": 5.334560482449647,
      "learning_rate": 1.089909443725744e-06,
      "loss": 5.2378,
      "step": 674
    },
    {
      "epoch": 0.08738712496358869,
      "grad_norm": 3.3562273561795313,
      "learning_rate": 1.0915265200517464e-06,
      "loss": 4.2126,
      "step": 675
    },
    {
      "epoch": 0.08738712496358869,
      "eval_PRM Accuracy": 0.37735849056603776,
      "eval_PRM F1": 0.4,
      "eval_PRM F1 AUC": 0.5238344683080147,
      "eval_PRM F1 Neg": 0.35294117647058826,
      "eval_PRM NPV": 0.22784810126582278,
      "eval_PRM Precision": 0.8148148148148148,
      "eval_PRM Recall": 0.26506024096385544,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.927734375,
      "eval_runtime": 14.9743,
      "eval_samples_per_second": 2.938,
      "eval_steps_per_second": 0.2,
      "step": 675
    },
    {
      "epoch": 0.08751658737094216,
      "grad_norm": 3.839414456101029,
      "learning_rate": 1.093143596377749e-06,
      "loss": 5.8862,
      "step": 676
    },
    {
      "epoch": 0.08764604977829563,
      "grad_norm": 1.8936817586298933,
      "learning_rate": 1.0947606727037518e-06,
      "loss": 3.3889,
      "step": 677
    },
    {
      "epoch": 0.0877755121856491,
      "grad_norm": 5.910427212711068,
      "learning_rate": 1.0963777490297544e-06,
      "loss": 7.0693,
      "step": 678
    },
    {
      "epoch": 0.08790497459300256,
      "grad_norm": 2.800044073307304,
      "learning_rate": 1.0979948253557569e-06,
      "loss": 4.5312,
      "step": 679
    },
    {
      "epoch": 0.08803443700035603,
      "grad_norm": 2.4777274689541633,
      "learning_rate": 1.0996119016817593e-06,
      "loss": 3.7324,
      "step": 680
    },
    {
      "epoch": 0.08803443700035603,
      "eval_PRM Accuracy": 0.3867924528301887,
      "eval_PRM F1": 0.4144144144144144,
      "eval_PRM F1 AUC": 0.5298585646935569,
      "eval_PRM F1 Neg": 0.3564356435643564,
      "eval_PRM NPV": 0.23076923076923078,
      "eval_PRM Precision": 0.8214285714285714,
      "eval_PRM Recall": 0.27710843373493976,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.927734375,
      "eval_runtime": 22.4193,
      "eval_samples_per_second": 1.963,
      "eval_steps_per_second": 0.134,
      "step": 680
    },
    {
      "epoch": 0.08816389940770948,
      "grad_norm": 5.07762505715575,
      "learning_rate": 1.101228978007762e-06,
      "loss": 5.6816,
      "step": 681
    },
    {
      "epoch": 0.08829336181506295,
      "grad_norm": 4.122952485942595,
      "learning_rate": 1.1028460543337647e-06,
      "loss": 4.4331,
      "step": 682
    },
    {
      "epoch": 0.08842282422241642,
      "grad_norm": 5.731260260030666,
      "learning_rate": 1.1044631306597671e-06,
      "loss": 5.0352,
      "step": 683
    },
    {
      "epoch": 0.08855228662976988,
      "grad_norm": 5.955792580882385,
      "learning_rate": 1.1060802069857698e-06,
      "loss": 5.5566,
      "step": 684
    },
    {
      "epoch": 0.08868174903712335,
      "grad_norm": 2.8779771736185853,
      "learning_rate": 1.1076972833117725e-06,
      "loss": 4.3464,
      "step": 685
    },
    {
      "epoch": 0.08868174903712335,
      "eval_PRM Accuracy": 0.37735849056603776,
      "eval_PRM F1": 0.4,
      "eval_PRM F1 AUC": 0.5238344683080147,
      "eval_PRM F1 Neg": 0.35294117647058826,
      "eval_PRM NPV": 0.22784810126582278,
      "eval_PRM Precision": 0.8148148148148148,
      "eval_PRM Recall": 0.26506024096385544,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9293323755264282,
      "eval_runtime": 14.2059,
      "eval_samples_per_second": 3.097,
      "eval_steps_per_second": 0.211,
      "step": 685
    },
    {
      "epoch": 0.0888112114444768,
      "grad_norm": 1.6034673869806007,
      "learning_rate": 1.1093143596377752e-06,
      "loss": 3.4072,
      "step": 686
    },
    {
      "epoch": 0.08894067385183027,
      "grad_norm": 4.895853758787289,
      "learning_rate": 1.1109314359637776e-06,
      "loss": 6.3848,
      "step": 687
    },
    {
      "epoch": 0.08907013625918374,
      "grad_norm": 3.283050375386971,
      "learning_rate": 1.11254851228978e-06,
      "loss": 4.6675,
      "step": 688
    },
    {
      "epoch": 0.0891995986665372,
      "grad_norm": 3.408758314451046,
      "learning_rate": 1.1141655886157827e-06,
      "loss": 3.4902,
      "step": 689
    },
    {
      "epoch": 0.08932906107389067,
      "grad_norm": 4.9801633043257905,
      "learning_rate": 1.1157826649417854e-06,
      "loss": 5.9888,
      "step": 690
    },
    {
      "epoch": 0.08932906107389067,
      "eval_PRM Accuracy": 0.36792452830188677,
      "eval_PRM F1": 0.3853211009174312,
      "eval_PRM F1 AUC": 0.5178103719224726,
      "eval_PRM F1 Neg": 0.34951456310679613,
      "eval_PRM NPV": 0.225,
      "eval_PRM Precision": 0.8076923076923077,
      "eval_PRM Recall": 0.25301204819277107,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9329723119735718,
      "eval_runtime": 14.6811,
      "eval_samples_per_second": 2.997,
      "eval_steps_per_second": 0.204,
      "step": 690
    },
    {
      "epoch": 0.08945852348124414,
      "grad_norm": 3.2296193036231067,
      "learning_rate": 1.1173997412677879e-06,
      "loss": 3.9331,
      "step": 691
    },
    {
      "epoch": 0.0895879858885976,
      "grad_norm": 4.489150987405942,
      "learning_rate": 1.1190168175937905e-06,
      "loss": 4.6777,
      "step": 692
    },
    {
      "epoch": 0.08971744829595106,
      "grad_norm": 3.433086525653378,
      "learning_rate": 1.120633893919793e-06,
      "loss": 5.6968,
      "step": 693
    },
    {
      "epoch": 0.08984691070330453,
      "grad_norm": 2.3814291124617575,
      "learning_rate": 1.1222509702457957e-06,
      "loss": 3.2441,
      "step": 694
    },
    {
      "epoch": 0.089976373110658,
      "grad_norm": 3.4705775775257206,
      "learning_rate": 1.1238680465717983e-06,
      "loss": 4.6189,
      "step": 695
    },
    {
      "epoch": 0.089976373110658,
      "eval_PRM Accuracy": 0.3584905660377358,
      "eval_PRM F1": 0.37037037037037035,
      "eval_PRM F1 AUC": 0.5117862755369303,
      "eval_PRM F1 Neg": 0.34615384615384615,
      "eval_PRM NPV": 0.2222222222222222,
      "eval_PRM Precision": 0.8,
      "eval_PRM Recall": 0.24096385542168675,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9323508739471436,
      "eval_runtime": 14.3791,
      "eval_samples_per_second": 3.06,
      "eval_steps_per_second": 0.209,
      "step": 695
    },
    {
      "epoch": 0.09010583551801146,
      "grad_norm": 1.4031427780005894,
      "learning_rate": 1.1254851228978008e-06,
      "loss": 2.8269,
      "step": 696
    },
    {
      "epoch": 0.09023529792536492,
      "grad_norm": 4.84226040077209,
      "learning_rate": 1.1271021992238035e-06,
      "loss": 6.5381,
      "step": 697
    },
    {
      "epoch": 0.09036476033271838,
      "grad_norm": 2.9830415474694476,
      "learning_rate": 1.1287192755498062e-06,
      "loss": 4.415,
      "step": 698
    },
    {
      "epoch": 0.09049422274007185,
      "grad_norm": 3.3445875539537977,
      "learning_rate": 1.1303363518758086e-06,
      "loss": 4.6343,
      "step": 699
    },
    {
      "epoch": 0.09062368514742532,
      "grad_norm": 3.340910486222267,
      "learning_rate": 1.131953428201811e-06,
      "loss": 4.5891,
      "step": 700
    },
    {
      "epoch": 0.09062368514742532,
      "eval_PRM Accuracy": 0.36792452830188677,
      "eval_PRM F1": 0.3853211009174312,
      "eval_PRM F1 AUC": 0.5178103719224726,
      "eval_PRM F1 Neg": 0.34951456310679613,
      "eval_PRM NPV": 0.225,
      "eval_PRM Precision": 0.8076923076923077,
      "eval_PRM Recall": 0.25301204819277107,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9374112486839294,
      "eval_runtime": 14.1662,
      "eval_samples_per_second": 3.106,
      "eval_steps_per_second": 0.212,
      "step": 700
    },
    {
      "epoch": 0.09075314755477878,
      "grad_norm": 3.291208314334283,
      "learning_rate": 1.1335705045278137e-06,
      "loss": 4.3103,
      "step": 701
    },
    {
      "epoch": 0.09088260996213225,
      "grad_norm": 3.7902054579688427,
      "learning_rate": 1.1351875808538164e-06,
      "loss": 4.4829,
      "step": 702
    },
    {
      "epoch": 0.0910120723694857,
      "grad_norm": 3.8765634704855967,
      "learning_rate": 1.136804657179819e-06,
      "loss": 4.8525,
      "step": 703
    },
    {
      "epoch": 0.09114153477683917,
      "grad_norm": 3.477765893551626,
      "learning_rate": 1.1384217335058215e-06,
      "loss": 4.377,
      "step": 704
    },
    {
      "epoch": 0.09127099718419264,
      "grad_norm": 4.316002808739543,
      "learning_rate": 1.1400388098318242e-06,
      "loss": 5.2102,
      "step": 705
    },
    {
      "epoch": 0.09127099718419264,
      "eval_PRM Accuracy": 0.3584905660377358,
      "eval_PRM F1": 0.37037037037037035,
      "eval_PRM F1 AUC": 0.5117862755369303,
      "eval_PRM F1 Neg": 0.34615384615384615,
      "eval_PRM NPV": 0.2222222222222222,
      "eval_PRM Precision": 0.8,
      "eval_PRM Recall": 0.24096385542168675,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9383878111839294,
      "eval_runtime": 15.1852,
      "eval_samples_per_second": 2.898,
      "eval_steps_per_second": 0.198,
      "step": 705
    },
    {
      "epoch": 0.09140045959154611,
      "grad_norm": 5.333892145532132,
      "learning_rate": 1.1416558861578267e-06,
      "loss": 6.0728,
      "step": 706
    },
    {
      "epoch": 0.09152992199889957,
      "grad_norm": 4.117906116304779,
      "learning_rate": 1.1432729624838293e-06,
      "loss": 4.6228,
      "step": 707
    },
    {
      "epoch": 0.09165938440625303,
      "grad_norm": 5.105444072776924,
      "learning_rate": 1.144890038809832e-06,
      "loss": 6.3406,
      "step": 708
    },
    {
      "epoch": 0.0917888468136065,
      "grad_norm": 2.8326974320088505,
      "learning_rate": 1.1465071151358345e-06,
      "loss": 3.3999,
      "step": 709
    },
    {
      "epoch": 0.09191830922095996,
      "grad_norm": 4.041014890890666,
      "learning_rate": 1.1481241914618371e-06,
      "loss": 5.6228,
      "step": 710
    },
    {
      "epoch": 0.09191830922095996,
      "eval_PRM Accuracy": 0.3490566037735849,
      "eval_PRM F1": 0.35514018691588783,
      "eval_PRM F1 AUC": 0.5057621791513882,
      "eval_PRM F1 Neg": 0.34285714285714286,
      "eval_PRM NPV": 0.21951219512195122,
      "eval_PRM Precision": 0.7916666666666666,
      "eval_PRM Recall": 0.2289156626506024,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9406960010528564,
      "eval_runtime": 15.2254,
      "eval_samples_per_second": 2.89,
      "eval_steps_per_second": 0.197,
      "step": 710
    },
    {
      "epoch": 0.09204777162831343,
      "grad_norm": 4.397996167048664,
      "learning_rate": 1.1497412677878398e-06,
      "loss": 5.6943,
      "step": 711
    },
    {
      "epoch": 0.0921772340356669,
      "grad_norm": 3.8179337999486886,
      "learning_rate": 1.1513583441138423e-06,
      "loss": 4.103,
      "step": 712
    },
    {
      "epoch": 0.09230669644302036,
      "grad_norm": 3.8626218672791293,
      "learning_rate": 1.1529754204398447e-06,
      "loss": 4.5566,
      "step": 713
    },
    {
      "epoch": 0.09243615885037382,
      "grad_norm": 3.3837870996613777,
      "learning_rate": 1.1545924967658474e-06,
      "loss": 4.0381,
      "step": 714
    },
    {
      "epoch": 0.09256562125772728,
      "grad_norm": 3.5064318702253794,
      "learning_rate": 1.15620957309185e-06,
      "loss": 4.5518,
      "step": 715
    },
    {
      "epoch": 0.09256562125772728,
      "eval_PRM Accuracy": 0.3584905660377358,
      "eval_PRM F1": 0.37037037037037035,
      "eval_PRM F1 AUC": 0.5117862755369303,
      "eval_PRM F1 Neg": 0.34615384615384615,
      "eval_PRM NPV": 0.2222222222222222,
      "eval_PRM Precision": 0.8,
      "eval_PRM Recall": 0.24096385542168675,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9413174986839294,
      "eval_runtime": 13.7446,
      "eval_samples_per_second": 3.201,
      "eval_steps_per_second": 0.218,
      "step": 715
    },
    {
      "epoch": 0.09269508366508075,
      "grad_norm": 1.4254962179750599,
      "learning_rate": 1.1578266494178527e-06,
      "loss": 2.6558,
      "step": 716
    },
    {
      "epoch": 0.09282454607243422,
      "grad_norm": 2.7181580173659254,
      "learning_rate": 1.1594437257438552e-06,
      "loss": 3.9551,
      "step": 717
    },
    {
      "epoch": 0.09295400847978769,
      "grad_norm": 2.1720428930117883,
      "learning_rate": 1.1610608020698579e-06,
      "loss": 3.2974,
      "step": 718
    },
    {
      "epoch": 0.09308347088714114,
      "grad_norm": 3.335922023004612,
      "learning_rate": 1.1626778783958603e-06,
      "loss": 4.5981,
      "step": 719
    },
    {
      "epoch": 0.09321293329449461,
      "grad_norm": 2.5720838066439913,
      "learning_rate": 1.164294954721863e-06,
      "loss": 3.5359,
      "step": 720
    },
    {
      "epoch": 0.09321293329449461,
      "eval_PRM Accuracy": 0.3490566037735849,
      "eval_PRM F1": 0.35514018691588783,
      "eval_PRM F1 AUC": 0.5057621791513882,
      "eval_PRM F1 Neg": 0.34285714285714286,
      "eval_PRM NPV": 0.21951219512195122,
      "eval_PRM Precision": 0.7916666666666666,
      "eval_PRM Recall": 0.2289156626506024,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9451349377632141,
      "eval_runtime": 14.817,
      "eval_samples_per_second": 2.97,
      "eval_steps_per_second": 0.202,
      "step": 720
    },
    {
      "epoch": 0.09334239570184807,
      "grad_norm": 4.412096999745951,
      "learning_rate": 1.1659120310478655e-06,
      "loss": 5.0664,
      "step": 721
    },
    {
      "epoch": 0.09347185810920154,
      "grad_norm": 4.043391021692804,
      "learning_rate": 1.1675291073738681e-06,
      "loss": 4.8105,
      "step": 722
    },
    {
      "epoch": 0.09360132051655501,
      "grad_norm": 3.8852944956904394,
      "learning_rate": 1.1691461836998708e-06,
      "loss": 4.6001,
      "step": 723
    },
    {
      "epoch": 0.09373078292390848,
      "grad_norm": 3.61702897503039,
      "learning_rate": 1.1707632600258735e-06,
      "loss": 4.9185,
      "step": 724
    },
    {
      "epoch": 0.09386024533126193,
      "grad_norm": 3.770457660094383,
      "learning_rate": 1.172380336351876e-06,
      "loss": 4.9395,
      "step": 725
    },
    {
      "epoch": 0.09386024533126193,
      "eval_PRM Accuracy": 0.3490566037735849,
      "eval_PRM F1": 0.35514018691588783,
      "eval_PRM F1 AUC": 0.5057621791513882,
      "eval_PRM F1 Neg": 0.34285714285714286,
      "eval_PRM NPV": 0.21951219512195122,
      "eval_PRM Precision": 0.7916666666666666,
      "eval_PRM Recall": 0.2289156626506024,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9439808130264282,
      "eval_runtime": 14.3355,
      "eval_samples_per_second": 3.069,
      "eval_steps_per_second": 0.209,
      "step": 725
    },
    {
      "epoch": 0.0939897077386154,
      "grad_norm": 2.422357203887003,
      "learning_rate": 1.1739974126778784e-06,
      "loss": 3.4121,
      "step": 726
    },
    {
      "epoch": 0.09411917014596886,
      "grad_norm": 3.2641001532898124,
      "learning_rate": 1.175614489003881e-06,
      "loss": 4.0156,
      "step": 727
    },
    {
      "epoch": 0.09424863255332233,
      "grad_norm": 3.056422956455596,
      "learning_rate": 1.1772315653298837e-06,
      "loss": 3.2949,
      "step": 728
    },
    {
      "epoch": 0.0943780949606758,
      "grad_norm": 4.2993300475915355,
      "learning_rate": 1.1788486416558862e-06,
      "loss": 5.209,
      "step": 729
    },
    {
      "epoch": 0.09450755736802925,
      "grad_norm": 3.7276126322748278,
      "learning_rate": 1.1804657179818889e-06,
      "loss": 4.8374,
      "step": 730
    },
    {
      "epoch": 0.09450755736802925,
      "eval_PRM Accuracy": 0.3490566037735849,
      "eval_PRM F1": 0.35514018691588783,
      "eval_PRM F1 AUC": 0.5057621791513882,
      "eval_PRM F1 Neg": 0.34285714285714286,
      "eval_PRM NPV": 0.21951219512195122,
      "eval_PRM Precision": 0.7916666666666666,
      "eval_PRM Recall": 0.2289156626506024,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9463778138160706,
      "eval_runtime": 13.9681,
      "eval_samples_per_second": 3.15,
      "eval_steps_per_second": 0.215,
      "step": 730
    },
    {
      "epoch": 0.09463701977538272,
      "grad_norm": 3.42735558167999,
      "learning_rate": 1.1820827943078915e-06,
      "loss": 4.6589,
      "step": 731
    },
    {
      "epoch": 0.09476648218273619,
      "grad_norm": 2.7507527801419918,
      "learning_rate": 1.183699870633894e-06,
      "loss": 3.436,
      "step": 732
    },
    {
      "epoch": 0.09489594459008965,
      "grad_norm": 2.8666509025023625,
      "learning_rate": 1.1853169469598967e-06,
      "loss": 4.2251,
      "step": 733
    },
    {
      "epoch": 0.09502540699744312,
      "grad_norm": 4.1237553399909475,
      "learning_rate": 1.1869340232858991e-06,
      "loss": 4.8445,
      "step": 734
    },
    {
      "epoch": 0.09515486940479659,
      "grad_norm": 4.049456218972724,
      "learning_rate": 1.1885510996119018e-06,
      "loss": 5.0388,
      "step": 735
    },
    {
      "epoch": 0.09515486940479659,
      "eval_PRM Accuracy": 0.3490566037735849,
      "eval_PRM F1": 0.35514018691588783,
      "eval_PRM F1 AUC": 0.5057621791513882,
      "eval_PRM F1 Neg": 0.34285714285714286,
      "eval_PRM NPV": 0.21951219512195122,
      "eval_PRM Precision": 0.7916666666666666,
      "eval_PRM Recall": 0.2289156626506024,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9477983117103577,
      "eval_runtime": 14.6157,
      "eval_samples_per_second": 3.01,
      "eval_steps_per_second": 0.205,
      "step": 735
    },
    {
      "epoch": 0.09528433181215004,
      "grad_norm": 2.8678986489506197,
      "learning_rate": 1.1901681759379045e-06,
      "loss": 3.9961,
      "step": 736
    },
    {
      "epoch": 0.09541379421950351,
      "grad_norm": 4.527628480815001,
      "learning_rate": 1.191785252263907e-06,
      "loss": 6.0391,
      "step": 737
    },
    {
      "epoch": 0.09554325662685698,
      "grad_norm": 2.6501271154285684,
      "learning_rate": 1.1934023285899094e-06,
      "loss": 3.6812,
      "step": 738
    },
    {
      "epoch": 0.09567271903421044,
      "grad_norm": 6.202858843561828,
      "learning_rate": 1.195019404915912e-06,
      "loss": 6.6831,
      "step": 739
    },
    {
      "epoch": 0.09580218144156391,
      "grad_norm": 3.4141775212994996,
      "learning_rate": 1.1966364812419147e-06,
      "loss": 3.7017,
      "step": 740
    },
    {
      "epoch": 0.09580218144156391,
      "eval_PRM Accuracy": 0.3490566037735849,
      "eval_PRM F1": 0.35514018691588783,
      "eval_PRM F1 AUC": 0.5057621791513882,
      "eval_PRM F1 Neg": 0.34285714285714286,
      "eval_PRM NPV": 0.21951219512195122,
      "eval_PRM Precision": 0.7916666666666666,
      "eval_PRM Recall": 0.2289156626506024,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9508167505264282,
      "eval_runtime": 14.3734,
      "eval_samples_per_second": 3.061,
      "eval_steps_per_second": 0.209,
      "step": 740
    },
    {
      "epoch": 0.09593164384891736,
      "grad_norm": 3.3812200832346067,
      "learning_rate": 1.1982535575679174e-06,
      "loss": 4.1841,
      "step": 741
    },
    {
      "epoch": 0.09606110625627083,
      "grad_norm": 2.839737276307249,
      "learning_rate": 1.1998706338939199e-06,
      "loss": 4.333,
      "step": 742
    },
    {
      "epoch": 0.0961905686636243,
      "grad_norm": 3.7117497819259992,
      "learning_rate": 1.2014877102199225e-06,
      "loss": 4.0034,
      "step": 743
    },
    {
      "epoch": 0.09632003107097777,
      "grad_norm": 1.3147220409495661,
      "learning_rate": 1.2031047865459252e-06,
      "loss": 2.5254,
      "step": 744
    },
    {
      "epoch": 0.09644949347833123,
      "grad_norm": 3.06151101114409,
      "learning_rate": 1.2047218628719277e-06,
      "loss": 3.8535,
      "step": 745
    },
    {
      "epoch": 0.09644949347833123,
      "eval_PRM Accuracy": 0.3490566037735849,
      "eval_PRM F1": 0.35514018691588783,
      "eval_PRM F1 AUC": 0.5057621791513882,
      "eval_PRM F1 Neg": 0.34285714285714286,
      "eval_PRM NPV": 0.21951219512195122,
      "eval_PRM Precision": 0.7916666666666666,
      "eval_PRM Recall": 0.2289156626506024,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9534801244735718,
      "eval_runtime": 14.1651,
      "eval_samples_per_second": 3.106,
      "eval_steps_per_second": 0.212,
      "step": 745
    },
    {
      "epoch": 0.0965789558856847,
      "grad_norm": 2.777819300257185,
      "learning_rate": 1.2063389391979301e-06,
      "loss": 3.231,
      "step": 746
    },
    {
      "epoch": 0.09670841829303815,
      "grad_norm": 2.230023110319025,
      "learning_rate": 1.2079560155239328e-06,
      "loss": 3.4385,
      "step": 747
    },
    {
      "epoch": 0.09683788070039162,
      "grad_norm": 4.259154743643278,
      "learning_rate": 1.2095730918499355e-06,
      "loss": 4.5786,
      "step": 748
    },
    {
      "epoch": 0.09696734310774509,
      "grad_norm": 2.4180331921350446,
      "learning_rate": 1.2111901681759381e-06,
      "loss": 3.585,
      "step": 749
    },
    {
      "epoch": 0.09709680551509856,
      "grad_norm": 1.8441265887393403,
      "learning_rate": 1.2128072445019406e-06,
      "loss": 3.1641,
      "step": 750
    },
    {
      "epoch": 0.09709680551509856,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.33962264150943394,
      "eval_PRM F1 AUC": 0.499738082765846,
      "eval_PRM F1 Neg": 0.33962264150943394,
      "eval_PRM NPV": 0.21686746987951808,
      "eval_PRM Precision": 0.782608695652174,
      "eval_PRM Recall": 0.21686746987951808,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9540128111839294,
      "eval_runtime": 13.4035,
      "eval_samples_per_second": 3.283,
      "eval_steps_per_second": 0.224,
      "step": 750
    },
    {
      "epoch": 0.09722626792245202,
      "grad_norm": 2.5901463716558855,
      "learning_rate": 1.214424320827943e-06,
      "loss": 3.4058,
      "step": 751
    },
    {
      "epoch": 0.09735573032980548,
      "grad_norm": 4.921828335737435,
      "learning_rate": 1.2160413971539457e-06,
      "loss": 5.6348,
      "step": 752
    },
    {
      "epoch": 0.09748519273715894,
      "grad_norm": 5.598306921120383,
      "learning_rate": 1.2176584734799484e-06,
      "loss": 5.4517,
      "step": 753
    },
    {
      "epoch": 0.09761465514451241,
      "grad_norm": 3.0004463970660735,
      "learning_rate": 1.219275549805951e-06,
      "loss": 3.6851,
      "step": 754
    },
    {
      "epoch": 0.09774411755186588,
      "grad_norm": 3.023154824328046,
      "learning_rate": 1.2208926261319535e-06,
      "loss": 3.8838,
      "step": 755
    },
    {
      "epoch": 0.09774411755186588,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.33962264150943394,
      "eval_PRM F1 AUC": 0.499738082765846,
      "eval_PRM F1 Neg": 0.33962264150943394,
      "eval_PRM NPV": 0.21686746987951808,
      "eval_PRM Precision": 0.782608695652174,
      "eval_PRM Recall": 0.21686746987951808,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9588068127632141,
      "eval_runtime": 14.57,
      "eval_samples_per_second": 3.02,
      "eval_steps_per_second": 0.206,
      "step": 755
    },
    {
      "epoch": 0.09787357995921935,
      "grad_norm": 3.3491400623033645,
      "learning_rate": 1.2225097024579562e-06,
      "loss": 4.5225,
      "step": 756
    },
    {
      "epoch": 0.09800304236657281,
      "grad_norm": 2.4926367191547794,
      "learning_rate": 1.2241267787839586e-06,
      "loss": 3.6133,
      "step": 757
    },
    {
      "epoch": 0.09813250477392627,
      "grad_norm": 5.250129071006987,
      "learning_rate": 1.2257438551099613e-06,
      "loss": 5.2407,
      "step": 758
    },
    {
      "epoch": 0.09826196718127973,
      "grad_norm": 4.938303635815838,
      "learning_rate": 1.2273609314359638e-06,
      "loss": 5.2261,
      "step": 759
    },
    {
      "epoch": 0.0983914295886332,
      "grad_norm": 2.892426500916422,
      "learning_rate": 1.2289780077619664e-06,
      "loss": 4.4419,
      "step": 760
    },
    {
      "epoch": 0.0983914295886332,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.33962264150943394,
      "eval_PRM F1 AUC": 0.499738082765846,
      "eval_PRM F1 Neg": 0.33962264150943394,
      "eval_PRM NPV": 0.21686746987951808,
      "eval_PRM Precision": 0.782608695652174,
      "eval_PRM Recall": 0.21686746987951808,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9600496888160706,
      "eval_runtime": 14.8324,
      "eval_samples_per_second": 2.966,
      "eval_steps_per_second": 0.202,
      "step": 760
    },
    {
      "epoch": 0.09852089199598667,
      "grad_norm": 4.178198556986774,
      "learning_rate": 1.2305950840879691e-06,
      "loss": 5.0015,
      "step": 761
    },
    {
      "epoch": 0.09865035440334013,
      "grad_norm": 3.0622134145854276,
      "learning_rate": 1.2322121604139718e-06,
      "loss": 4.0886,
      "step": 762
    },
    {
      "epoch": 0.09877981681069359,
      "grad_norm": 5.34613239740492,
      "learning_rate": 1.2338292367399742e-06,
      "loss": 5.3184,
      "step": 763
    },
    {
      "epoch": 0.09890927921804706,
      "grad_norm": 3.8116626392907333,
      "learning_rate": 1.2354463130659767e-06,
      "loss": 5.4507,
      "step": 764
    },
    {
      "epoch": 0.09903874162540052,
      "grad_norm": 2.9408501223754486,
      "learning_rate": 1.2370633893919794e-06,
      "loss": 3.7075,
      "step": 765
    },
    {
      "epoch": 0.09903874162540052,
      "eval_PRM Accuracy": 0.3490566037735849,
      "eval_PRM F1": 0.35514018691588783,
      "eval_PRM F1 AUC": 0.5057621791513882,
      "eval_PRM F1 Neg": 0.34285714285714286,
      "eval_PRM NPV": 0.21951219512195122,
      "eval_PRM Precision": 0.7916666666666666,
      "eval_PRM Recall": 0.2289156626506024,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9600496888160706,
      "eval_runtime": 14.1306,
      "eval_samples_per_second": 3.114,
      "eval_steps_per_second": 0.212,
      "step": 765
    },
    {
      "epoch": 0.09916820403275399,
      "grad_norm": 6.713956462957203,
      "learning_rate": 1.238680465717982e-06,
      "loss": 6.6514,
      "step": 766
    },
    {
      "epoch": 0.09929766644010746,
      "grad_norm": 3.7341104005926904,
      "learning_rate": 1.2402975420439845e-06,
      "loss": 4.0283,
      "step": 767
    },
    {
      "epoch": 0.09942712884746092,
      "grad_norm": 4.316832455260363,
      "learning_rate": 1.2419146183699872e-06,
      "loss": 4.8613,
      "step": 768
    },
    {
      "epoch": 0.09955659125481438,
      "grad_norm": 4.507170174269931,
      "learning_rate": 1.2435316946959898e-06,
      "loss": 5.1484,
      "step": 769
    },
    {
      "epoch": 0.09968605366216785,
      "grad_norm": 6.561329419759221,
      "learning_rate": 1.2451487710219923e-06,
      "loss": 6.2466,
      "step": 770
    },
    {
      "epoch": 0.09968605366216785,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.33962264150943394,
      "eval_PRM F1 AUC": 0.499738082765846,
      "eval_PRM F1 Neg": 0.33962264150943394,
      "eval_PRM NPV": 0.21686746987951808,
      "eval_PRM Precision": 0.782608695652174,
      "eval_PRM Recall": 0.21686746987951808,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9601384997367859,
      "eval_runtime": 14.9473,
      "eval_samples_per_second": 2.944,
      "eval_steps_per_second": 0.201,
      "step": 770
    },
    {
      "epoch": 0.09981551606952131,
      "grad_norm": 6.469504466095085,
      "learning_rate": 1.246765847347995e-06,
      "loss": 6.5288,
      "step": 771
    },
    {
      "epoch": 0.09994497847687478,
      "grad_norm": 4.861047041568011,
      "learning_rate": 1.2483829236739974e-06,
      "loss": 5.3623,
      "step": 772
    },
    {
      "epoch": 0.10007444088422825,
      "grad_norm": 5.996185337759073,
      "learning_rate": 1.25e-06,
      "loss": 6.0386,
      "step": 773
    },
    {
      "epoch": 0.1002039032915817,
      "grad_norm": 2.1163860189694,
      "learning_rate": 1.249999936165541e-06,
      "loss": 3.3276,
      "step": 774
    },
    {
      "epoch": 0.10033336569893517,
      "grad_norm": 5.838219278177589,
      "learning_rate": 1.2499997446621767e-06,
      "loss": 5.8081,
      "step": 775
    },
    {
      "epoch": 0.10033336569893517,
      "eval_PRM Accuracy": 0.3490566037735849,
      "eval_PRM F1": 0.35514018691588783,
      "eval_PRM F1 AUC": 0.5057621791513882,
      "eval_PRM F1 Neg": 0.34285714285714286,
      "eval_PRM NPV": 0.21951219512195122,
      "eval_PRM Precision": 0.7916666666666666,
      "eval_PRM Recall": 0.2289156626506024,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9632457494735718,
      "eval_runtime": 14.1007,
      "eval_samples_per_second": 3.12,
      "eval_steps_per_second": 0.213,
      "step": 775
    },
    {
      "epoch": 0.10046282810628863,
      "grad_norm": 2.2235757549701107,
      "learning_rate": 1.2499994254899465e-06,
      "loss": 3.8042,
      "step": 776
    },
    {
      "epoch": 0.1005922905136421,
      "grad_norm": 4.50003441955947,
      "learning_rate": 1.2499989786489155e-06,
      "loss": 5.3418,
      "step": 777
    },
    {
      "epoch": 0.10072175292099557,
      "grad_norm": 5.251162335967897,
      "learning_rate": 1.249998404139175e-06,
      "loss": 6.5532,
      "step": 778
    },
    {
      "epoch": 0.10085121532834904,
      "grad_norm": 3.6217613110813027,
      "learning_rate": 1.2499977019608421e-06,
      "loss": 4.2666,
      "step": 779
    },
    {
      "epoch": 0.10098067773570249,
      "grad_norm": 6.651999256435771,
      "learning_rate": 1.2499968721140606e-06,
      "loss": 5.873,
      "step": 780
    },
    {
      "epoch": 0.10098067773570249,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.33962264150943394,
      "eval_PRM F1 AUC": 0.499738082765846,
      "eval_PRM F1 Neg": 0.33962264150943394,
      "eval_PRM NPV": 0.21686746987951808,
      "eval_PRM Precision": 0.782608695652174,
      "eval_PRM Recall": 0.21686746987951808,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9674183130264282,
      "eval_runtime": 13.1967,
      "eval_samples_per_second": 3.334,
      "eval_steps_per_second": 0.227,
      "step": 780
    },
    {
      "epoch": 0.10111014014305596,
      "grad_norm": 1.941792858648424,
      "learning_rate": 1.2499959145989998e-06,
      "loss": 3.707,
      "step": 781
    },
    {
      "epoch": 0.10123960255040942,
      "grad_norm": 1.5217448607721415,
      "learning_rate": 1.2499948294158554e-06,
      "loss": 2.7529,
      "step": 782
    },
    {
      "epoch": 0.10136906495776289,
      "grad_norm": 6.527780107406759,
      "learning_rate": 1.249993616564849e-06,
      "loss": 6.4536,
      "step": 783
    },
    {
      "epoch": 0.10149852736511636,
      "grad_norm": 4.506007033579019,
      "learning_rate": 1.2499922760462283e-06,
      "loss": 4.9507,
      "step": 784
    },
    {
      "epoch": 0.10162798977246981,
      "grad_norm": 4.163520430297341,
      "learning_rate": 1.2499908078602673e-06,
      "loss": 4.0503,
      "step": 785
    },
    {
      "epoch": 0.10162798977246981,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.33962264150943394,
      "eval_PRM F1 AUC": 0.499738082765846,
      "eval_PRM F1 Neg": 0.33962264150943394,
      "eval_PRM NPV": 0.21686746987951808,
      "eval_PRM Precision": 0.782608695652174,
      "eval_PRM Recall": 0.21686746987951808,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9685724377632141,
      "eval_runtime": 13.1939,
      "eval_samples_per_second": 3.335,
      "eval_steps_per_second": 0.227,
      "step": 785
    },
    {
      "epoch": 0.10175745217982328,
      "grad_norm": 4.349718521857898,
      "learning_rate": 1.249989212007266e-06,
      "loss": 4.1675,
      "step": 786
    },
    {
      "epoch": 0.10188691458717675,
      "grad_norm": 3.9889110504572254,
      "learning_rate": 1.2499874884875499e-06,
      "loss": 4.5693,
      "step": 787
    },
    {
      "epoch": 0.10201637699453021,
      "grad_norm": 4.382411355963895,
      "learning_rate": 1.2499856373014715e-06,
      "loss": 5.7251,
      "step": 788
    },
    {
      "epoch": 0.10214583940188368,
      "grad_norm": 5.396796100788135,
      "learning_rate": 1.2499836584494087e-06,
      "loss": 5.2334,
      "step": 789
    },
    {
      "epoch": 0.10227530180923715,
      "grad_norm": 1.6618567265559632,
      "learning_rate": 1.2499815519317659e-06,
      "loss": 2.562,
      "step": 790
    },
    {
      "epoch": 0.10227530180923715,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3238095238095238,
      "eval_PRM F1 AUC": 0.4937139863803038,
      "eval_PRM F1 Neg": 0.3364485981308411,
      "eval_PRM NPV": 0.21428571428571427,
      "eval_PRM Precision": 0.7727272727272727,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9685724377632141,
      "eval_runtime": 14.7448,
      "eval_samples_per_second": 2.984,
      "eval_steps_per_second": 0.203,
      "step": 790
    },
    {
      "epoch": 0.1024047642165906,
      "grad_norm": 1.4166413772022242,
      "learning_rate": 1.2499793177489731e-06,
      "loss": 3.0095,
      "step": 791
    },
    {
      "epoch": 0.10253422662394407,
      "grad_norm": 3.252997108285663,
      "learning_rate": 1.249976955901487e-06,
      "loss": 4.3838,
      "step": 792
    },
    {
      "epoch": 0.10266368903129754,
      "grad_norm": 2.702674167176806,
      "learning_rate": 1.2499744663897902e-06,
      "loss": 4.156,
      "step": 793
    },
    {
      "epoch": 0.102793151438651,
      "grad_norm": 3.7037228796285944,
      "learning_rate": 1.2499718492143908e-06,
      "loss": 4.1729,
      "step": 794
    },
    {
      "epoch": 0.10292261384600447,
      "grad_norm": 4.174273470864699,
      "learning_rate": 1.2499691043758235e-06,
      "loss": 4.1172,
      "step": 795
    },
    {
      "epoch": 0.10292261384600447,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.33962264150943394,
      "eval_PRM F1 AUC": 0.499738082765846,
      "eval_PRM F1 Neg": 0.33962264150943394,
      "eval_PRM NPV": 0.21686746987951808,
      "eval_PRM Precision": 0.782608695652174,
      "eval_PRM Recall": 0.21686746987951808,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9720348119735718,
      "eval_runtime": 13.008,
      "eval_samples_per_second": 3.383,
      "eval_steps_per_second": 0.231,
      "step": 795
    },
    {
      "epoch": 0.10305207625335792,
      "grad_norm": 3.068136111633999,
      "learning_rate": 1.2499662318746493e-06,
      "loss": 4.3054,
      "step": 796
    },
    {
      "epoch": 0.10318153866071139,
      "grad_norm": 4.518625817828032,
      "learning_rate": 1.2499632317114545e-06,
      "loss": 4.9478,
      "step": 797
    },
    {
      "epoch": 0.10331100106806486,
      "grad_norm": 3.2489296991270202,
      "learning_rate": 1.2499601038868525e-06,
      "loss": 4.1729,
      "step": 798
    },
    {
      "epoch": 0.10344046347541833,
      "grad_norm": 4.141141218186075,
      "learning_rate": 1.2499568484014818e-06,
      "loss": 3.8232,
      "step": 799
    },
    {
      "epoch": 0.1035699258827718,
      "grad_norm": 4.319817471531137,
      "learning_rate": 1.2499534652560074e-06,
      "loss": 5.1089,
      "step": 800
    },
    {
      "epoch": 0.1035699258827718,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3238095238095238,
      "eval_PRM F1 AUC": 0.4937139863803038,
      "eval_PRM F1 Neg": 0.3364485981308411,
      "eval_PRM NPV": 0.21428571428571427,
      "eval_PRM Precision": 0.7727272727272727,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9743430614471436,
      "eval_runtime": 15.319,
      "eval_samples_per_second": 2.872,
      "eval_steps_per_second": 0.196,
      "step": 800
    },
    {
      "epoch": 0.10369938829012526,
      "grad_norm": 2.9014865333487654,
      "learning_rate": 1.2499499544511205e-06,
      "loss": 4.2224,
      "step": 801
    },
    {
      "epoch": 0.10382885069747871,
      "grad_norm": 4.725470683614037,
      "learning_rate": 1.2499463159875382e-06,
      "loss": 5.0999,
      "step": 802
    },
    {
      "epoch": 0.10395831310483218,
      "grad_norm": 4.298496724704148,
      "learning_rate": 1.249942549866004e-06,
      "loss": 4.4946,
      "step": 803
    },
    {
      "epoch": 0.10408777551218565,
      "grad_norm": 5.453434201047757,
      "learning_rate": 1.2499386560872869e-06,
      "loss": 5.1079,
      "step": 804
    },
    {
      "epoch": 0.10421723791953912,
      "grad_norm": 7.189806185237564,
      "learning_rate": 1.2499346346521823e-06,
      "loss": 6.4219,
      "step": 805
    },
    {
      "epoch": 0.10421723791953912,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.33962264150943394,
      "eval_PRM F1 AUC": 0.499738082765846,
      "eval_PRM F1 Neg": 0.33962264150943394,
      "eval_PRM NPV": 0.21686746987951808,
      "eval_PRM Precision": 0.782608695652174,
      "eval_PRM Recall": 0.21686746987951808,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9779829382896423,
      "eval_runtime": 15.2178,
      "eval_samples_per_second": 2.891,
      "eval_steps_per_second": 0.197,
      "step": 805
    },
    {
      "epoch": 0.10434670032689258,
      "grad_norm": 2.727236830674406,
      "learning_rate": 1.249930485561512e-06,
      "loss": 3.8376,
      "step": 806
    },
    {
      "epoch": 0.10447616273424604,
      "grad_norm": 3.3423763458003077,
      "learning_rate": 1.2499262088161227e-06,
      "loss": 3.5952,
      "step": 807
    },
    {
      "epoch": 0.1046056251415995,
      "grad_norm": 2.6734807471168778,
      "learning_rate": 1.249921804416889e-06,
      "loss": 3.7695,
      "step": 808
    },
    {
      "epoch": 0.10473508754895297,
      "grad_norm": 3.7526485759162194,
      "learning_rate": 1.24991727236471e-06,
      "loss": 4.993,
      "step": 809
    },
    {
      "epoch": 0.10486454995630644,
      "grad_norm": 1.8145850363033316,
      "learning_rate": 1.2499126126605117e-06,
      "loss": 2.9448,
      "step": 810
    },
    {
      "epoch": 0.10486454995630644,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3238095238095238,
      "eval_PRM F1 AUC": 0.4937139863803038,
      "eval_PRM F1 Neg": 0.3364485981308411,
      "eval_PRM NPV": 0.21428571428571427,
      "eval_PRM Precision": 0.7727272727272727,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9822443127632141,
      "eval_runtime": 14.2835,
      "eval_samples_per_second": 3.08,
      "eval_steps_per_second": 0.21,
      "step": 810
    },
    {
      "epoch": 0.1049940123636599,
      "grad_norm": 2.451000170082706,
      "learning_rate": 1.2499078253052457e-06,
      "loss": 3.1382,
      "step": 811
    },
    {
      "epoch": 0.10512347477101337,
      "grad_norm": 2.65749940033498,
      "learning_rate": 1.2499029102998902e-06,
      "loss": 3.6758,
      "step": 812
    },
    {
      "epoch": 0.10525293717836683,
      "grad_norm": 3.503890377293656,
      "learning_rate": 1.2498978676454491e-06,
      "loss": 4.5347,
      "step": 813
    },
    {
      "epoch": 0.1053823995857203,
      "grad_norm": 5.026059938552501,
      "learning_rate": 1.2498926973429524e-06,
      "loss": 5.0806,
      "step": 814
    },
    {
      "epoch": 0.10551186199307376,
      "grad_norm": 5.269285147372982,
      "learning_rate": 1.2498873993934564e-06,
      "loss": 6.0195,
      "step": 815
    },
    {
      "epoch": 0.10551186199307376,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.33962264150943394,
      "eval_PRM F1 AUC": 0.499738082765846,
      "eval_PRM F1 Neg": 0.33962264150943394,
      "eval_PRM NPV": 0.21686746987951808,
      "eval_PRM Precision": 0.782608695652174,
      "eval_PRM Recall": 0.21686746987951808,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9808238744735718,
      "eval_runtime": 14.1579,
      "eval_samples_per_second": 3.108,
      "eval_steps_per_second": 0.212,
      "step": 815
    },
    {
      "epoch": 0.10564132440042723,
      "grad_norm": 3.753898840344595,
      "learning_rate": 1.2498819737980431e-06,
      "loss": 4.2593,
      "step": 816
    },
    {
      "epoch": 0.1057707868077807,
      "grad_norm": 4.208880584724217,
      "learning_rate": 1.2498764205578206e-06,
      "loss": 4.4756,
      "step": 817
    },
    {
      "epoch": 0.10590024921513416,
      "grad_norm": 3.9561165975523243,
      "learning_rate": 1.249870739673924e-06,
      "loss": 4.1987,
      "step": 818
    },
    {
      "epoch": 0.10602971162248762,
      "grad_norm": 6.03267265679179,
      "learning_rate": 1.249864931147513e-06,
      "loss": 5.2749,
      "step": 819
    },
    {
      "epoch": 0.10615917402984108,
      "grad_norm": 1.7585393235758906,
      "learning_rate": 1.2498589949797742e-06,
      "loss": 3.4028,
      "step": 820
    },
    {
      "epoch": 0.10615917402984108,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3238095238095238,
      "eval_PRM F1 AUC": 0.4937139863803038,
      "eval_PRM F1 Neg": 0.3364485981308411,
      "eval_PRM NPV": 0.21428571428571427,
      "eval_PRM Precision": 0.7727272727272727,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9831321239471436,
      "eval_runtime": 14.5652,
      "eval_samples_per_second": 3.021,
      "eval_steps_per_second": 0.206,
      "step": 820
    },
    {
      "epoch": 0.10628863643719455,
      "grad_norm": 1.9249061148750881,
      "learning_rate": 1.2498529311719207e-06,
      "loss": 3.2039,
      "step": 821
    },
    {
      "epoch": 0.10641809884454802,
      "grad_norm": 4.411019319171776,
      "learning_rate": 1.2498467397251908e-06,
      "loss": 4.5864,
      "step": 822
    },
    {
      "epoch": 0.10654756125190148,
      "grad_norm": 3.0029360297239838,
      "learning_rate": 1.249840420640849e-06,
      "loss": 3.7687,
      "step": 823
    },
    {
      "epoch": 0.10667702365925494,
      "grad_norm": 7.898229005186057,
      "learning_rate": 1.2498339739201865e-06,
      "loss": 6.8633,
      "step": 824
    },
    {
      "epoch": 0.1068064860666084,
      "grad_norm": 2.7280314836868014,
      "learning_rate": 1.24982739956452e-06,
      "loss": 3.6343,
      "step": 825
    },
    {
      "epoch": 0.1068064860666084,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3238095238095238,
      "eval_PRM F1 AUC": 0.4937139863803038,
      "eval_PRM F1 Neg": 0.3364485981308411,
      "eval_PRM NPV": 0.21428571428571427,
      "eval_PRM Precision": 0.7727272727272727,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9847301244735718,
      "eval_runtime": 14.6771,
      "eval_samples_per_second": 2.998,
      "eval_steps_per_second": 0.204,
      "step": 825
    },
    {
      "epoch": 0.10693594847396187,
      "grad_norm": 3.461807695266605,
      "learning_rate": 1.2498206975751924e-06,
      "loss": 4.4724,
      "step": 826
    },
    {
      "epoch": 0.10706541088131534,
      "grad_norm": 4.235149535224669,
      "learning_rate": 1.2498138679535728e-06,
      "loss": 4.4229,
      "step": 827
    },
    {
      "epoch": 0.10719487328866881,
      "grad_norm": 1.86838832461481,
      "learning_rate": 1.2498069107010563e-06,
      "loss": 2.9741,
      "step": 828
    },
    {
      "epoch": 0.10732433569602227,
      "grad_norm": 3.1641926300180683,
      "learning_rate": 1.2497998258190641e-06,
      "loss": 4.3408,
      "step": 829
    },
    {
      "epoch": 0.10745379810337573,
      "grad_norm": 3.9061723174447787,
      "learning_rate": 1.2497926133090432e-06,
      "loss": 4.1934,
      "step": 830
    },
    {
      "epoch": 0.10745379810337573,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3238095238095238,
      "eval_PRM F1 AUC": 0.4937139863803038,
      "eval_PRM F1 Neg": 0.3364485981308411,
      "eval_PRM NPV": 0.21428571428571427,
      "eval_PRM Precision": 0.7727272727272727,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9897016882896423,
      "eval_runtime": 14.44,
      "eval_samples_per_second": 3.047,
      "eval_steps_per_second": 0.208,
      "step": 830
    },
    {
      "epoch": 0.1075832605107292,
      "grad_norm": 2.628502672134275,
      "learning_rate": 1.2497852731724673e-06,
      "loss": 3.6946,
      "step": 831
    },
    {
      "epoch": 0.10771272291808266,
      "grad_norm": 4.882329550794728,
      "learning_rate": 1.2497778054108353e-06,
      "loss": 5.0615,
      "step": 832
    },
    {
      "epoch": 0.10784218532543613,
      "grad_norm": 5.284342736005965,
      "learning_rate": 1.2497702100256732e-06,
      "loss": 5.6089,
      "step": 833
    },
    {
      "epoch": 0.1079716477327896,
      "grad_norm": 2.4447431854678503,
      "learning_rate": 1.2497624870185319e-06,
      "loss": 3.5361,
      "step": 834
    },
    {
      "epoch": 0.10810111014014305,
      "grad_norm": 4.223731188362758,
      "learning_rate": 1.2497546363909894e-06,
      "loss": 3.7925,
      "step": 835
    },
    {
      "epoch": 0.10810111014014305,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3238095238095238,
      "eval_PRM F1 AUC": 0.4937139863803038,
      "eval_PRM F1 Neg": 0.3364485981308411,
      "eval_PRM NPV": 0.21428571428571427,
      "eval_PRM Precision": 0.7727272727272727,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9895241260528564,
      "eval_runtime": 14.3143,
      "eval_samples_per_second": 3.074,
      "eval_steps_per_second": 0.21,
      "step": 835
    },
    {
      "epoch": 0.10823057254749652,
      "grad_norm": 5.529833391656926,
      "learning_rate": 1.249746658144649e-06,
      "loss": 6.0088,
      "step": 836
    },
    {
      "epoch": 0.10836003495484998,
      "grad_norm": 4.951873947610698,
      "learning_rate": 1.2497385522811408e-06,
      "loss": 5.6558,
      "step": 837
    },
    {
      "epoch": 0.10848949736220345,
      "grad_norm": 2.973231011524131,
      "learning_rate": 1.2497303188021205e-06,
      "loss": 3.115,
      "step": 838
    },
    {
      "epoch": 0.10861895976955692,
      "grad_norm": 2.4066000784751544,
      "learning_rate": 1.2497219577092696e-06,
      "loss": 3.5962,
      "step": 839
    },
    {
      "epoch": 0.10874842217691039,
      "grad_norm": 3.5217087789139656,
      "learning_rate": 1.2497134690042965e-06,
      "loss": 4.5015,
      "step": 840
    },
    {
      "epoch": 0.10874842217691039,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3238095238095238,
      "eval_PRM F1 AUC": 0.4937139863803038,
      "eval_PRM F1 Neg": 0.3364485981308411,
      "eval_PRM NPV": 0.21428571428571427,
      "eval_PRM Precision": 0.7727272727272727,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9920987486839294,
      "eval_runtime": 13.7161,
      "eval_samples_per_second": 3.208,
      "eval_steps_per_second": 0.219,
      "step": 840
    },
    {
      "epoch": 0.10887788458426384,
      "grad_norm": 2.852173896730775,
      "learning_rate": 1.249704852688935e-06,
      "loss": 3.4556,
      "step": 841
    },
    {
      "epoch": 0.10900734699161731,
      "grad_norm": 3.4229498964932374,
      "learning_rate": 1.249696108764945e-06,
      "loss": 3.895,
      "step": 842
    },
    {
      "epoch": 0.10913680939897077,
      "grad_norm": 4.3670371394492795,
      "learning_rate": 1.249687237234113e-06,
      "loss": 3.9854,
      "step": 843
    },
    {
      "epoch": 0.10926627180632424,
      "grad_norm": 6.42217369367119,
      "learning_rate": 1.2496782380982509e-06,
      "loss": 6.5903,
      "step": 844
    },
    {
      "epoch": 0.10939573421367771,
      "grad_norm": 4.845601383962569,
      "learning_rate": 1.249669111359197e-06,
      "loss": 4.6348,
      "step": 845
    },
    {
      "epoch": 0.10939573421367771,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.3269230769230769,
      "eval_PRM F1 AUC": 0.5154531168150864,
      "eval_PRM F1 Neg": 0.35185185185185186,
      "eval_PRM NPV": 0.2235294117647059,
      "eval_PRM Precision": 0.8095238095238095,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.8260869565217391,
      "eval_loss": 0.9969815611839294,
      "eval_runtime": 14.0664,
      "eval_samples_per_second": 3.128,
      "eval_steps_per_second": 0.213,
      "step": 845
    },
    {
      "epoch": 0.10952519662103116,
      "grad_norm": 3.9497601962992634,
      "learning_rate": 1.2496598570188156e-06,
      "loss": 4.9312,
      "step": 846
    },
    {
      "epoch": 0.10965465902838463,
      "grad_norm": 3.731438089599192,
      "learning_rate": 1.249650475078997e-06,
      "loss": 3.9316,
      "step": 847
    },
    {
      "epoch": 0.1097841214357381,
      "grad_norm": 1.7497549960975611,
      "learning_rate": 1.249640965541658e-06,
      "loss": 2.9502,
      "step": 848
    },
    {
      "epoch": 0.10991358384309156,
      "grad_norm": 4.325633256350921,
      "learning_rate": 1.2496313284087408e-06,
      "loss": 4.395,
      "step": 849
    },
    {
      "epoch": 0.11004304625044503,
      "grad_norm": 6.382616472343234,
      "learning_rate": 1.249621563682214e-06,
      "loss": 5.3101,
      "step": 850
    },
    {
      "epoch": 0.11004304625044503,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3238095238095238,
      "eval_PRM F1 AUC": 0.4937139863803038,
      "eval_PRM F1 Neg": 0.3364485981308411,
      "eval_PRM NPV": 0.21428571428571427,
      "eval_PRM Precision": 0.7727272727272727,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 0.9966264367103577,
      "eval_runtime": 15.1859,
      "eval_samples_per_second": 2.897,
      "eval_steps_per_second": 0.198,
      "step": 850
    },
    {
      "epoch": 0.1101725086577985,
      "grad_norm": 2.8336601664198966,
      "learning_rate": 1.2496116713640725e-06,
      "loss": 3.7319,
      "step": 851
    },
    {
      "epoch": 0.11030197106515195,
      "grad_norm": 1.8389531897806024,
      "learning_rate": 1.2496016514563367e-06,
      "loss": 3.0261,
      "step": 852
    },
    {
      "epoch": 0.11043143347250542,
      "grad_norm": 3.6222250371246347,
      "learning_rate": 1.2495915039610534e-06,
      "loss": 4.6567,
      "step": 853
    },
    {
      "epoch": 0.11056089587985889,
      "grad_norm": 5.188574916796592,
      "learning_rate": 1.2495812288802959e-06,
      "loss": 5.4341,
      "step": 854
    },
    {
      "epoch": 0.11069035828721235,
      "grad_norm": 2.309330039567924,
      "learning_rate": 1.2495708262161623e-06,
      "loss": 3.1453,
      "step": 855
    },
    {
      "epoch": 0.11069035828721235,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3238095238095238,
      "eval_PRM F1 AUC": 0.4937139863803038,
      "eval_PRM F1 Neg": 0.3364485981308411,
      "eval_PRM NPV": 0.21428571428571427,
      "eval_PRM Precision": 0.7727272727272727,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.782608695652174,
      "eval_loss": 1.0021306276321411,
      "eval_runtime": 15.2091,
      "eval_samples_per_second": 2.893,
      "eval_steps_per_second": 0.197,
      "step": 855
    },
    {
      "epoch": 0.11081982069456582,
      "grad_norm": 6.078198113679103,
      "learning_rate": 1.249560295970778e-06,
      "loss": 5.3374,
      "step": 856
    },
    {
      "epoch": 0.11094928310191927,
      "grad_norm": 5.553502738982046,
      "learning_rate": 1.2495496381462942e-06,
      "loss": 4.9473,
      "step": 857
    },
    {
      "epoch": 0.11107874550927274,
      "grad_norm": 3.997020265294788,
      "learning_rate": 1.2495388527448875e-06,
      "loss": 4.4836,
      "step": 858
    },
    {
      "epoch": 0.11120820791662621,
      "grad_norm": 2.179289610827541,
      "learning_rate": 1.2495279397687615e-06,
      "loss": 3.5059,
      "step": 859
    },
    {
      "epoch": 0.11133767032397968,
      "grad_norm": 5.901634015894297,
      "learning_rate": 1.2495168992201454e-06,
      "loss": 5.603,
      "step": 860
    },
    {
      "epoch": 0.11133767032397968,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.3269230769230769,
      "eval_PRM F1 AUC": 0.5154531168150864,
      "eval_PRM F1 Neg": 0.35185185185185186,
      "eval_PRM NPV": 0.2235294117647059,
      "eval_PRM Precision": 0.8095238095238095,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.8260869565217391,
      "eval_loss": 0.9991121888160706,
      "eval_runtime": 15.0122,
      "eval_samples_per_second": 2.931,
      "eval_steps_per_second": 0.2,
      "step": 860
    },
    {
      "epoch": 0.11146713273133314,
      "grad_norm": 1.5009745040246112,
      "learning_rate": 1.249505731101294e-06,
      "loss": 3.178,
      "step": 861
    },
    {
      "epoch": 0.11159659513868661,
      "grad_norm": 7.412442578727905,
      "learning_rate": 1.249494435414489e-06,
      "loss": 6.4973,
      "step": 862
    },
    {
      "epoch": 0.11172605754604006,
      "grad_norm": 8.60535577512405,
      "learning_rate": 1.2494830121620375e-06,
      "loss": 7.4678,
      "step": 863
    },
    {
      "epoch": 0.11185551995339353,
      "grad_norm": 3.926298901427391,
      "learning_rate": 1.249471461346273e-06,
      "loss": 4.481,
      "step": 864
    },
    {
      "epoch": 0.111984982360747,
      "grad_norm": 3.599262243600864,
      "learning_rate": 1.249459782969555e-06,
      "loss": 3.9001,
      "step": 865
    },
    {
      "epoch": 0.111984982360747,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3106796116504854,
      "eval_PRM F1 AUC": 0.5094290204295442,
      "eval_PRM F1 Neg": 0.3486238532110092,
      "eval_PRM NPV": 0.22093023255813954,
      "eval_PRM Precision": 0.8,
      "eval_PRM Recall": 0.1927710843373494,
      "eval_PRM Specificty": 0.8260869565217391,
      "eval_loss": 1.0030184984207153,
      "eval_runtime": 14.7638,
      "eval_samples_per_second": 2.98,
      "eval_steps_per_second": 0.203,
      "step": 865
    },
    {
      "epoch": 0.11211444476810047,
      "grad_norm": 5.375324572377989,
      "learning_rate": 1.2494479770342694e-06,
      "loss": 4.5283,
      "step": 866
    },
    {
      "epoch": 0.11224390717545393,
      "grad_norm": 3.5982816431956754,
      "learning_rate": 1.2494360435428275e-06,
      "loss": 3.8792,
      "step": 867
    },
    {
      "epoch": 0.11237336958280739,
      "grad_norm": 5.852933486181772,
      "learning_rate": 1.2494239824976668e-06,
      "loss": 4.9756,
      "step": 868
    },
    {
      "epoch": 0.11250283199016085,
      "grad_norm": 3.426276644536621,
      "learning_rate": 1.249411793901251e-06,
      "loss": 3.6125,
      "step": 869
    },
    {
      "epoch": 0.11263229439751432,
      "grad_norm": 2.4398585844201466,
      "learning_rate": 1.2493994777560703e-06,
      "loss": 3.0264,
      "step": 870
    },
    {
      "epoch": 0.11263229439751432,
      "eval_PRM Accuracy": 0.3490566037735849,
      "eval_PRM F1": 0.3300970873786408,
      "eval_PRM F1 AUC": 0.537192247249869,
      "eval_PRM F1 Neg": 0.3669724770642202,
      "eval_PRM NPV": 0.23255813953488372,
      "eval_PRM Precision": 0.85,
      "eval_PRM Recall": 0.20481927710843373,
      "eval_PRM Specificty": 0.8695652173913043,
      "eval_loss": 1.0076349973678589,
      "eval_runtime": 13.7873,
      "eval_samples_per_second": 3.191,
      "eval_steps_per_second": 0.218,
      "step": 870
    },
    {
      "epoch": 0.11276175680486779,
      "grad_norm": 3.5402306806103865,
      "learning_rate": 1.2493870340646403e-06,
      "loss": 3.7861,
      "step": 871
    },
    {
      "epoch": 0.11289121921222126,
      "grad_norm": 2.712796681020074,
      "learning_rate": 1.2493744628295027e-06,
      "loss": 3.2959,
      "step": 872
    },
    {
      "epoch": 0.11302068161957472,
      "grad_norm": 5.453100818840841,
      "learning_rate": 1.2493617640532257e-06,
      "loss": 4.9917,
      "step": 873
    },
    {
      "epoch": 0.11315014402692818,
      "grad_norm": 3.8104240445796487,
      "learning_rate": 1.2493489377384029e-06,
      "loss": 4.9126,
      "step": 874
    },
    {
      "epoch": 0.11327960643428164,
      "grad_norm": 4.717319395191185,
      "learning_rate": 1.249335983887655e-06,
      "loss": 4.3391,
      "step": 875
    },
    {
      "epoch": 0.11327960643428164,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3106796116504854,
      "eval_PRM F1 AUC": 0.5094290204295442,
      "eval_PRM F1 Neg": 0.3486238532110092,
      "eval_PRM NPV": 0.22093023255813954,
      "eval_PRM Precision": 0.8,
      "eval_PRM Recall": 0.1927710843373494,
      "eval_PRM Specificty": 0.8260869565217391,
      "eval_loss": 1.0116299390792847,
      "eval_runtime": 13.9696,
      "eval_samples_per_second": 3.15,
      "eval_steps_per_second": 0.215,
      "step": 875
    },
    {
      "epoch": 0.11340906884163511,
      "grad_norm": 4.493686327669407,
      "learning_rate": 1.2493229025036275e-06,
      "loss": 4.5591,
      "step": 876
    },
    {
      "epoch": 0.11353853124898858,
      "grad_norm": 3.0504005702181916,
      "learning_rate": 1.2493096935889926e-06,
      "loss": 4.3931,
      "step": 877
    },
    {
      "epoch": 0.11366799365634205,
      "grad_norm": 2.5019561657338936,
      "learning_rate": 1.2492963571464487e-06,
      "loss": 3.7275,
      "step": 878
    },
    {
      "epoch": 0.1137974560636955,
      "grad_norm": 4.110635049966043,
      "learning_rate": 1.2492828931787199e-06,
      "loss": 4.6113,
      "step": 879
    },
    {
      "epoch": 0.11392691847104897,
      "grad_norm": 3.9883417213824335,
      "learning_rate": 1.2492693016885564e-06,
      "loss": 4.3462,
      "step": 880
    },
    {
      "epoch": 0.11392691847104897,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.3106796116504854,
      "eval_PRM F1 AUC": 0.5094290204295442,
      "eval_PRM F1 Neg": 0.3486238532110092,
      "eval_PRM NPV": 0.22093023255813954,
      "eval_PRM Precision": 0.8,
      "eval_PRM Recall": 0.1927710843373494,
      "eval_PRM Specificty": 0.8260869565217391,
      "eval_loss": 1.0116299390792847,
      "eval_runtime": 14.106,
      "eval_samples_per_second": 3.119,
      "eval_steps_per_second": 0.213,
      "step": 880
    },
    {
      "epoch": 0.11405638087840243,
      "grad_norm": 5.739692247577143,
      "learning_rate": 1.249255582678735e-06,
      "loss": 4.5376,
      "step": 881
    },
    {
      "epoch": 0.1141858432857559,
      "grad_norm": 3.062193103997161,
      "learning_rate": 1.2492417361520575e-06,
      "loss": 4.3318,
      "step": 882
    },
    {
      "epoch": 0.11431530569310937,
      "grad_norm": 3.906370106428911,
      "learning_rate": 1.2492277621113528e-06,
      "loss": 3.2939,
      "step": 883
    },
    {
      "epoch": 0.11444476810046283,
      "grad_norm": 7.935009829442984,
      "learning_rate": 1.249213660559475e-06,
      "loss": 6.3003,
      "step": 884
    },
    {
      "epoch": 0.11457423050781629,
      "grad_norm": 4.802993612758119,
      "learning_rate": 1.249199431499305e-06,
      "loss": 4.7627,
      "step": 885
    },
    {
      "epoch": 0.11457423050781629,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.3,
      "eval_PRM F1 AUC": 0.5468831849135674,
      "eval_PRM F1 Neg": 0.375,
      "eval_PRM NPV": 0.23595505617977527,
      "eval_PRM Precision": 0.8823529411764706,
      "eval_PRM Recall": 0.18072289156626506,
      "eval_PRM Specificty": 0.9130434782608695,
      "eval_loss": 1.0138493776321411,
      "eval_runtime": 15.1959,
      "eval_samples_per_second": 2.896,
      "eval_steps_per_second": 0.197,
      "step": 885
    },
    {
      "epoch": 0.11470369291516976,
      "grad_norm": 3.4235349694067083,
      "learning_rate": 1.249185074933749e-06,
      "loss": 3.2212,
      "step": 886
    },
    {
      "epoch": 0.11483315532252322,
      "grad_norm": 3.1475045189217314,
      "learning_rate": 1.2491705908657395e-06,
      "loss": 3.5601,
      "step": 887
    },
    {
      "epoch": 0.11496261772987669,
      "grad_norm": 3.3576060084887924,
      "learning_rate": 1.2491559792982358e-06,
      "loss": 3.9238,
      "step": 888
    },
    {
      "epoch": 0.11509208013723016,
      "grad_norm": 3.7841363012296423,
      "learning_rate": 1.2491412402342223e-06,
      "loss": 3.9614,
      "step": 889
    },
    {
      "epoch": 0.11522154254458361,
      "grad_norm": 3.601368336096151,
      "learning_rate": 1.2491263736767095e-06,
      "loss": 4.1152,
      "step": 890
    },
    {
      "epoch": 0.11522154254458361,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.3137254901960784,
      "eval_PRM F1 AUC": 0.5311681508643268,
      "eval_PRM F1 Neg": 0.36363636363636365,
      "eval_PRM NPV": 0.22988505747126436,
      "eval_PRM Precision": 0.8421052631578947,
      "eval_PRM Recall": 0.1927710843373494,
      "eval_PRM Specificty": 0.8695652173913043,
      "eval_loss": 1.018998622894287,
      "eval_runtime": 14.9255,
      "eval_samples_per_second": 2.948,
      "eval_steps_per_second": 0.201,
      "step": 890
    },
    {
      "epoch": 0.11535100495193708,
      "grad_norm": 3.4178557242458534,
      "learning_rate": 1.2491113796287347e-06,
      "loss": 4.3242,
      "step": 891
    },
    {
      "epoch": 0.11548046735929055,
      "grad_norm": 2.974190535857782,
      "learning_rate": 1.2490962580933602e-06,
      "loss": 3.377,
      "step": 892
    },
    {
      "epoch": 0.11560992976664401,
      "grad_norm": 4.724263966045136,
      "learning_rate": 1.2490810090736752e-06,
      "loss": 4.3235,
      "step": 893
    },
    {
      "epoch": 0.11573939217399748,
      "grad_norm": 2.242875054399141,
      "learning_rate": 1.2490656325727945e-06,
      "loss": 3.1953,
      "step": 894
    },
    {
      "epoch": 0.11586885458135095,
      "grad_norm": 2.529182875396455,
      "learning_rate": 1.2490501285938592e-06,
      "loss": 3.3425,
      "step": 895
    },
    {
      "epoch": 0.11586885458135095,
      "eval_PRM Accuracy": 0.33962264150943394,
      "eval_PRM F1": 0.3,
      "eval_PRM F1 AUC": 0.5468831849135674,
      "eval_PRM F1 Neg": 0.375,
      "eval_PRM NPV": 0.23595505617977527,
      "eval_PRM Precision": 0.8823529411764706,
      "eval_PRM Recall": 0.18072289156626506,
      "eval_PRM Specificty": 0.9130434782608695,
      "eval_loss": 1.0213068723678589,
      "eval_runtime": 14.8251,
      "eval_samples_per_second": 2.968,
      "eval_steps_per_second": 0.202,
      "step": 895
    },
    {
      "epoch": 0.1159983169887044,
      "grad_norm": 2.2319499551112747,
      "learning_rate": 1.2490344971400363e-06,
      "loss": 3.2275,
      "step": 896
    },
    {
      "epoch": 0.11612777939605787,
      "grad_norm": 2.966351334593291,
      "learning_rate": 1.2490187382145185e-06,
      "loss": 3.373,
      "step": 897
    },
    {
      "epoch": 0.11625724180341133,
      "grad_norm": 5.077339885312606,
      "learning_rate": 1.2490028518205256e-06,
      "loss": 4.2671,
      "step": 898
    },
    {
      "epoch": 0.1163867042107648,
      "grad_norm": 1.444467093764416,
      "learning_rate": 1.2489868379613017e-06,
      "loss": 2.5872,
      "step": 899
    },
    {
      "epoch": 0.11651616661811827,
      "grad_norm": 5.850737496333685,
      "learning_rate": 1.248970696640119e-06,
      "loss": 4.689,
      "step": 900
    },
    {
      "epoch": 0.11651616661811827,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.2828282828282828,
      "eval_PRM F1 AUC": 0.5408590885280251,
      "eval_PRM F1 Neg": 0.37168141592920356,
      "eval_PRM NPV": 0.23333333333333334,
      "eval_PRM Precision": 0.875,
      "eval_PRM Recall": 0.1686746987951807,
      "eval_PRM Specificty": 0.9130434782608695,
      "eval_loss": 1.0205966234207153,
      "eval_runtime": 15.7606,
      "eval_samples_per_second": 2.792,
      "eval_steps_per_second": 0.19,
      "step": 900
    },
    {
      "epoch": 0.11664562902547172,
      "grad_norm": 6.6209305728707974,
      "learning_rate": 1.248954427860274e-06,
      "loss": 5.437,
      "step": 901
    },
    {
      "epoch": 0.11677509143282519,
      "grad_norm": 5.155904054749944,
      "learning_rate": 1.2489380316250898e-06,
      "loss": 5.8501,
      "step": 902
    },
    {
      "epoch": 0.11690455384017866,
      "grad_norm": 1.8097533710532838,
      "learning_rate": 1.2489215079379164e-06,
      "loss": 2.7844,
      "step": 903
    },
    {
      "epoch": 0.11703401624753212,
      "grad_norm": 5.140810409033703,
      "learning_rate": 1.2489048568021284e-06,
      "loss": 5.583,
      "step": 904
    },
    {
      "epoch": 0.11716347865488559,
      "grad_norm": 2.708403766038811,
      "learning_rate": 1.2488880782211277e-06,
      "loss": 3.1641,
      "step": 905
    },
    {
      "epoch": 0.11716347865488559,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.2828282828282828,
      "eval_PRM F1 AUC": 0.5408590885280251,
      "eval_PRM F1 Neg": 0.37168141592920356,
      "eval_PRM NPV": 0.23333333333333334,
      "eval_PRM Precision": 0.875,
      "eval_PRM Recall": 0.1686746987951807,
      "eval_PRM Specificty": 0.9130434782608695,
      "eval_loss": 1.0230823755264282,
      "eval_runtime": 14.2319,
      "eval_samples_per_second": 3.092,
      "eval_steps_per_second": 0.211,
      "step": 905
    },
    {
      "epoch": 0.11729294106223906,
      "grad_norm": 2.9003673908208167,
      "learning_rate": 1.2488711721983413e-06,
      "loss": 4.0337,
      "step": 906
    },
    {
      "epoch": 0.11742240346959251,
      "grad_norm": 4.41386619831515,
      "learning_rate": 1.2488541387372226e-06,
      "loss": 4.2759,
      "step": 907
    },
    {
      "epoch": 0.11755186587694598,
      "grad_norm": 5.497296513095215,
      "learning_rate": 1.2488369778412513e-06,
      "loss": 5.5381,
      "step": 908
    },
    {
      "epoch": 0.11768132828429945,
      "grad_norm": 4.057718708605249,
      "learning_rate": 1.2488196895139323e-06,
      "loss": 4.2947,
      "step": 909
    },
    {
      "epoch": 0.11781079069165291,
      "grad_norm": 4.075251374539448,
      "learning_rate": 1.2488022737587978e-06,
      "loss": 4.1887,
      "step": 910
    },
    {
      "epoch": 0.11781079069165291,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.26804123711340205,
      "eval_PRM F1 AUC": 0.5565741225772656,
      "eval_PRM F1 Neg": 0.3826086956521739,
      "eval_PRM NPV": 0.2391304347826087,
      "eval_PRM Precision": 0.9285714285714286,
      "eval_PRM Recall": 0.1566265060240964,
      "eval_PRM Specificty": 0.9565217391304348,
      "eval_loss": 1.0245915651321411,
      "eval_runtime": 13.6452,
      "eval_samples_per_second": 3.225,
      "eval_steps_per_second": 0.22,
      "step": 910
    },
    {
      "epoch": 0.11794025309900638,
      "grad_norm": 4.546605874265857,
      "learning_rate": 1.248784730579405e-06,
      "loss": 4.9141,
      "step": 911
    },
    {
      "epoch": 0.11806971550635983,
      "grad_norm": 3.5002972929514824,
      "learning_rate": 1.2487670599793372e-06,
      "loss": 3.5303,
      "step": 912
    },
    {
      "epoch": 0.1181991779137133,
      "grad_norm": 4.543518740402077,
      "learning_rate": 1.2487492619622045e-06,
      "loss": 4.6104,
      "step": 913
    },
    {
      "epoch": 0.11832864032106677,
      "grad_norm": 1.5433945063954684,
      "learning_rate": 1.2487313365316419e-06,
      "loss": 2.6094,
      "step": 914
    },
    {
      "epoch": 0.11845810272842024,
      "grad_norm": 1.5873363221733159,
      "learning_rate": 1.2487132836913117e-06,
      "loss": 2.8586,
      "step": 915
    },
    {
      "epoch": 0.11845810272842024,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.25,
      "eval_PRM F1 AUC": 0.5505500261917234,
      "eval_PRM F1 Neg": 0.3793103448275862,
      "eval_PRM NPV": 0.23655913978494625,
      "eval_PRM Precision": 0.9230769230769231,
      "eval_PRM Recall": 0.14457831325301204,
      "eval_PRM Specificty": 0.9565217391304348,
      "eval_loss": 1.0308948755264282,
      "eval_runtime": 13.85,
      "eval_samples_per_second": 3.177,
      "eval_steps_per_second": 0.217,
      "step": 915
    },
    {
      "epoch": 0.1185875651357737,
      "grad_norm": 4.7418148262939885,
      "learning_rate": 1.248695103444901e-06,
      "loss": 4.1821,
      "step": 916
    },
    {
      "epoch": 0.11871702754312717,
      "grad_norm": 3.693319011165287,
      "learning_rate": 1.2486767957961237e-06,
      "loss": 3.6196,
      "step": 917
    },
    {
      "epoch": 0.11884648995048062,
      "grad_norm": 2.649385339123106,
      "learning_rate": 1.2486583607487197e-06,
      "loss": 3.5439,
      "step": 918
    },
    {
      "epoch": 0.11897595235783409,
      "grad_norm": 5.939044051126767,
      "learning_rate": 1.2486397983064543e-06,
      "loss": 5.5059,
      "step": 919
    },
    {
      "epoch": 0.11910541476518756,
      "grad_norm": 2.8394658154401795,
      "learning_rate": 1.2486211084731197e-06,
      "loss": 3.8086,
      "step": 920
    },
    {
      "epoch": 0.11910541476518756,
      "eval_PRM Accuracy": 0.3490566037735849,
      "eval_PRM F1": 0.30303030303030304,
      "eval_PRM F1 AUC": 0.56862231534835,
      "eval_PRM F1 Neg": 0.3893805309734513,
      "eval_PRM NPV": 0.24444444444444444,
      "eval_PRM Precision": 0.9375,
      "eval_PRM Recall": 0.18072289156626506,
      "eval_PRM Specificty": 0.9565217391304348,
      "eval_loss": 1.033647060394287,
      "eval_runtime": 15.5497,
      "eval_samples_per_second": 2.83,
      "eval_steps_per_second": 0.193,
      "step": 920
    },
    {
      "epoch": 0.11923487717254103,
      "grad_norm": 2.669443359973629,
      "learning_rate": 1.2486022912525333e-06,
      "loss": 3.0354,
      "step": 921
    },
    {
      "epoch": 0.1193643395798945,
      "grad_norm": 2.6859606289086204,
      "learning_rate": 1.2485833466485392e-06,
      "loss": 3.7344,
      "step": 922
    },
    {
      "epoch": 0.11949380198724795,
      "grad_norm": 2.3894788312649036,
      "learning_rate": 1.248564274665007e-06,
      "loss": 3.2959,
      "step": 923
    },
    {
      "epoch": 0.11962326439460141,
      "grad_norm": 2.079663998889526,
      "learning_rate": 1.2485450753058326e-06,
      "loss": 3.0674,
      "step": 924
    },
    {
      "epoch": 0.11975272680195488,
      "grad_norm": 6.620720008250604,
      "learning_rate": 1.248525748574938e-06,
      "loss": 5.6985,
      "step": 925
    },
    {
      "epoch": 0.11975272680195488,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.26804123711340205,
      "eval_PRM F1 AUC": 0.5565741225772656,
      "eval_PRM F1 Neg": 0.3826086956521739,
      "eval_PRM NPV": 0.2391304347826087,
      "eval_PRM Precision": 0.9285714285714286,
      "eval_PRM Recall": 0.1566265060240964,
      "eval_PRM Specificty": 0.9565217391304348,
      "eval_loss": 1.0389736890792847,
      "eval_runtime": 15.1577,
      "eval_samples_per_second": 2.903,
      "eval_steps_per_second": 0.198,
      "step": 925
    },
    {
      "epoch": 0.11988218920930835,
      "grad_norm": 4.121052505537082,
      "learning_rate": 1.2485062944762707e-06,
      "loss": 3.767,
      "step": 926
    },
    {
      "epoch": 0.12001165161666182,
      "grad_norm": 3.949851494932497,
      "learning_rate": 1.2484867130138051e-06,
      "loss": 4.7007,
      "step": 927
    },
    {
      "epoch": 0.12014111402401528,
      "grad_norm": 4.8733058805450185,
      "learning_rate": 1.2484670041915408e-06,
      "loss": 4.9292,
      "step": 928
    },
    {
      "epoch": 0.12027057643136874,
      "grad_norm": 2.755096418475927,
      "learning_rate": 1.248447168013504e-06,
      "loss": 3.5366,
      "step": 929
    },
    {
      "epoch": 0.1204000388387222,
      "grad_norm": 2.534427885267366,
      "learning_rate": 1.2484272044837462e-06,
      "loss": 3.7122,
      "step": 930
    },
    {
      "epoch": 0.1204000388387222,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.25263157894736843,
      "eval_PRM F1 AUC": 0.572289156626506,
      "eval_PRM F1 Neg": 0.39316239316239315,
      "eval_PRM NPV": 0.24468085106382978,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.14457831325301204,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0392400026321411,
      "eval_runtime": 15.12,
      "eval_samples_per_second": 2.91,
      "eval_steps_per_second": 0.198,
      "step": 930
    },
    {
      "epoch": 0.12052950124607567,
      "grad_norm": 3.350680168750641,
      "learning_rate": 1.2484071136063456e-06,
      "loss": 4.415,
      "step": 931
    },
    {
      "epoch": 0.12065896365342914,
      "grad_norm": 5.518309124632107,
      "learning_rate": 1.2483868953854061e-06,
      "loss": 4.4155,
      "step": 932
    },
    {
      "epoch": 0.1207884260607826,
      "grad_norm": 3.6445339199367734,
      "learning_rate": 1.248366549825058e-06,
      "loss": 3.9683,
      "step": 933
    },
    {
      "epoch": 0.12091788846813606,
      "grad_norm": 5.318081452162914,
      "learning_rate": 1.248346076929457e-06,
      "loss": 4.0686,
      "step": 934
    },
    {
      "epoch": 0.12104735087548953,
      "grad_norm": 7.0084529244510385,
      "learning_rate": 1.2483254767027848e-06,
      "loss": 5.2681,
      "step": 935
    },
    {
      "epoch": 0.12104735087548953,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.25,
      "eval_PRM F1 AUC": 0.5505500261917234,
      "eval_PRM F1 Neg": 0.3793103448275862,
      "eval_PRM NPV": 0.23655913978494625,
      "eval_PRM Precision": 0.9230769230769231,
      "eval_PRM Recall": 0.14457831325301204,
      "eval_PRM Specificty": 0.9565217391304348,
      "eval_loss": 1.0420809984207153,
      "eval_runtime": 14.9876,
      "eval_samples_per_second": 2.936,
      "eval_steps_per_second": 0.2,
      "step": 935
    },
    {
      "epoch": 0.121176813282843,
      "grad_norm": 8.067725993235062,
      "learning_rate": 1.2483047491492501e-06,
      "loss": 5.5825,
      "step": 936
    },
    {
      "epoch": 0.12130627569019646,
      "grad_norm": 4.398388312771275,
      "learning_rate": 1.2482838942730862e-06,
      "loss": 4.6582,
      "step": 937
    },
    {
      "epoch": 0.12143573809754993,
      "grad_norm": 8.748390665190263,
      "learning_rate": 1.248262912078554e-06,
      "loss": 5.6055,
      "step": 938
    },
    {
      "epoch": 0.1215652005049034,
      "grad_norm": 2.396884372702812,
      "learning_rate": 1.2482418025699386e-06,
      "loss": 3.0403,
      "step": 939
    },
    {
      "epoch": 0.12169466291225685,
      "grad_norm": 1.7716886889052004,
      "learning_rate": 1.2482205657515528e-06,
      "loss": 2.9629,
      "step": 940
    },
    {
      "epoch": 0.12169466291225685,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.25,
      "eval_PRM F1 AUC": 0.5505500261917234,
      "eval_PRM F1 Neg": 0.3793103448275862,
      "eval_PRM NPV": 0.23655913978494625,
      "eval_PRM Precision": 0.9230769230769231,
      "eval_PRM Recall": 0.14457831325301204,
      "eval_PRM Specificty": 0.9565217391304348,
      "eval_loss": 1.0450993776321411,
      "eval_runtime": 14.9449,
      "eval_samples_per_second": 2.944,
      "eval_steps_per_second": 0.201,
      "step": 940
    },
    {
      "epoch": 0.12182412531961032,
      "grad_norm": 4.808822983703,
      "learning_rate": 1.2481992016277341e-06,
      "loss": 4.8062,
      "step": 941
    },
    {
      "epoch": 0.12195358772696378,
      "grad_norm": 5.5255127021190775,
      "learning_rate": 1.248177710202847e-06,
      "loss": 4.7842,
      "step": 942
    },
    {
      "epoch": 0.12208305013431725,
      "grad_norm": 8.477913096089036,
      "learning_rate": 1.2481560914812812e-06,
      "loss": 7.334,
      "step": 943
    },
    {
      "epoch": 0.12221251254167072,
      "grad_norm": 5.035702792043305,
      "learning_rate": 1.248134345467453e-06,
      "loss": 4.7322,
      "step": 944
    },
    {
      "epoch": 0.12234197494902417,
      "grad_norm": 2.471288339182732,
      "learning_rate": 1.2481124721658043e-06,
      "loss": 2.9038,
      "step": 945
    },
    {
      "epoch": 0.12234197494902417,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.25,
      "eval_PRM F1 AUC": 0.5505500261917234,
      "eval_PRM F1 Neg": 0.3793103448275862,
      "eval_PRM NPV": 0.23655913978494625,
      "eval_PRM Precision": 0.9230769230769231,
      "eval_PRM Recall": 0.14457831325301204,
      "eval_PRM Specificty": 0.9565217391304348,
      "eval_loss": 1.0478515625,
      "eval_runtime": 13.4994,
      "eval_samples_per_second": 3.259,
      "eval_steps_per_second": 0.222,
      "step": 945
    },
    {
      "epoch": 0.12247143735637764,
      "grad_norm": 3.1523675503177246,
      "learning_rate": 1.2480904715808035e-06,
      "loss": 3.916,
      "step": 946
    },
    {
      "epoch": 0.1226008997637311,
      "grad_norm": 4.728132501670559,
      "learning_rate": 1.2480683437169441e-06,
      "loss": 4.5342,
      "step": 947
    },
    {
      "epoch": 0.12273036217108457,
      "grad_norm": 1.5046766742673054,
      "learning_rate": 1.2480460885787465e-06,
      "loss": 2.7151,
      "step": 948
    },
    {
      "epoch": 0.12285982457843804,
      "grad_norm": 6.40134935479295,
      "learning_rate": 1.248023706170757e-06,
      "loss": 5.7119,
      "step": 949
    },
    {
      "epoch": 0.12298928698579151,
      "grad_norm": 4.457082019511461,
      "learning_rate": 1.2480011964975472e-06,
      "loss": 4.4854,
      "step": 950
    },
    {
      "epoch": 0.12298928698579151,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.25,
      "eval_PRM F1 AUC": 0.5505500261917234,
      "eval_PRM F1 Neg": 0.3793103448275862,
      "eval_PRM NPV": 0.23655913978494625,
      "eval_PRM Precision": 0.9230769230769231,
      "eval_PRM Recall": 0.14457831325301204,
      "eval_PRM Specificty": 0.9565217391304348,
      "eval_loss": 1.0516690015792847,
      "eval_runtime": 14.6554,
      "eval_samples_per_second": 3.002,
      "eval_steps_per_second": 0.205,
      "step": 950
    },
    {
      "epoch": 0.12311874939314496,
      "grad_norm": 1.6090340656282216,
      "learning_rate": 1.2479785595637152e-06,
      "loss": 2.9792,
      "step": 951
    },
    {
      "epoch": 0.12324821180049843,
      "grad_norm": 2.9840250428661466,
      "learning_rate": 1.2479557953738854e-06,
      "loss": 3.25,
      "step": 952
    },
    {
      "epoch": 0.1233776742078519,
      "grad_norm": 5.476687802538889,
      "learning_rate": 1.2479329039327078e-06,
      "loss": 4.4148,
      "step": 953
    },
    {
      "epoch": 0.12350713661520536,
      "grad_norm": 4.3153723449268675,
      "learning_rate": 1.247909885244858e-06,
      "loss": 4.3208,
      "step": 954
    },
    {
      "epoch": 0.12363659902255883,
      "grad_norm": 3.6232463768075824,
      "learning_rate": 1.2478867393150387e-06,
      "loss": 3.6406,
      "step": 955
    },
    {
      "epoch": 0.12363659902255883,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.25263157894736843,
      "eval_PRM F1 AUC": 0.572289156626506,
      "eval_PRM F1 Neg": 0.39316239316239315,
      "eval_PRM NPV": 0.24468085106382978,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.14457831325301204,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0526455640792847,
      "eval_runtime": 13.9618,
      "eval_samples_per_second": 3.151,
      "eval_steps_per_second": 0.215,
      "step": 955
    },
    {
      "epoch": 0.12376606142991228,
      "grad_norm": 2.7932189865694106,
      "learning_rate": 1.247863466147977e-06,
      "loss": 3.1865,
      "step": 956
    },
    {
      "epoch": 0.12389552383726575,
      "grad_norm": 5.355610282953668,
      "learning_rate": 1.247840065748428e-06,
      "loss": 4.791,
      "step": 957
    },
    {
      "epoch": 0.12402498624461922,
      "grad_norm": 3.858147150203406,
      "learning_rate": 1.247816538121171e-06,
      "loss": 3.6785,
      "step": 958
    },
    {
      "epoch": 0.12415444865197268,
      "grad_norm": 2.707026376077688,
      "learning_rate": 1.247792883271012e-06,
      "loss": 3.4209,
      "step": 959
    },
    {
      "epoch": 0.12428391105932615,
      "grad_norm": 3.74142751010886,
      "learning_rate": 1.2477691012027835e-06,
      "loss": 4.124,
      "step": 960
    },
    {
      "epoch": 0.12428391105932615,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0560191869735718,
      "eval_runtime": 14.1069,
      "eval_samples_per_second": 3.119,
      "eval_steps_per_second": 0.213,
      "step": 960
    },
    {
      "epoch": 0.12441337346667962,
      "grad_norm": 5.926902047536574,
      "learning_rate": 1.247745191921343e-06,
      "loss": 5.7275,
      "step": 961
    },
    {
      "epoch": 0.12454283587403307,
      "grad_norm": 3.4307092732067392,
      "learning_rate": 1.2477211554315744e-06,
      "loss": 3.6108,
      "step": 962
    },
    {
      "epoch": 0.12467229828138654,
      "grad_norm": 2.8287339202573696,
      "learning_rate": 1.247696991738388e-06,
      "loss": 3.7573,
      "step": 963
    },
    {
      "epoch": 0.12480176068874001,
      "grad_norm": 4.414601818495286,
      "learning_rate": 1.2476727008467195e-06,
      "loss": 5.5747,
      "step": 964
    },
    {
      "epoch": 0.12493122309609347,
      "grad_norm": 6.8284215950680585,
      "learning_rate": 1.2476482827615308e-06,
      "loss": 5.2762,
      "step": 965
    },
    {
      "epoch": 0.12493122309609347,
      "eval_PRM Accuracy": 0.330188679245283,
      "eval_PRM F1": 0.25263157894736843,
      "eval_PRM F1 AUC": 0.572289156626506,
      "eval_PRM F1 Neg": 0.39316239316239315,
      "eval_PRM NPV": 0.24468085106382978,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.14457831325301204,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.05859375,
      "eval_runtime": 13.9419,
      "eval_samples_per_second": 3.156,
      "eval_steps_per_second": 0.215,
      "step": 965
    },
    {
      "epoch": 0.12506068550344693,
      "grad_norm": 2.8675498151650767,
      "learning_rate": 1.2476237374878098e-06,
      "loss": 3.4995,
      "step": 966
    },
    {
      "epoch": 0.1251901479108004,
      "grad_norm": 4.540907752630362,
      "learning_rate": 1.2475990650305705e-06,
      "loss": 4.3252,
      "step": 967
    },
    {
      "epoch": 0.12531961031815386,
      "grad_norm": 3.8781901142816855,
      "learning_rate": 1.2475742653948527e-06,
      "loss": 4.3799,
      "step": 968
    },
    {
      "epoch": 0.12544907272550734,
      "grad_norm": 2.421194687012083,
      "learning_rate": 1.2475493385857222e-06,
      "loss": 3.0205,
      "step": 969
    },
    {
      "epoch": 0.1255785351328608,
      "grad_norm": 5.3313143451088765,
      "learning_rate": 1.2475242846082706e-06,
      "loss": 4.7822,
      "step": 970
    },
    {
      "epoch": 0.1255785351328608,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0603693723678589,
      "eval_runtime": 15.3503,
      "eval_samples_per_second": 2.866,
      "eval_steps_per_second": 0.195,
      "step": 970
    },
    {
      "epoch": 0.12570799754021425,
      "grad_norm": 8.339895406926132,
      "learning_rate": 1.247499103467616e-06,
      "loss": 5.8936,
      "step": 971
    },
    {
      "epoch": 0.12583745994756773,
      "grad_norm": 5.317072011717124,
      "learning_rate": 1.247473795168902e-06,
      "loss": 4.4404,
      "step": 972
    },
    {
      "epoch": 0.12596692235492118,
      "grad_norm": 2.7815267226558134,
      "learning_rate": 1.2474483597172985e-06,
      "loss": 3.186,
      "step": 973
    },
    {
      "epoch": 0.12609638476227467,
      "grad_norm": 3.456682002237875,
      "learning_rate": 1.247422797118001e-06,
      "loss": 3.3911,
      "step": 974
    },
    {
      "epoch": 0.12622584716962812,
      "grad_norm": 7.608816983260536,
      "learning_rate": 1.2473971073762312e-06,
      "loss": 5.8828,
      "step": 975
    },
    {
      "epoch": 0.12622584716962812,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.064009189605713,
      "eval_runtime": 15.1268,
      "eval_samples_per_second": 2.909,
      "eval_steps_per_second": 0.198,
      "step": 975
    },
    {
      "epoch": 0.12635530957698157,
      "grad_norm": 3.5969502525506396,
      "learning_rate": 1.2473712904972369e-06,
      "loss": 3.7305,
      "step": 976
    },
    {
      "epoch": 0.12648477198433505,
      "grad_norm": 4.646756236458957,
      "learning_rate": 1.2473453464862914e-06,
      "loss": 4.8125,
      "step": 977
    },
    {
      "epoch": 0.1266142343916885,
      "grad_norm": 4.720806105640083,
      "learning_rate": 1.247319275348695e-06,
      "loss": 5.4941,
      "step": 978
    },
    {
      "epoch": 0.126743696799042,
      "grad_norm": 4.248075920549624,
      "learning_rate": 1.2472930770897722e-06,
      "loss": 3.8857,
      "step": 979
    },
    {
      "epoch": 0.12687315920639544,
      "grad_norm": 2.8605290332286293,
      "learning_rate": 1.2472667517148754e-06,
      "loss": 3.0388,
      "step": 980
    },
    {
      "epoch": 0.12687315920639544,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0687144994735718,
      "eval_runtime": 14.0645,
      "eval_samples_per_second": 3.128,
      "eval_steps_per_second": 0.213,
      "step": 980
    },
    {
      "epoch": 0.1270026216137489,
      "grad_norm": 2.854697893128177,
      "learning_rate": 1.2472402992293817e-06,
      "loss": 3.5352,
      "step": 981
    },
    {
      "epoch": 0.12713208402110238,
      "grad_norm": 3.126921632202384,
      "learning_rate": 1.2472137196386946e-06,
      "loss": 3.4629,
      "step": 982
    },
    {
      "epoch": 0.12726154642845583,
      "grad_norm": 3.1697296364764194,
      "learning_rate": 1.2471870129482436e-06,
      "loss": 3.7405,
      "step": 983
    },
    {
      "epoch": 0.1273910088358093,
      "grad_norm": 5.887579579012297,
      "learning_rate": 1.2471601791634843e-06,
      "loss": 4.1284,
      "step": 984
    },
    {
      "epoch": 0.12752047124316276,
      "grad_norm": 5.737257026895172,
      "learning_rate": 1.2471332182898976e-06,
      "loss": 4.3589,
      "step": 985
    },
    {
      "epoch": 0.12752047124316276,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0690696239471436,
      "eval_runtime": 14.6042,
      "eval_samples_per_second": 3.013,
      "eval_steps_per_second": 0.205,
      "step": 985
    },
    {
      "epoch": 0.12764993365051625,
      "grad_norm": 3.954361826605491,
      "learning_rate": 1.247106130332991e-06,
      "loss": 4.3667,
      "step": 986
    },
    {
      "epoch": 0.1277793960578697,
      "grad_norm": 5.238380771473893,
      "learning_rate": 1.2470789152982976e-06,
      "loss": 4.2905,
      "step": 987
    },
    {
      "epoch": 0.12790885846522315,
      "grad_norm": 3.683852590339371,
      "learning_rate": 1.247051573191377e-06,
      "loss": 4.1904,
      "step": 988
    },
    {
      "epoch": 0.12803832087257663,
      "grad_norm": 3.800600473229702,
      "learning_rate": 1.247024104017814e-06,
      "loss": 4.1104,
      "step": 989
    },
    {
      "epoch": 0.1281677832799301,
      "grad_norm": 3.8831777476402936,
      "learning_rate": 1.24699650778322e-06,
      "loss": 3.3628,
      "step": 990
    },
    {
      "epoch": 0.1281677832799301,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0739524364471436,
      "eval_runtime": 15.813,
      "eval_samples_per_second": 2.783,
      "eval_steps_per_second": 0.19,
      "step": 990
    },
    {
      "epoch": 0.12829724568728357,
      "grad_norm": 3.8936271763020236,
      "learning_rate": 1.246968784493232e-06,
      "loss": 4.4331,
      "step": 991
    },
    {
      "epoch": 0.12842670809463702,
      "grad_norm": 5.1837856890403415,
      "learning_rate": 1.2469409341535132e-06,
      "loss": 4.3799,
      "step": 992
    },
    {
      "epoch": 0.12855617050199047,
      "grad_norm": 2.5855666146103746,
      "learning_rate": 1.246912956769752e-06,
      "loss": 3.2197,
      "step": 993
    },
    {
      "epoch": 0.12868563290934396,
      "grad_norm": 3.0354135540536085,
      "learning_rate": 1.2468848523476642e-06,
      "loss": 3.606,
      "step": 994
    },
    {
      "epoch": 0.1288150953166974,
      "grad_norm": 7.425821204767459,
      "learning_rate": 1.24685662089299e-06,
      "loss": 4.8413,
      "step": 995
    },
    {
      "epoch": 0.1288150953166974,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.074751377105713,
      "eval_runtime": 13.9792,
      "eval_samples_per_second": 3.148,
      "eval_steps_per_second": 0.215,
      "step": 995
    },
    {
      "epoch": 0.1289445577240509,
      "grad_norm": 6.017254284461653,
      "learning_rate": 1.2468282624114965e-06,
      "loss": 5.498,
      "step": 996
    },
    {
      "epoch": 0.12907402013140434,
      "grad_norm": 2.397532893122352,
      "learning_rate": 1.2467997769089768e-06,
      "loss": 3.0735,
      "step": 997
    },
    {
      "epoch": 0.1292034825387578,
      "grad_norm": 3.5378089273062403,
      "learning_rate": 1.2467711643912493e-06,
      "loss": 3.6301,
      "step": 998
    },
    {
      "epoch": 0.12933294494611128,
      "grad_norm": 4.768367773201599,
      "learning_rate": 1.2467424248641585e-06,
      "loss": 4.478,
      "step": 999
    },
    {
      "epoch": 0.12946240735346473,
      "grad_norm": 5.105482652077706,
      "learning_rate": 1.2467135583335757e-06,
      "loss": 4.4482,
      "step": 1000
    },
    {
      "epoch": 0.12946240735346473,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.078657627105713,
      "eval_runtime": 14.0803,
      "eval_samples_per_second": 3.125,
      "eval_steps_per_second": 0.213,
      "step": 1000
    },
    {
      "epoch": 0.1295918697608182,
      "grad_norm": 2.8120885811909413,
      "learning_rate": 1.2466845648053966e-06,
      "loss": 3.8772,
      "step": 1001
    },
    {
      "epoch": 0.12972133216817167,
      "grad_norm": 5.756038565932446,
      "learning_rate": 1.2466554442855443e-06,
      "loss": 4.4307,
      "step": 1002
    },
    {
      "epoch": 0.12985079457552512,
      "grad_norm": 1.8124096125196645,
      "learning_rate": 1.2466261967799672e-06,
      "loss": 2.8479,
      "step": 1003
    },
    {
      "epoch": 0.1299802569828786,
      "grad_norm": 4.381773001776485,
      "learning_rate": 1.2465968222946397e-06,
      "loss": 5.085,
      "step": 1004
    },
    {
      "epoch": 0.13010971939023205,
      "grad_norm": 5.382581508115201,
      "learning_rate": 1.2465673208355619e-06,
      "loss": 4.7542,
      "step": 1005
    },
    {
      "epoch": 0.13010971939023205,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0799893140792847,
      "eval_runtime": 14.2326,
      "eval_samples_per_second": 3.092,
      "eval_steps_per_second": 0.211,
      "step": 1005
    },
    {
      "epoch": 0.13023918179758553,
      "grad_norm": 4.17922696281196,
      "learning_rate": 1.2465376924087603e-06,
      "loss": 4.1602,
      "step": 1006
    },
    {
      "epoch": 0.130368644204939,
      "grad_norm": 1.3444658394429028,
      "learning_rate": 1.246507937020287e-06,
      "loss": 2.7651,
      "step": 1007
    },
    {
      "epoch": 0.13049810661229247,
      "grad_norm": 2.557842837131199,
      "learning_rate": 1.2464780546762201e-06,
      "loss": 3.1748,
      "step": 1008
    },
    {
      "epoch": 0.13062756901964592,
      "grad_norm": 4.350603012131716,
      "learning_rate": 1.2464480453826638e-06,
      "loss": 4.8105,
      "step": 1009
    },
    {
      "epoch": 0.13075703142699938,
      "grad_norm": 2.3737512312825664,
      "learning_rate": 1.2464179091457482e-06,
      "loss": 3.3989,
      "step": 1010
    },
    {
      "epoch": 0.13075703142699938,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0841618776321411,
      "eval_runtime": 13.5446,
      "eval_samples_per_second": 3.249,
      "eval_steps_per_second": 0.221,
      "step": 1010
    },
    {
      "epoch": 0.13088649383435286,
      "grad_norm": 5.674478451377292,
      "learning_rate": 1.246387645971629e-06,
      "loss": 5.355,
      "step": 1011
    },
    {
      "epoch": 0.1310159562417063,
      "grad_norm": 4.088708377835137,
      "learning_rate": 1.246357255866488e-06,
      "loss": 3.7275,
      "step": 1012
    },
    {
      "epoch": 0.1311454186490598,
      "grad_norm": 6.34277906172523,
      "learning_rate": 1.2463267388365332e-06,
      "loss": 5.2432,
      "step": 1013
    },
    {
      "epoch": 0.13127488105641325,
      "grad_norm": 3.3532623510005397,
      "learning_rate": 1.2462960948879985e-06,
      "loss": 3.5669,
      "step": 1014
    },
    {
      "epoch": 0.1314043434637667,
      "grad_norm": 4.065983366275227,
      "learning_rate": 1.246265324027143e-06,
      "loss": 3.603,
      "step": 1015
    },
    {
      "epoch": 0.1314043434637667,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.08984375,
      "eval_runtime": 14.51,
      "eval_samples_per_second": 3.032,
      "eval_steps_per_second": 0.207,
      "step": 1015
    },
    {
      "epoch": 0.13153380587112018,
      "grad_norm": 4.871446791271936,
      "learning_rate": 1.2462344262602528e-06,
      "loss": 4.6147,
      "step": 1016
    },
    {
      "epoch": 0.13166326827847363,
      "grad_norm": 6.884465471392841,
      "learning_rate": 1.2462034015936392e-06,
      "loss": 5.7163,
      "step": 1017
    },
    {
      "epoch": 0.13179273068582711,
      "grad_norm": 3.540900098909635,
      "learning_rate": 1.2461722500336397e-06,
      "loss": 3.4534,
      "step": 1018
    },
    {
      "epoch": 0.13192219309318057,
      "grad_norm": 6.846871483011824,
      "learning_rate": 1.2461409715866172e-06,
      "loss": 5.3828,
      "step": 1019
    },
    {
      "epoch": 0.13205165550053402,
      "grad_norm": 2.226206677374046,
      "learning_rate": 1.2461095662589615e-06,
      "loss": 3.1299,
      "step": 1020
    },
    {
      "epoch": 0.13205165550053402,
      "eval_PRM Accuracy": 0.3113207547169811,
      "eval_PRM F1": 0.21505376344086022,
      "eval_PRM F1 AUC": 0.5602409638554217,
      "eval_PRM F1 Neg": 0.3865546218487395,
      "eval_PRM NPV": 0.23958333333333334,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.12048192771084337,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0929509401321411,
      "eval_runtime": 14.5103,
      "eval_samples_per_second": 3.032,
      "eval_steps_per_second": 0.207,
      "step": 1020
    },
    {
      "epoch": 0.1321811179078875,
      "grad_norm": 4.176354096852999,
      "learning_rate": 1.2460780340570877e-06,
      "loss": 4.3921,
      "step": 1021
    },
    {
      "epoch": 0.13231058031524096,
      "grad_norm": 4.958554188670066,
      "learning_rate": 1.2460463749874366e-06,
      "loss": 5.5342,
      "step": 1022
    },
    {
      "epoch": 0.13244004272259444,
      "grad_norm": 3.9563417452721175,
      "learning_rate": 1.2460145890564754e-06,
      "loss": 4.0376,
      "step": 1023
    },
    {
      "epoch": 0.1325695051299479,
      "grad_norm": 3.0904427146375215,
      "learning_rate": 1.245982676270697e-06,
      "loss": 3.6658,
      "step": 1024
    },
    {
      "epoch": 0.13269896753730134,
      "grad_norm": 4.385039892457533,
      "learning_rate": 1.24595063663662e-06,
      "loss": 4.2363,
      "step": 1025
    },
    {
      "epoch": 0.13269896753730134,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0927734375,
      "eval_runtime": 14.4587,
      "eval_samples_per_second": 3.043,
      "eval_steps_per_second": 0.207,
      "step": 1025
    },
    {
      "epoch": 0.13282842994465482,
      "grad_norm": 1.3935763226393512,
      "learning_rate": 1.2459184701607896e-06,
      "loss": 2.5908,
      "step": 1026
    },
    {
      "epoch": 0.13295789235200828,
      "grad_norm": 4.042920496148491,
      "learning_rate": 1.2458861768497763e-06,
      "loss": 3.9229,
      "step": 1027
    },
    {
      "epoch": 0.13308735475936176,
      "grad_norm": 5.486748298031092,
      "learning_rate": 1.2458537567101765e-06,
      "loss": 4.395,
      "step": 1028
    },
    {
      "epoch": 0.1332168171667152,
      "grad_norm": 6.810921713396431,
      "learning_rate": 1.2458212097486128e-06,
      "loss": 5.709,
      "step": 1029
    },
    {
      "epoch": 0.1333462795740687,
      "grad_norm": 2.9961899744639338,
      "learning_rate": 1.2457885359717334e-06,
      "loss": 3.9429,
      "step": 1030
    },
    {
      "epoch": 0.1333462795740687,
      "eval_PRM Accuracy": 0.3113207547169811,
      "eval_PRM F1": 0.21505376344086022,
      "eval_PRM F1 AUC": 0.5602409638554217,
      "eval_PRM F1 Neg": 0.3865546218487395,
      "eval_PRM NPV": 0.23958333333333334,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.12048192771084337,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.096235752105713,
      "eval_runtime": 14.9658,
      "eval_samples_per_second": 2.94,
      "eval_steps_per_second": 0.2,
      "step": 1030
    },
    {
      "epoch": 0.13347574198142215,
      "grad_norm": 3.9360839882171677,
      "learning_rate": 1.245755735386213e-06,
      "loss": 4.5273,
      "step": 1031
    },
    {
      "epoch": 0.1336052043887756,
      "grad_norm": 3.5663962125428466,
      "learning_rate": 1.2457228079987512e-06,
      "loss": 4.2832,
      "step": 1032
    },
    {
      "epoch": 0.13373466679612908,
      "grad_norm": 2.482586747992713,
      "learning_rate": 1.2456897538160745e-06,
      "loss": 2.9658,
      "step": 1033
    },
    {
      "epoch": 0.13386412920348253,
      "grad_norm": 3.8268853657919926,
      "learning_rate": 1.2456565728449349e-06,
      "loss": 3.5974,
      "step": 1034
    },
    {
      "epoch": 0.13399359161083602,
      "grad_norm": 8.482190043082205,
      "learning_rate": 1.24562326509211e-06,
      "loss": 5.9277,
      "step": 1035
    },
    {
      "epoch": 0.13399359161083602,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.0987216234207153,
      "eval_runtime": 14.1289,
      "eval_samples_per_second": 3.114,
      "eval_steps_per_second": 0.212,
      "step": 1035
    },
    {
      "epoch": 0.13412305401818947,
      "grad_norm": 2.0206646424945482,
      "learning_rate": 1.2455898305644038e-06,
      "loss": 3.0186,
      "step": 1036
    },
    {
      "epoch": 0.13425251642554292,
      "grad_norm": 5.020851343704947,
      "learning_rate": 1.2455562692686459e-06,
      "loss": 3.8423,
      "step": 1037
    },
    {
      "epoch": 0.1343819788328964,
      "grad_norm": 5.132360994585813,
      "learning_rate": 1.245522581211692e-06,
      "loss": 4.5366,
      "step": 1038
    },
    {
      "epoch": 0.13451144124024986,
      "grad_norm": 1.477783236080554,
      "learning_rate": 1.2454887664004233e-06,
      "loss": 3.0269,
      "step": 1039
    },
    {
      "epoch": 0.13464090364760334,
      "grad_norm": 2.237540431285223,
      "learning_rate": 1.2454548248417475e-06,
      "loss": 3.4336,
      "step": 1040
    },
    {
      "epoch": 0.13464090364760334,
      "eval_PRM Accuracy": 0.3113207547169811,
      "eval_PRM F1": 0.21505376344086022,
      "eval_PRM F1 AUC": 0.5602409638554217,
      "eval_PRM F1 Neg": 0.3865546218487395,
      "eval_PRM NPV": 0.23958333333333334,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.12048192771084337,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1044921875,
      "eval_runtime": 14.3873,
      "eval_samples_per_second": 3.058,
      "eval_steps_per_second": 0.209,
      "step": 1040
    },
    {
      "epoch": 0.1347703660549568,
      "grad_norm": 3.8002126953119744,
      "learning_rate": 1.2454207565425973e-06,
      "loss": 4.0537,
      "step": 1041
    },
    {
      "epoch": 0.13489982846231025,
      "grad_norm": 2.7480266287135913,
      "learning_rate": 1.2453865615099327e-06,
      "loss": 3.1523,
      "step": 1042
    },
    {
      "epoch": 0.13502929086966373,
      "grad_norm": 2.9818602767874163,
      "learning_rate": 1.245352239750738e-06,
      "loss": 3.1567,
      "step": 1043
    },
    {
      "epoch": 0.13515875327701718,
      "grad_norm": 1.7889731630388341,
      "learning_rate": 1.2453177912720244e-06,
      "loss": 3.0554,
      "step": 1044
    },
    {
      "epoch": 0.13528821568437066,
      "grad_norm": 2.137420505975883,
      "learning_rate": 1.2452832160808286e-06,
      "loss": 2.8738,
      "step": 1045
    },
    {
      "epoch": 0.13528821568437066,
      "eval_PRM Accuracy": 0.32075471698113206,
      "eval_PRM F1": 0.23404255319148937,
      "eval_PRM F1 AUC": 0.5662650602409639,
      "eval_PRM F1 Neg": 0.3898305084745763,
      "eval_PRM NPV": 0.24210526315789474,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.13253012048192772,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1065341234207153,
      "eval_runtime": 14.6277,
      "eval_samples_per_second": 3.008,
      "eval_steps_per_second": 0.205,
      "step": 1045
    },
    {
      "epoch": 0.13541767809172411,
      "grad_norm": 1.9691915406197775,
      "learning_rate": 1.2452485141842135e-06,
      "loss": 2.9021,
      "step": 1046
    },
    {
      "epoch": 0.13554714049907757,
      "grad_norm": 2.7479659357203112,
      "learning_rate": 1.2452136855892674e-06,
      "loss": 3.9976,
      "step": 1047
    },
    {
      "epoch": 0.13567660290643105,
      "grad_norm": 5.13953469653412,
      "learning_rate": 1.2451787303031051e-06,
      "loss": 4.9111,
      "step": 1048
    },
    {
      "epoch": 0.1358060653137845,
      "grad_norm": 6.077565305237441,
      "learning_rate": 1.2451436483328663e-06,
      "loss": 5.792,
      "step": 1049
    },
    {
      "epoch": 0.13593552772113798,
      "grad_norm": 5.968102615606787,
      "learning_rate": 1.2451084396857178e-06,
      "loss": 4.2515,
      "step": 1050
    },
    {
      "epoch": 0.13593552772113798,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.111328125,
      "eval_runtime": 14.4539,
      "eval_samples_per_second": 3.044,
      "eval_steps_per_second": 0.208,
      "step": 1050
    },
    {
      "epoch": 0.13606499012849144,
      "grad_norm": 4.935480110101971,
      "learning_rate": 1.2450731043688515e-06,
      "loss": 3.7729,
      "step": 1051
    },
    {
      "epoch": 0.13619445253584492,
      "grad_norm": 4.213728724481223,
      "learning_rate": 1.2450376423894853e-06,
      "loss": 3.9072,
      "step": 1052
    },
    {
      "epoch": 0.13632391494319837,
      "grad_norm": 4.412709520527811,
      "learning_rate": 1.245002053754863e-06,
      "loss": 4.5903,
      "step": 1053
    },
    {
      "epoch": 0.13645337735055182,
      "grad_norm": 4.761015015207559,
      "learning_rate": 1.2449663384722543e-06,
      "loss": 5.1465,
      "step": 1054
    },
    {
      "epoch": 0.1365828397579053,
      "grad_norm": 3.025138872548357,
      "learning_rate": 1.244930496548955e-06,
      "loss": 3.2041,
      "step": 1055
    },
    {
      "epoch": 0.1365828397579053,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1125710010528564,
      "eval_runtime": 14.0121,
      "eval_samples_per_second": 3.14,
      "eval_steps_per_second": 0.214,
      "step": 1055
    },
    {
      "epoch": 0.13671230216525876,
      "grad_norm": 4.3390895925318995,
      "learning_rate": 1.2448945279922863e-06,
      "loss": 4.3188,
      "step": 1056
    },
    {
      "epoch": 0.13684176457261224,
      "grad_norm": 4.890149310380324,
      "learning_rate": 1.2448584328095955e-06,
      "loss": 4.2554,
      "step": 1057
    },
    {
      "epoch": 0.1369712269799657,
      "grad_norm": 4.870306366542534,
      "learning_rate": 1.2448222110082557e-06,
      "loss": 4.3779,
      "step": 1058
    },
    {
      "epoch": 0.13710068938731915,
      "grad_norm": 4.237106036947216,
      "learning_rate": 1.2447858625956663e-06,
      "loss": 3.938,
      "step": 1059
    },
    {
      "epoch": 0.13723015179467263,
      "grad_norm": 2.137305045729939,
      "learning_rate": 1.244749387579252e-06,
      "loss": 3.2622,
      "step": 1060
    },
    {
      "epoch": 0.13723015179467263,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1148792505264282,
      "eval_runtime": 13.4891,
      "eval_samples_per_second": 3.262,
      "eval_steps_per_second": 0.222,
      "step": 1060
    },
    {
      "epoch": 0.13735961420202608,
      "grad_norm": 6.874154844975598,
      "learning_rate": 1.2447127859664633e-06,
      "loss": 5.7681,
      "step": 1061
    },
    {
      "epoch": 0.13748907660937956,
      "grad_norm": 5.518812046908119,
      "learning_rate": 1.2446760577647774e-06,
      "loss": 4.4961,
      "step": 1062
    },
    {
      "epoch": 0.13761853901673302,
      "grad_norm": 5.635553577183266,
      "learning_rate": 1.2446392029816962e-06,
      "loss": 4.4473,
      "step": 1063
    },
    {
      "epoch": 0.13774800142408647,
      "grad_norm": 5.895651136352229,
      "learning_rate": 1.2446022216247482e-06,
      "loss": 5.8628,
      "step": 1064
    },
    {
      "epoch": 0.13787746383143995,
      "grad_norm": 9.150235423220652,
      "learning_rate": 1.244565113701488e-06,
      "loss": 6.271,
      "step": 1065
    },
    {
      "epoch": 0.13787746383143995,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1170099973678589,
      "eval_runtime": 14.3522,
      "eval_samples_per_second": 3.066,
      "eval_steps_per_second": 0.209,
      "step": 1065
    },
    {
      "epoch": 0.1380069262387934,
      "grad_norm": 2.3926213237978335,
      "learning_rate": 1.2445278792194952e-06,
      "loss": 3.1616,
      "step": 1066
    },
    {
      "epoch": 0.13813638864614688,
      "grad_norm": 3.0041194779372837,
      "learning_rate": 1.2444905181863756e-06,
      "loss": 3.6079,
      "step": 1067
    },
    {
      "epoch": 0.13826585105350034,
      "grad_norm": 7.102381942612581,
      "learning_rate": 1.2444530306097612e-06,
      "loss": 5.105,
      "step": 1068
    },
    {
      "epoch": 0.1383953134608538,
      "grad_norm": 3.051103199781432,
      "learning_rate": 1.2444154164973097e-06,
      "loss": 3.9746,
      "step": 1069
    },
    {
      "epoch": 0.13852477586820727,
      "grad_norm": 4.155816663827047,
      "learning_rate": 1.2443776758567045e-06,
      "loss": 4.3833,
      "step": 1070
    },
    {
      "epoch": 0.13852477586820727,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1204723119735718,
      "eval_runtime": 16.336,
      "eval_samples_per_second": 2.693,
      "eval_steps_per_second": 0.184,
      "step": 1070
    },
    {
      "epoch": 0.13865423827556073,
      "grad_norm": 2.909044633792644,
      "learning_rate": 1.2443398086956545e-06,
      "loss": 3.7295,
      "step": 1071
    },
    {
      "epoch": 0.1387837006829142,
      "grad_norm": 6.810916498117859,
      "learning_rate": 1.2443018150218954e-06,
      "loss": 5.7261,
      "step": 1072
    },
    {
      "epoch": 0.13891316309026766,
      "grad_norm": 2.5246667893864694,
      "learning_rate": 1.2442636948431877e-06,
      "loss": 2.7729,
      "step": 1073
    },
    {
      "epoch": 0.13904262549762114,
      "grad_norm": 6.535081404466834,
      "learning_rate": 1.2442254481673185e-06,
      "loss": 5.9629,
      "step": 1074
    },
    {
      "epoch": 0.1391720879049746,
      "grad_norm": 4.1115567819025065,
      "learning_rate": 1.2441870750021003e-06,
      "loss": 4.0261,
      "step": 1075
    },
    {
      "epoch": 0.1391720879049746,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1260653734207153,
      "eval_runtime": 15.2691,
      "eval_samples_per_second": 2.882,
      "eval_steps_per_second": 0.196,
      "step": 1075
    },
    {
      "epoch": 0.13930155031232805,
      "grad_norm": 6.11749785808605,
      "learning_rate": 1.244148575355372e-06,
      "loss": 5.7808,
      "step": 1076
    },
    {
      "epoch": 0.13943101271968153,
      "grad_norm": 3.844043882823037,
      "learning_rate": 1.2441099492349973e-06,
      "loss": 4.0398,
      "step": 1077
    },
    {
      "epoch": 0.13956047512703498,
      "grad_norm": 3.999562530348418,
      "learning_rate": 1.2440711966488666e-06,
      "loss": 4.2388,
      "step": 1078
    },
    {
      "epoch": 0.13968993753438846,
      "grad_norm": 1.525558009055224,
      "learning_rate": 1.2440323176048964e-06,
      "loss": 2.6057,
      "step": 1079
    },
    {
      "epoch": 0.13981939994174192,
      "grad_norm": 3.6527491474817726,
      "learning_rate": 1.2439933121110278e-06,
      "loss": 3.2852,
      "step": 1080
    },
    {
      "epoch": 0.13981939994174192,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1299716234207153,
      "eval_runtime": 15.1106,
      "eval_samples_per_second": 2.912,
      "eval_steps_per_second": 0.199,
      "step": 1080
    },
    {
      "epoch": 0.13994886234909537,
      "grad_norm": 2.7800560748337824,
      "learning_rate": 1.2439541801752288e-06,
      "loss": 3.6885,
      "step": 1081
    },
    {
      "epoch": 0.14007832475644885,
      "grad_norm": 5.713285844726542,
      "learning_rate": 1.2439149218054928e-06,
      "loss": 5.0088,
      "step": 1082
    },
    {
      "epoch": 0.1402077871638023,
      "grad_norm": 4.513255233030094,
      "learning_rate": 1.2438755370098393e-06,
      "loss": 3.6943,
      "step": 1083
    },
    {
      "epoch": 0.1403372495711558,
      "grad_norm": 3.9719742935338926,
      "learning_rate": 1.2438360257963134e-06,
      "loss": 3.8955,
      "step": 1084
    },
    {
      "epoch": 0.14046671197850924,
      "grad_norm": 4.401220468575003,
      "learning_rate": 1.2437963881729858e-06,
      "loss": 4.0269,
      "step": 1085
    },
    {
      "epoch": 0.14046671197850924,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1299716234207153,
      "eval_runtime": 15.277,
      "eval_samples_per_second": 2.88,
      "eval_steps_per_second": 0.196,
      "step": 1085
    },
    {
      "epoch": 0.1405961743858627,
      "grad_norm": 2.7065091243488153,
      "learning_rate": 1.2437566241479536e-06,
      "loss": 3.5391,
      "step": 1086
    },
    {
      "epoch": 0.14072563679321617,
      "grad_norm": 5.291122797587988,
      "learning_rate": 1.2437167337293393e-06,
      "loss": 4.2891,
      "step": 1087
    },
    {
      "epoch": 0.14085509920056963,
      "grad_norm": 4.4398178392708285,
      "learning_rate": 1.2436767169252915e-06,
      "loss": 4.2461,
      "step": 1088
    },
    {
      "epoch": 0.1409845616079231,
      "grad_norm": 4.159299303691462,
      "learning_rate": 1.243636573743984e-06,
      "loss": 4.019,
      "step": 1089
    },
    {
      "epoch": 0.14111402401527656,
      "grad_norm": 5.255831610541715,
      "learning_rate": 1.2435963041936173e-06,
      "loss": 3.8638,
      "step": 1090
    },
    {
      "epoch": 0.14111402401527656,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1344993114471436,
      "eval_runtime": 20.5228,
      "eval_samples_per_second": 2.144,
      "eval_steps_per_second": 0.146,
      "step": 1090
    },
    {
      "epoch": 0.14124348642263002,
      "grad_norm": 4.271540775688631,
      "learning_rate": 1.2435559082824169e-06,
      "loss": 3.6372,
      "step": 1091
    },
    {
      "epoch": 0.1413729488299835,
      "grad_norm": 3.4427712967584525,
      "learning_rate": 1.2435153860186347e-06,
      "loss": 3.7021,
      "step": 1092
    },
    {
      "epoch": 0.14150241123733695,
      "grad_norm": 5.287523391353101,
      "learning_rate": 1.2434747374105482e-06,
      "loss": 3.95,
      "step": 1093
    },
    {
      "epoch": 0.14163187364469043,
      "grad_norm": 4.709416318166009,
      "learning_rate": 1.2434339624664608e-06,
      "loss": 4.4761,
      "step": 1094
    },
    {
      "epoch": 0.14176133605204388,
      "grad_norm": 3.3367113935601274,
      "learning_rate": 1.2433930611947015e-06,
      "loss": 4.1387,
      "step": 1095
    },
    {
      "epoch": 0.14176133605204388,
      "eval_PRM Accuracy": 0.29245283018867924,
      "eval_PRM F1": 0.17582417582417584,
      "eval_PRM F1 AUC": 0.5481927710843374,
      "eval_PRM F1 Neg": 0.38016528925619836,
      "eval_PRM NPV": 0.23469387755102042,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.0963855421686747,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1363636255264282,
      "eval_runtime": 14.3226,
      "eval_samples_per_second": 3.072,
      "eval_steps_per_second": 0.209,
      "step": 1095
    },
    {
      "epoch": 0.14189079845939737,
      "grad_norm": 2.6404084900379385,
      "learning_rate": 1.243352033603625e-06,
      "loss": 3.2983,
      "step": 1096
    },
    {
      "epoch": 0.14202026086675082,
      "grad_norm": 3.192939910029371,
      "learning_rate": 1.2433108797016124e-06,
      "loss": 3.3809,
      "step": 1097
    },
    {
      "epoch": 0.14214972327410427,
      "grad_norm": 3.7648727238369535,
      "learning_rate": 1.2432695994970698e-06,
      "loss": 3.9678,
      "step": 1098
    },
    {
      "epoch": 0.14227918568145775,
      "grad_norm": 6.6314968524971585,
      "learning_rate": 1.24322819299843e-06,
      "loss": 5.1104,
      "step": 1099
    },
    {
      "epoch": 0.1424086480888112,
      "grad_norm": 4.524153810847838,
      "learning_rate": 1.2431866602141506e-06,
      "loss": 4.5664,
      "step": 1100
    },
    {
      "epoch": 0.1424086480888112,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1424893140792847,
      "eval_runtime": 16.1796,
      "eval_samples_per_second": 2.719,
      "eval_steps_per_second": 0.185,
      "step": 1100
    },
    {
      "epoch": 0.1425381104961647,
      "grad_norm": 2.771733681929019,
      "learning_rate": 1.243145001152716e-06,
      "loss": 3.4568,
      "step": 1101
    },
    {
      "epoch": 0.14266757290351814,
      "grad_norm": 1.8409856025083404,
      "learning_rate": 1.2431032158226356e-06,
      "loss": 2.9897,
      "step": 1102
    },
    {
      "epoch": 0.1427970353108716,
      "grad_norm": 6.330061608212643,
      "learning_rate": 1.243061304232445e-06,
      "loss": 5.5767,
      "step": 1103
    },
    {
      "epoch": 0.14292649771822508,
      "grad_norm": 5.633492517164207,
      "learning_rate": 1.2430192663907054e-06,
      "loss": 4.2026,
      "step": 1104
    },
    {
      "epoch": 0.14305596012557853,
      "grad_norm": 5.2708392710741885,
      "learning_rate": 1.2429771023060042e-06,
      "loss": 4.5283,
      "step": 1105
    },
    {
      "epoch": 0.14305596012557853,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1439098119735718,
      "eval_runtime": 14.3628,
      "eval_samples_per_second": 3.063,
      "eval_steps_per_second": 0.209,
      "step": 1105
    },
    {
      "epoch": 0.143185422532932,
      "grad_norm": 2.700956332923536,
      "learning_rate": 1.2429348119869536e-06,
      "loss": 3.0215,
      "step": 1106
    },
    {
      "epoch": 0.14331488494028546,
      "grad_norm": 4.310952952859002,
      "learning_rate": 1.2428923954421932e-06,
      "loss": 3.9307,
      "step": 1107
    },
    {
      "epoch": 0.14344434734763892,
      "grad_norm": 3.311799919455459,
      "learning_rate": 1.2428498526803866e-06,
      "loss": 3.7446,
      "step": 1108
    },
    {
      "epoch": 0.1435738097549924,
      "grad_norm": 8.659301414845961,
      "learning_rate": 1.2428071837102243e-06,
      "loss": 5.3613,
      "step": 1109
    },
    {
      "epoch": 0.14370327216234585,
      "grad_norm": 3.2744913060764107,
      "learning_rate": 1.2427643885404225e-06,
      "loss": 3.5542,
      "step": 1110
    },
    {
      "epoch": 0.14370327216234585,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1500355005264282,
      "eval_runtime": 15.3362,
      "eval_samples_per_second": 2.869,
      "eval_steps_per_second": 0.196,
      "step": 1110
    },
    {
      "epoch": 0.14383273456969933,
      "grad_norm": 5.068685889283908,
      "learning_rate": 1.2427214671797228e-06,
      "loss": 4.2529,
      "step": 1111
    },
    {
      "epoch": 0.1439621969770528,
      "grad_norm": 3.4333095693992433,
      "learning_rate": 1.2426784196368927e-06,
      "loss": 3.9536,
      "step": 1112
    },
    {
      "epoch": 0.14409165938440624,
      "grad_norm": 3.9324156577274665,
      "learning_rate": 1.242635245920726e-06,
      "loss": 3.8291,
      "step": 1113
    },
    {
      "epoch": 0.14422112179175972,
      "grad_norm": 2.5062848368018926,
      "learning_rate": 1.242591946040041e-06,
      "loss": 3.4534,
      "step": 1114
    },
    {
      "epoch": 0.14435058419911317,
      "grad_norm": 2.33993888046791,
      "learning_rate": 1.242548520003683e-06,
      "loss": 3.0461,
      "step": 1115
    },
    {
      "epoch": 0.14435058419911317,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.150923252105713,
      "eval_runtime": 13.6039,
      "eval_samples_per_second": 3.234,
      "eval_steps_per_second": 0.221,
      "step": 1115
    },
    {
      "epoch": 0.14448004660646666,
      "grad_norm": 5.0594952402452495,
      "learning_rate": 1.2425049678205228e-06,
      "loss": 4.5825,
      "step": 1116
    },
    {
      "epoch": 0.1446095090138201,
      "grad_norm": 5.093242852258919,
      "learning_rate": 1.2424612894994566e-06,
      "loss": 4.644,
      "step": 1117
    },
    {
      "epoch": 0.1447389714211736,
      "grad_norm": 5.840839193932989,
      "learning_rate": 1.2424174850494068e-06,
      "loss": 5.0425,
      "step": 1118
    },
    {
      "epoch": 0.14486843382852704,
      "grad_norm": 6.281788255156286,
      "learning_rate": 1.2423735544793209e-06,
      "loss": 5.2183,
      "step": 1119
    },
    {
      "epoch": 0.1449978962358805,
      "grad_norm": 2.1010621793288986,
      "learning_rate": 1.2423294977981732e-06,
      "loss": 3.2271,
      "step": 1120
    },
    {
      "epoch": 0.1449978962358805,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.154296875,
      "eval_runtime": 14.9627,
      "eval_samples_per_second": 2.941,
      "eval_steps_per_second": 0.2,
      "step": 1120
    },
    {
      "epoch": 0.14512735864323398,
      "grad_norm": 3.7053175155266773,
      "learning_rate": 1.2422853150149628e-06,
      "loss": 3.752,
      "step": 1121
    },
    {
      "epoch": 0.14525682105058743,
      "grad_norm": 2.5972642614167105,
      "learning_rate": 1.2422410061387148e-06,
      "loss": 2.7944,
      "step": 1122
    },
    {
      "epoch": 0.1453862834579409,
      "grad_norm": 3.436084182132075,
      "learning_rate": 1.2421965711784805e-06,
      "loss": 4.0332,
      "step": 1123
    },
    {
      "epoch": 0.14551574586529437,
      "grad_norm": 2.643748949239247,
      "learning_rate": 1.2421520101433367e-06,
      "loss": 3.2324,
      "step": 1124
    },
    {
      "epoch": 0.14564520827264782,
      "grad_norm": 5.8756578305350935,
      "learning_rate": 1.2421073230423854e-06,
      "loss": 5.3511,
      "step": 1125
    },
    {
      "epoch": 0.14564520827264782,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1598899364471436,
      "eval_runtime": 14.9721,
      "eval_samples_per_second": 2.939,
      "eval_steps_per_second": 0.2,
      "step": 1125
    },
    {
      "epoch": 0.1457746706800013,
      "grad_norm": 4.300233658047289,
      "learning_rate": 1.2420625098847555e-06,
      "loss": 4.0046,
      "step": 1126
    },
    {
      "epoch": 0.14590413308735475,
      "grad_norm": 3.0812323358378713,
      "learning_rate": 1.2420175706796005e-06,
      "loss": 3.9351,
      "step": 1127
    },
    {
      "epoch": 0.14603359549470823,
      "grad_norm": 3.290230686059332,
      "learning_rate": 1.2419725054361004e-06,
      "loss": 4.2256,
      "step": 1128
    },
    {
      "epoch": 0.1461630579020617,
      "grad_norm": 3.01406253574201,
      "learning_rate": 1.2419273141634607e-06,
      "loss": 3.865,
      "step": 1129
    },
    {
      "epoch": 0.14629252030941514,
      "grad_norm": 4.141674327338006,
      "learning_rate": 1.2418819968709123e-06,
      "loss": 3.7329,
      "step": 1130
    },
    {
      "epoch": 0.14629252030941514,
      "eval_PRM Accuracy": 0.29245283018867924,
      "eval_PRM F1": 0.17582417582417584,
      "eval_PRM F1 AUC": 0.5481927710843374,
      "eval_PRM F1 Neg": 0.38016528925619836,
      "eval_PRM NPV": 0.23469387755102042,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.0963855421686747,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1622425317764282,
      "eval_runtime": 15.0433,
      "eval_samples_per_second": 2.925,
      "eval_steps_per_second": 0.199,
      "step": 1130
    },
    {
      "epoch": 0.14642198271676862,
      "grad_norm": 4.4611758828057555,
      "learning_rate": 1.2418365535677127e-06,
      "loss": 4.4844,
      "step": 1131
    },
    {
      "epoch": 0.14655144512412208,
      "grad_norm": 3.880257294864339,
      "learning_rate": 1.241790984263144e-06,
      "loss": 3.6812,
      "step": 1132
    },
    {
      "epoch": 0.14668090753147556,
      "grad_norm": 4.197403417992651,
      "learning_rate": 1.2417452889665153e-06,
      "loss": 3.4883,
      "step": 1133
    },
    {
      "epoch": 0.146810369938829,
      "grad_norm": 7.548336600024495,
      "learning_rate": 1.2416994676871601e-06,
      "loss": 5.3296,
      "step": 1134
    },
    {
      "epoch": 0.14693983234618246,
      "grad_norm": 5.314070244950245,
      "learning_rate": 1.241653520434439e-06,
      "loss": 4.2246,
      "step": 1135
    },
    {
      "epoch": 0.14693983234618246,
      "eval_PRM Accuracy": 0.29245283018867924,
      "eval_PRM F1": 0.17582417582417584,
      "eval_PRM F1 AUC": 0.5481927710843374,
      "eval_PRM F1 Neg": 0.38016528925619836,
      "eval_PRM NPV": 0.23469387755102042,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.0963855421686747,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.165482997894287,
      "eval_runtime": 14.2005,
      "eval_samples_per_second": 3.098,
      "eval_steps_per_second": 0.211,
      "step": 1135
    },
    {
      "epoch": 0.14706929475353595,
      "grad_norm": 4.010158133139364,
      "learning_rate": 1.2416074472177371e-06,
      "loss": 3.9932,
      "step": 1136
    },
    {
      "epoch": 0.1471987571608894,
      "grad_norm": 3.1664116446901214,
      "learning_rate": 1.2415612480464663e-06,
      "loss": 3.7764,
      "step": 1137
    },
    {
      "epoch": 0.14732821956824288,
      "grad_norm": 3.7452098581152593,
      "learning_rate": 1.2415149229300632e-06,
      "loss": 3.7412,
      "step": 1138
    },
    {
      "epoch": 0.14745768197559633,
      "grad_norm": 4.178176679330257,
      "learning_rate": 1.2414684718779913e-06,
      "loss": 4.8052,
      "step": 1139
    },
    {
      "epoch": 0.14758714438294981,
      "grad_norm": 3.736484115514027,
      "learning_rate": 1.2414218948997384e-06,
      "loss": 4.0894,
      "step": 1140
    },
    {
      "epoch": 0.14758714438294981,
      "eval_PRM Accuracy": 0.3018867924528302,
      "eval_PRM F1": 0.1956521739130435,
      "eval_PRM F1 AUC": 0.5542168674698795,
      "eval_PRM F1 Neg": 0.38333333333333336,
      "eval_PRM NPV": 0.23711340206185566,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.10843373493975904,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1712535619735718,
      "eval_runtime": 15.3806,
      "eval_samples_per_second": 2.861,
      "eval_steps_per_second": 0.195,
      "step": 1140
    },
    {
      "epoch": 0.14771660679030327,
      "grad_norm": 5.545612288497347,
      "learning_rate": 1.2413751920048193e-06,
      "loss": 4.3167,
      "step": 1141
    },
    {
      "epoch": 0.14784606919765672,
      "grad_norm": 5.885094342266873,
      "learning_rate": 1.241328363202774e-06,
      "loss": 5.2651,
      "step": 1142
    },
    {
      "epoch": 0.1479755316050102,
      "grad_norm": 9.48747168940096,
      "learning_rate": 1.2412814085031678e-06,
      "loss": 5.1099,
      "step": 1143
    },
    {
      "epoch": 0.14810499401236366,
      "grad_norm": 2.7175101031015734,
      "learning_rate": 1.2412343279155926e-06,
      "loss": 3.8564,
      "step": 1144
    },
    {
      "epoch": 0.14823445641971714,
      "grad_norm": 6.246399926253007,
      "learning_rate": 1.2411871214496655e-06,
      "loss": 5.4409,
      "step": 1145
    },
    {
      "epoch": 0.14823445641971714,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1707209348678589,
      "eval_runtime": 15.0676,
      "eval_samples_per_second": 2.92,
      "eval_steps_per_second": 0.199,
      "step": 1145
    },
    {
      "epoch": 0.1483639188270706,
      "grad_norm": 6.462318803632735,
      "learning_rate": 1.2411397891150293e-06,
      "loss": 4.7729,
      "step": 1146
    },
    {
      "epoch": 0.14849338123442404,
      "grad_norm": 3.3547904268249154,
      "learning_rate": 1.2410923309213525e-06,
      "loss": 3.3516,
      "step": 1147
    },
    {
      "epoch": 0.14862284364177752,
      "grad_norm": 6.0018818990826865,
      "learning_rate": 1.2410447468783295e-06,
      "loss": 4.3984,
      "step": 1148
    },
    {
      "epoch": 0.14875230604913098,
      "grad_norm": 3.098226401723943,
      "learning_rate": 1.2409970369956804e-06,
      "loss": 3.4839,
      "step": 1149
    },
    {
      "epoch": 0.14888176845648446,
      "grad_norm": 6.805595417611291,
      "learning_rate": 1.2409492012831505e-06,
      "loss": 5.1567,
      "step": 1150
    },
    {
      "epoch": 0.14888176845648446,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.17431640625,
      "eval_runtime": 15.4352,
      "eval_samples_per_second": 2.851,
      "eval_steps_per_second": 0.194,
      "step": 1150
    },
    {
      "epoch": 0.1490112308638379,
      "grad_norm": 3.0085780082282,
      "learning_rate": 1.2409012397505118e-06,
      "loss": 3.1726,
      "step": 1151
    },
    {
      "epoch": 0.14914069327119137,
      "grad_norm": 4.219214951166914,
      "learning_rate": 1.2408531524075613e-06,
      "loss": 3.6909,
      "step": 1152
    },
    {
      "epoch": 0.14927015567854485,
      "grad_norm": 3.131571945617612,
      "learning_rate": 1.2408049392641215e-06,
      "loss": 3.3364,
      "step": 1153
    },
    {
      "epoch": 0.1493996180858983,
      "grad_norm": 4.558008815484859,
      "learning_rate": 1.2407566003300408e-06,
      "loss": 3.6157,
      "step": 1154
    },
    {
      "epoch": 0.14952908049325178,
      "grad_norm": 4.057693804613789,
      "learning_rate": 1.2407081356151937e-06,
      "loss": 4.2231,
      "step": 1155
    },
    {
      "epoch": 0.14952908049325178,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1780894994735718,
      "eval_runtime": 14.8844,
      "eval_samples_per_second": 2.956,
      "eval_steps_per_second": 0.202,
      "step": 1155
    },
    {
      "epoch": 0.14965854290060523,
      "grad_norm": 9.395021903942308,
      "learning_rate": 1.2406595451294804e-06,
      "loss": 5.3145,
      "step": 1156
    },
    {
      "epoch": 0.1497880053079587,
      "grad_norm": 4.739429747767751,
      "learning_rate": 1.240610828882826e-06,
      "loss": 4.2935,
      "step": 1157
    },
    {
      "epoch": 0.14991746771531217,
      "grad_norm": 6.013953177550472,
      "learning_rate": 1.2405619868851818e-06,
      "loss": 4.6714,
      "step": 1158
    },
    {
      "epoch": 0.15004693012266562,
      "grad_norm": 3.3936756358658933,
      "learning_rate": 1.240513019146525e-06,
      "loss": 4.0088,
      "step": 1159
    },
    {
      "epoch": 0.1501763925300191,
      "grad_norm": 4.790235402221722,
      "learning_rate": 1.2404639256768582e-06,
      "loss": 4.2891,
      "step": 1160
    },
    {
      "epoch": 0.1501763925300191,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.182661533355713,
      "eval_runtime": 15.9452,
      "eval_samples_per_second": 2.759,
      "eval_steps_per_second": 0.188,
      "step": 1160
    },
    {
      "epoch": 0.15030585493737256,
      "grad_norm": 2.7456538713104166,
      "learning_rate": 1.2404147064862096e-06,
      "loss": 3.2068,
      "step": 1161
    },
    {
      "epoch": 0.15043531734472604,
      "grad_norm": 2.5993777167552086,
      "learning_rate": 1.2403653615846334e-06,
      "loss": 3.2368,
      "step": 1162
    },
    {
      "epoch": 0.1505647797520795,
      "grad_norm": 2.075738190996234,
      "learning_rate": 1.240315890982209e-06,
      "loss": 2.9714,
      "step": 1163
    },
    {
      "epoch": 0.15069424215943295,
      "grad_norm": 3.7640632118845887,
      "learning_rate": 1.2402662946890424e-06,
      "loss": 4.0061,
      "step": 1164
    },
    {
      "epoch": 0.15082370456678643,
      "grad_norm": 4.867948556551821,
      "learning_rate": 1.2402165727152638e-06,
      "loss": 4.1699,
      "step": 1165
    },
    {
      "epoch": 0.15082370456678643,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1837269067764282,
      "eval_runtime": 14.2197,
      "eval_samples_per_second": 3.094,
      "eval_steps_per_second": 0.211,
      "step": 1165
    },
    {
      "epoch": 0.15095316697413988,
      "grad_norm": 7.848922824337514,
      "learning_rate": 1.2401667250710306e-06,
      "loss": 5.6499,
      "step": 1166
    },
    {
      "epoch": 0.15108262938149336,
      "grad_norm": 1.97181000796218,
      "learning_rate": 1.2401167517665249e-06,
      "loss": 3.3882,
      "step": 1167
    },
    {
      "epoch": 0.15121209178884681,
      "grad_norm": 6.725735942527819,
      "learning_rate": 1.240066652811955e-06,
      "loss": 5.21,
      "step": 1168
    },
    {
      "epoch": 0.15134155419620027,
      "grad_norm": 3.911916910319721,
      "learning_rate": 1.2400164282175541e-06,
      "loss": 3.739,
      "step": 1169
    },
    {
      "epoch": 0.15147101660355375,
      "grad_norm": 5.603532945962072,
      "learning_rate": 1.239966077993582e-06,
      "loss": 4.5073,
      "step": 1170
    },
    {
      "epoch": 0.15147101660355375,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1931818723678589,
      "eval_runtime": 15.3289,
      "eval_samples_per_second": 2.87,
      "eval_steps_per_second": 0.196,
      "step": 1170
    },
    {
      "epoch": 0.1516004790109072,
      "grad_norm": 3.8454810196359834,
      "learning_rate": 1.2399156021503236e-06,
      "loss": 3.4502,
      "step": 1171
    },
    {
      "epoch": 0.15172994141826068,
      "grad_norm": 4.184152439986204,
      "learning_rate": 1.23986500069809e-06,
      "loss": 4.5081,
      "step": 1172
    },
    {
      "epoch": 0.15185940382561414,
      "grad_norm": 5.974191003909128,
      "learning_rate": 1.2398142736472168e-06,
      "loss": 5.334,
      "step": 1173
    },
    {
      "epoch": 0.1519888662329676,
      "grad_norm": 1.7684190414875718,
      "learning_rate": 1.2397634210080669e-06,
      "loss": 3.0549,
      "step": 1174
    },
    {
      "epoch": 0.15211832864032107,
      "grad_norm": 3.354370515992573,
      "learning_rate": 1.2397124427910275e-06,
      "loss": 3.8457,
      "step": 1175
    },
    {
      "epoch": 0.15211832864032107,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1973544359207153,
      "eval_runtime": 14.2302,
      "eval_samples_per_second": 3.092,
      "eval_steps_per_second": 0.211,
      "step": 1175
    },
    {
      "epoch": 0.15224779104767452,
      "grad_norm": 9.703247303413816,
      "learning_rate": 1.2396613390065118e-06,
      "loss": 5.6182,
      "step": 1176
    },
    {
      "epoch": 0.152377253455028,
      "grad_norm": 4.2978327684094175,
      "learning_rate": 1.239610109664959e-06,
      "loss": 3.9922,
      "step": 1177
    },
    {
      "epoch": 0.15250671586238146,
      "grad_norm": 3.3861931761892725,
      "learning_rate": 1.2395587547768339e-06,
      "loss": 3.3247,
      "step": 1178
    },
    {
      "epoch": 0.1526361782697349,
      "grad_norm": 2.194702128769876,
      "learning_rate": 1.2395072743526267e-06,
      "loss": 3.1234,
      "step": 1179
    },
    {
      "epoch": 0.1527656406770884,
      "grad_norm": 3.6377293597590983,
      "learning_rate": 1.239455668402853e-06,
      "loss": 3.77,
      "step": 1180
    },
    {
      "epoch": 0.1527656406770884,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.1990412473678589,
      "eval_runtime": 13.9685,
      "eval_samples_per_second": 3.15,
      "eval_steps_per_second": 0.215,
      "step": 1180
    },
    {
      "epoch": 0.15289510308444185,
      "grad_norm": 2.371487956136997,
      "learning_rate": 1.2394039369380547e-06,
      "loss": 3.1914,
      "step": 1181
    },
    {
      "epoch": 0.15302456549179533,
      "grad_norm": 2.1579526763975125,
      "learning_rate": 1.2393520799687989e-06,
      "loss": 3.189,
      "step": 1182
    },
    {
      "epoch": 0.15315402789914878,
      "grad_norm": 5.358489453115939,
      "learning_rate": 1.2393000975056782e-06,
      "loss": 4.2988,
      "step": 1183
    },
    {
      "epoch": 0.15328349030650226,
      "grad_norm": 6.584407517694455,
      "learning_rate": 1.2392479895593116e-06,
      "loss": 4.229,
      "step": 1184
    },
    {
      "epoch": 0.15341295271385572,
      "grad_norm": 8.309404189634096,
      "learning_rate": 1.2391957561403425e-06,
      "loss": 6.2046,
      "step": 1185
    },
    {
      "epoch": 0.15341295271385572,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2014826536178589,
      "eval_runtime": 14.1547,
      "eval_samples_per_second": 3.109,
      "eval_steps_per_second": 0.212,
      "step": 1185
    },
    {
      "epoch": 0.15354241512120917,
      "grad_norm": 4.74114632109422,
      "learning_rate": 1.2391433972594414e-06,
      "loss": 4.1753,
      "step": 1186
    },
    {
      "epoch": 0.15367187752856265,
      "grad_norm": 4.390193692733157,
      "learning_rate": 1.239090912927303e-06,
      "loss": 4.4556,
      "step": 1187
    },
    {
      "epoch": 0.1538013399359161,
      "grad_norm": 2.052791091352851,
      "learning_rate": 1.2390383031546488e-06,
      "loss": 2.9971,
      "step": 1188
    },
    {
      "epoch": 0.15393080234326958,
      "grad_norm": 2.4078215431835437,
      "learning_rate": 1.238985567952225e-06,
      "loss": 3.1335,
      "step": 1189
    },
    {
      "epoch": 0.15406026475062304,
      "grad_norm": 6.639587871619227,
      "learning_rate": 1.2389327073308043e-06,
      "loss": 5.4448,
      "step": 1190
    },
    {
      "epoch": 0.15406026475062304,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2037907838821411,
      "eval_runtime": 14.5114,
      "eval_samples_per_second": 3.032,
      "eval_steps_per_second": 0.207,
      "step": 1190
    },
    {
      "epoch": 0.1541897271579765,
      "grad_norm": 4.836779897554295,
      "learning_rate": 1.238879721301184e-06,
      "loss": 4.4336,
      "step": 1191
    },
    {
      "epoch": 0.15431918956532997,
      "grad_norm": 7.268946133144218,
      "learning_rate": 1.238826609874188e-06,
      "loss": 6.5786,
      "step": 1192
    },
    {
      "epoch": 0.15444865197268343,
      "grad_norm": 2.8229544082602063,
      "learning_rate": 1.2387733730606651e-06,
      "loss": 3.4456,
      "step": 1193
    },
    {
      "epoch": 0.1545781143800369,
      "grad_norm": 2.8855991056152677,
      "learning_rate": 1.2387200108714902e-06,
      "loss": 3.8381,
      "step": 1194
    },
    {
      "epoch": 0.15470757678739036,
      "grad_norm": 4.997642536024528,
      "learning_rate": 1.2386665233175638e-06,
      "loss": 4.0288,
      "step": 1195
    },
    {
      "epoch": 0.15470757678739036,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2100497484207153,
      "eval_runtime": 13.8741,
      "eval_samples_per_second": 3.171,
      "eval_steps_per_second": 0.216,
      "step": 1195
    },
    {
      "epoch": 0.15483703919474381,
      "grad_norm": 3.9264123046005794,
      "learning_rate": 1.2386129104098114e-06,
      "loss": 3.6814,
      "step": 1196
    },
    {
      "epoch": 0.1549665016020973,
      "grad_norm": 5.068382546763015,
      "learning_rate": 1.2385591721591844e-06,
      "loss": 4.186,
      "step": 1197
    },
    {
      "epoch": 0.15509596400945075,
      "grad_norm": 1.4889991711389337,
      "learning_rate": 1.2385053085766605e-06,
      "loss": 2.7527,
      "step": 1198
    },
    {
      "epoch": 0.15522542641680423,
      "grad_norm": 4.767650595764256,
      "learning_rate": 1.238451319673242e-06,
      "loss": 4.3545,
      "step": 1199
    },
    {
      "epoch": 0.15535488882415768,
      "grad_norm": 5.081639699506577,
      "learning_rate": 1.2383972054599575e-06,
      "loss": 4.8423,
      "step": 1200
    },
    {
      "epoch": 0.15535488882415768,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2083629369735718,
      "eval_runtime": 14.4094,
      "eval_samples_per_second": 3.054,
      "eval_steps_per_second": 0.208,
      "step": 1200
    },
    {
      "epoch": 0.15548435123151114,
      "grad_norm": 3.5174291881211217,
      "learning_rate": 1.2383429659478608e-06,
      "loss": 3.3845,
      "step": 1201
    },
    {
      "epoch": 0.15561381363886462,
      "grad_norm": 4.005290830675439,
      "learning_rate": 1.2382886011480312e-06,
      "loss": 3.6633,
      "step": 1202
    },
    {
      "epoch": 0.15574327604621807,
      "grad_norm": 6.014955308875428,
      "learning_rate": 1.238234111071574e-06,
      "loss": 4.2661,
      "step": 1203
    },
    {
      "epoch": 0.15587273845357155,
      "grad_norm": 1.9366912742329165,
      "learning_rate": 1.2381794957296201e-06,
      "loss": 2.7686,
      "step": 1204
    },
    {
      "epoch": 0.156002200860925,
      "grad_norm": 8.049063350431043,
      "learning_rate": 1.2381247551333255e-06,
      "loss": 5.6768,
      "step": 1205
    },
    {
      "epoch": 0.156002200860925,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2156871557235718,
      "eval_runtime": 15.5954,
      "eval_samples_per_second": 2.821,
      "eval_steps_per_second": 0.192,
      "step": 1205
    },
    {
      "epoch": 0.1561316632682785,
      "grad_norm": 1.8554917158094817,
      "learning_rate": 1.2380698892938726e-06,
      "loss": 3.0439,
      "step": 1206
    },
    {
      "epoch": 0.15626112567563194,
      "grad_norm": 4.908249250985227,
      "learning_rate": 1.238014898222468e-06,
      "loss": 4.4438,
      "step": 1207
    },
    {
      "epoch": 0.1563905880829854,
      "grad_norm": 2.756006043001854,
      "learning_rate": 1.2379597819303454e-06,
      "loss": 3.3386,
      "step": 1208
    },
    {
      "epoch": 0.15652005049033887,
      "grad_norm": 3.6296518464344203,
      "learning_rate": 1.237904540428763e-06,
      "loss": 3.5903,
      "step": 1209
    },
    {
      "epoch": 0.15664951289769233,
      "grad_norm": 1.3385748688924646,
      "learning_rate": 1.2378491737290053e-06,
      "loss": 2.8218,
      "step": 1210
    },
    {
      "epoch": 0.15664951289769233,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2168856859207153,
      "eval_runtime": 14.6688,
      "eval_samples_per_second": 3.0,
      "eval_steps_per_second": 0.205,
      "step": 1210
    },
    {
      "epoch": 0.1567789753050458,
      "grad_norm": 3.23412280352715,
      "learning_rate": 1.237793681842382e-06,
      "loss": 3.2202,
      "step": 1211
    },
    {
      "epoch": 0.15690843771239926,
      "grad_norm": 2.590075478031581,
      "learning_rate": 1.2377380647802283e-06,
      "loss": 3.4526,
      "step": 1212
    },
    {
      "epoch": 0.15703790011975272,
      "grad_norm": 3.7604559398392414,
      "learning_rate": 1.2376823225539056e-06,
      "loss": 3.9907,
      "step": 1213
    },
    {
      "epoch": 0.1571673625271062,
      "grad_norm": 3.145859799107146,
      "learning_rate": 1.2376264551747998e-06,
      "loss": 3.7087,
      "step": 1214
    },
    {
      "epoch": 0.15729682493445965,
      "grad_norm": 5.218020569732328,
      "learning_rate": 1.237570462654323e-06,
      "loss": 3.866,
      "step": 1215
    },
    {
      "epoch": 0.15729682493445965,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2225230932235718,
      "eval_runtime": 15.0032,
      "eval_samples_per_second": 2.933,
      "eval_steps_per_second": 0.2,
      "step": 1215
    },
    {
      "epoch": 0.15742628734181313,
      "grad_norm": 3.7469742939137856,
      "learning_rate": 1.2375143450039131e-06,
      "loss": 4.0188,
      "step": 1216
    },
    {
      "epoch": 0.15755574974916658,
      "grad_norm": 2.6400573546658612,
      "learning_rate": 1.2374581022350333e-06,
      "loss": 3.0256,
      "step": 1217
    },
    {
      "epoch": 0.15768521215652004,
      "grad_norm": 4.263157878094292,
      "learning_rate": 1.237401734359172e-06,
      "loss": 4.6477,
      "step": 1218
    },
    {
      "epoch": 0.15781467456387352,
      "grad_norm": 6.05031999725171,
      "learning_rate": 1.2373452413878438e-06,
      "loss": 4.3477,
      "step": 1219
    },
    {
      "epoch": 0.15794413697122697,
      "grad_norm": 5.309455621391533,
      "learning_rate": 1.237288623332588e-06,
      "loss": 4.7163,
      "step": 1220
    },
    {
      "epoch": 0.15794413697122697,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2266956567764282,
      "eval_runtime": 14.5441,
      "eval_samples_per_second": 3.025,
      "eval_steps_per_second": 0.206,
      "step": 1220
    },
    {
      "epoch": 0.15807359937858045,
      "grad_norm": 3.2215044538193234,
      "learning_rate": 1.2372318802049706e-06,
      "loss": 3.3384,
      "step": 1221
    },
    {
      "epoch": 0.1582030617859339,
      "grad_norm": 6.635073780420488,
      "learning_rate": 1.2371750120165821e-06,
      "loss": 4.4185,
      "step": 1222
    },
    {
      "epoch": 0.15833252419328736,
      "grad_norm": 5.25567473429258,
      "learning_rate": 1.2371180187790394e-06,
      "loss": 3.8003,
      "step": 1223
    },
    {
      "epoch": 0.15846198660064084,
      "grad_norm": 2.0632757968737163,
      "learning_rate": 1.2370609005039844e-06,
      "loss": 3.0242,
      "step": 1224
    },
    {
      "epoch": 0.1585914490079943,
      "grad_norm": 1.5882276585349184,
      "learning_rate": 1.2370036572030844e-06,
      "loss": 2.7036,
      "step": 1225
    },
    {
      "epoch": 0.1585914490079943,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2283381223678589,
      "eval_runtime": 14.3375,
      "eval_samples_per_second": 3.069,
      "eval_steps_per_second": 0.209,
      "step": 1225
    },
    {
      "epoch": 0.15872091141534778,
      "grad_norm": 4.8521324761303894,
      "learning_rate": 1.2369462888880327e-06,
      "loss": 3.5801,
      "step": 1226
    },
    {
      "epoch": 0.15885037382270123,
      "grad_norm": 3.207378128337535,
      "learning_rate": 1.2368887955705477e-06,
      "loss": 3.7197,
      "step": 1227
    },
    {
      "epoch": 0.1589798362300547,
      "grad_norm": 6.306439542762054,
      "learning_rate": 1.236831177262374e-06,
      "loss": 4.5679,
      "step": 1228
    },
    {
      "epoch": 0.15910929863740816,
      "grad_norm": 2.937161580711878,
      "learning_rate": 1.236773433975281e-06,
      "loss": 3.0605,
      "step": 1229
    },
    {
      "epoch": 0.15923876104476162,
      "grad_norm": 2.516742529686541,
      "learning_rate": 1.236715565721064e-06,
      "loss": 3.4512,
      "step": 1230
    },
    {
      "epoch": 0.15923876104476162,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2337535619735718,
      "eval_runtime": 14.1828,
      "eval_samples_per_second": 3.102,
      "eval_steps_per_second": 0.212,
      "step": 1230
    },
    {
      "epoch": 0.1593682234521151,
      "grad_norm": 4.335310410524557,
      "learning_rate": 1.236657572511544e-06,
      "loss": 3.855,
      "step": 1231
    },
    {
      "epoch": 0.15949768585946855,
      "grad_norm": 3.165587452637354,
      "learning_rate": 1.236599454358567e-06,
      "loss": 3.1357,
      "step": 1232
    },
    {
      "epoch": 0.15962714826682203,
      "grad_norm": 5.283993449053063,
      "learning_rate": 1.2365412112740048e-06,
      "loss": 4.6313,
      "step": 1233
    },
    {
      "epoch": 0.1597566106741755,
      "grad_norm": 4.918346096387316,
      "learning_rate": 1.2364828432697549e-06,
      "loss": 3.8223,
      "step": 1234
    },
    {
      "epoch": 0.15988607308152894,
      "grad_norm": 5.295910560956783,
      "learning_rate": 1.23642435035774e-06,
      "loss": 3.6594,
      "step": 1235
    },
    {
      "epoch": 0.15988607308152894,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2322443723678589,
      "eval_runtime": 13.8672,
      "eval_samples_per_second": 3.173,
      "eval_steps_per_second": 0.216,
      "step": 1235
    },
    {
      "epoch": 0.16001553548888242,
      "grad_norm": 5.161114186182553,
      "learning_rate": 1.2363657325499085e-06,
      "loss": 4.7429,
      "step": 1236
    },
    {
      "epoch": 0.16014499789623587,
      "grad_norm": 2.189717833652239,
      "learning_rate": 1.2363069898582345e-06,
      "loss": 3.3198,
      "step": 1237
    },
    {
      "epoch": 0.16027446030358936,
      "grad_norm": 5.000926250189146,
      "learning_rate": 1.236248122294717e-06,
      "loss": 4.1372,
      "step": 1238
    },
    {
      "epoch": 0.1604039227109428,
      "grad_norm": 4.22457709137243,
      "learning_rate": 1.236189129871381e-06,
      "loss": 3.9172,
      "step": 1239
    },
    {
      "epoch": 0.16053338511829626,
      "grad_norm": 4.48977911398757,
      "learning_rate": 1.2361300126002774e-06,
      "loss": 4.179,
      "step": 1240
    },
    {
      "epoch": 0.16053338511829626,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2400124073028564,
      "eval_runtime": 14.3539,
      "eval_samples_per_second": 3.065,
      "eval_steps_per_second": 0.209,
      "step": 1240
    },
    {
      "epoch": 0.16066284752564974,
      "grad_norm": 3.5557076852399754,
      "learning_rate": 1.2360707704934814e-06,
      "loss": 3.5537,
      "step": 1241
    },
    {
      "epoch": 0.1607923099330032,
      "grad_norm": 4.545641388873724,
      "learning_rate": 1.2360114035630947e-06,
      "loss": 3.4026,
      "step": 1242
    },
    {
      "epoch": 0.16092177234035668,
      "grad_norm": 2.9613659702626616,
      "learning_rate": 1.235951911821244e-06,
      "loss": 3.6074,
      "step": 1243
    },
    {
      "epoch": 0.16105123474771013,
      "grad_norm": 4.689840802864478,
      "learning_rate": 1.2358922952800822e-06,
      "loss": 4.6143,
      "step": 1244
    },
    {
      "epoch": 0.16118069715506358,
      "grad_norm": 3.6422182462756725,
      "learning_rate": 1.2358325539517867e-06,
      "loss": 3.7112,
      "step": 1245
    },
    {
      "epoch": 0.16118069715506358,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2424982786178589,
      "eval_runtime": 14.6766,
      "eval_samples_per_second": 2.998,
      "eval_steps_per_second": 0.204,
      "step": 1245
    },
    {
      "epoch": 0.16131015956241707,
      "grad_norm": 5.322698677692095,
      "learning_rate": 1.235772687848561e-06,
      "loss": 4.4839,
      "step": 1246
    },
    {
      "epoch": 0.16143962196977052,
      "grad_norm": 4.083340946445587,
      "learning_rate": 1.2357126969826342e-06,
      "loss": 3.7927,
      "step": 1247
    },
    {
      "epoch": 0.161569084377124,
      "grad_norm": 3.7809745288248524,
      "learning_rate": 1.2356525813662604e-06,
      "loss": 3.9868,
      "step": 1248
    },
    {
      "epoch": 0.16169854678447745,
      "grad_norm": 4.409665189402354,
      "learning_rate": 1.2355923410117193e-06,
      "loss": 3.783,
      "step": 1249
    },
    {
      "epoch": 0.16182800919183093,
      "grad_norm": 5.564918708338664,
      "learning_rate": 1.2355319759313167e-06,
      "loss": 4.75,
      "step": 1250
    },
    {
      "epoch": 0.16182800919183093,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2487126588821411,
      "eval_runtime": 14.722,
      "eval_samples_per_second": 2.989,
      "eval_steps_per_second": 0.204,
      "step": 1250
    },
    {
      "epoch": 0.1619574715991844,
      "grad_norm": 5.442702575649247,
      "learning_rate": 1.2354714861373828e-06,
      "loss": 4.0142,
      "step": 1251
    },
    {
      "epoch": 0.16208693400653784,
      "grad_norm": 2.6369781872365317,
      "learning_rate": 1.2354108716422744e-06,
      "loss": 3.54,
      "step": 1252
    },
    {
      "epoch": 0.16221639641389132,
      "grad_norm": 2.9925332747938342,
      "learning_rate": 1.235350132458373e-06,
      "loss": 3.5942,
      "step": 1253
    },
    {
      "epoch": 0.16234585882124478,
      "grad_norm": 5.268174377931261,
      "learning_rate": 1.2352892685980856e-06,
      "loss": 4.5454,
      "step": 1254
    },
    {
      "epoch": 0.16247532122859826,
      "grad_norm": 2.2686559422744974,
      "learning_rate": 1.2352282800738451e-06,
      "loss": 2.9609,
      "step": 1255
    },
    {
      "epoch": 0.16247532122859826,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2497336864471436,
      "eval_runtime": 13.888,
      "eval_samples_per_second": 3.168,
      "eval_steps_per_second": 0.216,
      "step": 1255
    },
    {
      "epoch": 0.1626047836359517,
      "grad_norm": 5.607461141405671,
      "learning_rate": 1.2351671668981098e-06,
      "loss": 4.8423,
      "step": 1256
    },
    {
      "epoch": 0.16273424604330516,
      "grad_norm": 5.850364440689246,
      "learning_rate": 1.2351059290833631e-06,
      "loss": 3.8564,
      "step": 1257
    },
    {
      "epoch": 0.16286370845065865,
      "grad_norm": 5.8228065719225235,
      "learning_rate": 1.235044566642114e-06,
      "loss": 4.0498,
      "step": 1258
    },
    {
      "epoch": 0.1629931708580121,
      "grad_norm": 3.526841380655274,
      "learning_rate": 1.2349830795868972e-06,
      "loss": 3.4312,
      "step": 1259
    },
    {
      "epoch": 0.16312263326536558,
      "grad_norm": 2.100369748774491,
      "learning_rate": 1.2349214679302726e-06,
      "loss": 3.0347,
      "step": 1260
    },
    {
      "epoch": 0.16312263326536558,
      "eval_PRM Accuracy": 0.2830188679245283,
      "eval_PRM F1": 0.15555555555555556,
      "eval_PRM F1 AUC": 0.5421686746987951,
      "eval_PRM F1 Neg": 0.3770491803278688,
      "eval_PRM NPV": 0.23232323232323232,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.08433734939759036,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.253373622894287,
      "eval_runtime": 14.4406,
      "eval_samples_per_second": 3.047,
      "eval_steps_per_second": 0.208,
      "step": 1260
    },
    {
      "epoch": 0.16325209567271903,
      "grad_norm": 2.7409557649454532,
      "learning_rate": 1.2348597316848257e-06,
      "loss": 3.1353,
      "step": 1261
    },
    {
      "epoch": 0.1633815580800725,
      "grad_norm": 5.024298758837595,
      "learning_rate": 1.2347978708631671e-06,
      "loss": 4.2914,
      "step": 1262
    },
    {
      "epoch": 0.16351102048742597,
      "grad_norm": 2.2232604215036864,
      "learning_rate": 1.2347358854779337e-06,
      "loss": 3.2371,
      "step": 1263
    },
    {
      "epoch": 0.16364048289477942,
      "grad_norm": 4.994951838441288,
      "learning_rate": 1.2346737755417867e-06,
      "loss": 5.2429,
      "step": 1264
    },
    {
      "epoch": 0.1637699453021329,
      "grad_norm": 5.528739361458071,
      "learning_rate": 1.2346115410674133e-06,
      "loss": 4.6021,
      "step": 1265
    },
    {
      "epoch": 0.1637699453021329,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2551935911178589,
      "eval_runtime": 13.7653,
      "eval_samples_per_second": 3.196,
      "eval_steps_per_second": 0.218,
      "step": 1265
    },
    {
      "epoch": 0.16389940770948636,
      "grad_norm": 4.479468314724481,
      "learning_rate": 1.2345491820675268e-06,
      "loss": 3.6948,
      "step": 1266
    },
    {
      "epoch": 0.1640288701168398,
      "grad_norm": 2.2989754140783627,
      "learning_rate": 1.2344866985548646e-06,
      "loss": 3.1912,
      "step": 1267
    },
    {
      "epoch": 0.1641583325241933,
      "grad_norm": 3.4286256750104673,
      "learning_rate": 1.2344240905421906e-06,
      "loss": 3.6475,
      "step": 1268
    },
    {
      "epoch": 0.16428779493154674,
      "grad_norm": 4.436999930092937,
      "learning_rate": 1.2343613580422934e-06,
      "loss": 3.7156,
      "step": 1269
    },
    {
      "epoch": 0.16441725733890022,
      "grad_norm": 3.6373496122515894,
      "learning_rate": 1.2342985010679879e-06,
      "loss": 4.1511,
      "step": 1270
    },
    {
      "epoch": 0.16441725733890022,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.261186122894287,
      "eval_runtime": 13.9727,
      "eval_samples_per_second": 3.149,
      "eval_steps_per_second": 0.215,
      "step": 1270
    },
    {
      "epoch": 0.16454671974625368,
      "grad_norm": 6.167071431690786,
      "learning_rate": 1.2342355196321135e-06,
      "loss": 4.5303,
      "step": 1271
    },
    {
      "epoch": 0.16467618215360716,
      "grad_norm": 5.847337076567583,
      "learning_rate": 1.2341724137475355e-06,
      "loss": 4.4424,
      "step": 1272
    },
    {
      "epoch": 0.1648056445609606,
      "grad_norm": 6.031013766434898,
      "learning_rate": 1.2341091834271449e-06,
      "loss": 4.2832,
      "step": 1273
    },
    {
      "epoch": 0.16493510696831407,
      "grad_norm": 3.126763617576407,
      "learning_rate": 1.2340458286838573e-06,
      "loss": 3.2061,
      "step": 1274
    },
    {
      "epoch": 0.16506456937566755,
      "grad_norm": 3.6081058992460684,
      "learning_rate": 1.2339823495306145e-06,
      "loss": 3.8472,
      "step": 1275
    },
    {
      "epoch": 0.16506456937566755,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.26416015625,
      "eval_runtime": 15.217,
      "eval_samples_per_second": 2.892,
      "eval_steps_per_second": 0.197,
      "step": 1275
    },
    {
      "epoch": 0.165194031783021,
      "grad_norm": 5.146830772092357,
      "learning_rate": 1.2339187459803831e-06,
      "loss": 4.2148,
      "step": 1276
    },
    {
      "epoch": 0.16532349419037448,
      "grad_norm": 3.9509885467502297,
      "learning_rate": 1.233855018046156e-06,
      "loss": 3.8145,
      "step": 1277
    },
    {
      "epoch": 0.16545295659772793,
      "grad_norm": 3.1281148078265506,
      "learning_rate": 1.23379116574095e-06,
      "loss": 3.6353,
      "step": 1278
    },
    {
      "epoch": 0.1655824190050814,
      "grad_norm": 2.3341125663017976,
      "learning_rate": 1.2337271890778093e-06,
      "loss": 2.8467,
      "step": 1279
    },
    {
      "epoch": 0.16571188141243487,
      "grad_norm": 6.860199796848311,
      "learning_rate": 1.2336630880698015e-06,
      "loss": 4.2173,
      "step": 1280
    },
    {
      "epoch": 0.16571188141243487,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2681552171707153,
      "eval_runtime": 15.3122,
      "eval_samples_per_second": 2.874,
      "eval_steps_per_second": 0.196,
      "step": 1280
    },
    {
      "epoch": 0.16584134381978832,
      "grad_norm": 2.1309762049815233,
      "learning_rate": 1.233598862730021e-06,
      "loss": 2.9844,
      "step": 1281
    },
    {
      "epoch": 0.1659708062271418,
      "grad_norm": 1.8571640811545418,
      "learning_rate": 1.2335345130715873e-06,
      "loss": 3.1099,
      "step": 1282
    },
    {
      "epoch": 0.16610026863449526,
      "grad_norm": 2.737781773506582,
      "learning_rate": 1.2334700391076445e-06,
      "loss": 3.0718,
      "step": 1283
    },
    {
      "epoch": 0.1662297310418487,
      "grad_norm": 3.7674622944115628,
      "learning_rate": 1.2334054408513636e-06,
      "loss": 3.7505,
      "step": 1284
    },
    {
      "epoch": 0.1663591934492022,
      "grad_norm": 1.9738861955669518,
      "learning_rate": 1.233340718315939e-06,
      "loss": 3.2979,
      "step": 1285
    },
    {
      "epoch": 0.1663591934492022,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2753018140792847,
      "eval_runtime": 13.6962,
      "eval_samples_per_second": 3.213,
      "eval_steps_per_second": 0.219,
      "step": 1285
    },
    {
      "epoch": 0.16648865585655565,
      "grad_norm": 3.836604791934664,
      "learning_rate": 1.2332758715145926e-06,
      "loss": 3.3862,
      "step": 1286
    },
    {
      "epoch": 0.16661811826390913,
      "grad_norm": 8.380308107723765,
      "learning_rate": 1.23321090046057e-06,
      "loss": 5.3901,
      "step": 1287
    },
    {
      "epoch": 0.16674758067126258,
      "grad_norm": 4.966178759450729,
      "learning_rate": 1.2331458051671432e-06,
      "loss": 4.5332,
      "step": 1288
    },
    {
      "epoch": 0.16687704307861603,
      "grad_norm": 1.543012905635976,
      "learning_rate": 1.233080585647609e-06,
      "loss": 2.7429,
      "step": 1289
    },
    {
      "epoch": 0.16700650548596951,
      "grad_norm": 3.495369542423535,
      "learning_rate": 1.2330152419152903e-06,
      "loss": 3.7144,
      "step": 1290
    },
    {
      "epoch": 0.16700650548596951,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2757457494735718,
      "eval_runtime": 13.8008,
      "eval_samples_per_second": 3.188,
      "eval_steps_per_second": 0.217,
      "step": 1290
    },
    {
      "epoch": 0.16713596789332297,
      "grad_norm": 4.883376146174935,
      "learning_rate": 1.2329497739835343e-06,
      "loss": 4.0574,
      "step": 1291
    },
    {
      "epoch": 0.16726543030067645,
      "grad_norm": 4.421732827938649,
      "learning_rate": 1.2328841818657145e-06,
      "loss": 4.2456,
      "step": 1292
    },
    {
      "epoch": 0.1673948927080299,
      "grad_norm": 2.4483818735743843,
      "learning_rate": 1.2328184655752292e-06,
      "loss": 3.2094,
      "step": 1293
    },
    {
      "epoch": 0.16752435511538338,
      "grad_norm": 2.029621505907904,
      "learning_rate": 1.2327526251255024e-06,
      "loss": 2.8154,
      "step": 1294
    },
    {
      "epoch": 0.16765381752273684,
      "grad_norm": 1.5805262062388568,
      "learning_rate": 1.2326866605299834e-06,
      "loss": 2.7754,
      "step": 1295
    },
    {
      "epoch": 0.16765381752273684,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2783203125,
      "eval_runtime": 14.3194,
      "eval_samples_per_second": 3.073,
      "eval_steps_per_second": 0.21,
      "step": 1295
    },
    {
      "epoch": 0.1677832799300903,
      "grad_norm": 3.2089745667494283,
      "learning_rate": 1.2326205718021466e-06,
      "loss": 3.4331,
      "step": 1296
    },
    {
      "epoch": 0.16791274233744377,
      "grad_norm": 4.429837592989631,
      "learning_rate": 1.232554358955492e-06,
      "loss": 3.4827,
      "step": 1297
    },
    {
      "epoch": 0.16804220474479722,
      "grad_norm": 5.786722189374458,
      "learning_rate": 1.2324880220035455e-06,
      "loss": 3.9502,
      "step": 1298
    },
    {
      "epoch": 0.1681716671521507,
      "grad_norm": 4.642693871719533,
      "learning_rate": 1.2324215609598568e-06,
      "loss": 4.0615,
      "step": 1299
    },
    {
      "epoch": 0.16830112955950416,
      "grad_norm": 4.978951094202258,
      "learning_rate": 1.2323549758380026e-06,
      "loss": 4.196,
      "step": 1300
    },
    {
      "epoch": 0.16830112955950416,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2800959348678589,
      "eval_runtime": 14.1293,
      "eval_samples_per_second": 3.114,
      "eval_steps_per_second": 0.212,
      "step": 1300
    },
    {
      "epoch": 0.1684305919668576,
      "grad_norm": 3.0785794770733723,
      "learning_rate": 1.232288266651584e-06,
      "loss": 3.0371,
      "step": 1301
    },
    {
      "epoch": 0.1685600543742111,
      "grad_norm": 3.8312546894832047,
      "learning_rate": 1.2322214334142279e-06,
      "loss": 3.6953,
      "step": 1302
    },
    {
      "epoch": 0.16868951678156455,
      "grad_norm": 4.189659697476781,
      "learning_rate": 1.232154476139586e-06,
      "loss": 3.6094,
      "step": 1303
    },
    {
      "epoch": 0.16881897918891803,
      "grad_norm": 2.557533080396666,
      "learning_rate": 1.232087394841336e-06,
      "loss": 3.3228,
      "step": 1304
    },
    {
      "epoch": 0.16894844159627148,
      "grad_norm": 2.045420861296134,
      "learning_rate": 1.2320201895331807e-06,
      "loss": 2.9675,
      "step": 1305
    },
    {
      "epoch": 0.16894844159627148,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2811611890792847,
      "eval_runtime": 14.8844,
      "eval_samples_per_second": 2.956,
      "eval_steps_per_second": 0.202,
      "step": 1305
    },
    {
      "epoch": 0.16907790400362493,
      "grad_norm": 5.105850138085465,
      "learning_rate": 1.2319528602288477e-06,
      "loss": 4.3081,
      "step": 1306
    },
    {
      "epoch": 0.16920736641097842,
      "grad_norm": 6.682389823064708,
      "learning_rate": 1.2318854069420906e-06,
      "loss": 4.4058,
      "step": 1307
    },
    {
      "epoch": 0.16933682881833187,
      "grad_norm": 5.176659050729813,
      "learning_rate": 1.2318178296866881e-06,
      "loss": 3.3955,
      "step": 1308
    },
    {
      "epoch": 0.16946629122568535,
      "grad_norm": 2.808313562781416,
      "learning_rate": 1.2317501284764443e-06,
      "loss": 2.9983,
      "step": 1309
    },
    {
      "epoch": 0.1695957536330388,
      "grad_norm": 4.748615457273775,
      "learning_rate": 1.2316823033251887e-06,
      "loss": 4.0342,
      "step": 1310
    },
    {
      "epoch": 0.1695957536330388,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2867542505264282,
      "eval_runtime": 14.0369,
      "eval_samples_per_second": 3.135,
      "eval_steps_per_second": 0.214,
      "step": 1310
    },
    {
      "epoch": 0.16972521604039228,
      "grad_norm": 2.7328742941141178,
      "learning_rate": 1.2316143542467756e-06,
      "loss": 3.1914,
      "step": 1311
    },
    {
      "epoch": 0.16985467844774574,
      "grad_norm": 2.375136785115866,
      "learning_rate": 1.2315462812550853e-06,
      "loss": 2.936,
      "step": 1312
    },
    {
      "epoch": 0.1699841408550992,
      "grad_norm": 5.309048653227103,
      "learning_rate": 1.2314780843640225e-06,
      "loss": 4.2681,
      "step": 1313
    },
    {
      "epoch": 0.17011360326245267,
      "grad_norm": 2.631556473990196,
      "learning_rate": 1.2314097635875185e-06,
      "loss": 3.1074,
      "step": 1314
    },
    {
      "epoch": 0.17024306566980613,
      "grad_norm": 6.010473160130362,
      "learning_rate": 1.2313413189395289e-06,
      "loss": 3.7925,
      "step": 1315
    },
    {
      "epoch": 0.17024306566980613,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2906605005264282,
      "eval_runtime": 14.1661,
      "eval_samples_per_second": 3.106,
      "eval_steps_per_second": 0.212,
      "step": 1315
    },
    {
      "epoch": 0.1703725280771596,
      "grad_norm": 1.2910117682548108,
      "learning_rate": 1.231272750434035e-06,
      "loss": 2.4407,
      "step": 1316
    },
    {
      "epoch": 0.17050199048451306,
      "grad_norm": 4.9717678405248,
      "learning_rate": 1.2312040580850432e-06,
      "loss": 4.2549,
      "step": 1317
    },
    {
      "epoch": 0.17063145289186651,
      "grad_norm": 3.8740745628307462,
      "learning_rate": 1.231135241906585e-06,
      "loss": 4.0288,
      "step": 1318
    },
    {
      "epoch": 0.17076091529922,
      "grad_norm": 8.644608909763942,
      "learning_rate": 1.2310663019127184e-06,
      "loss": 5.8799,
      "step": 1319
    },
    {
      "epoch": 0.17089037770657345,
      "grad_norm": 5.570092690640336,
      "learning_rate": 1.2309972381175249e-06,
      "loss": 4.2168,
      "step": 1320
    },
    {
      "epoch": 0.17089037770657345,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2937678098678589,
      "eval_runtime": 13.9362,
      "eval_samples_per_second": 3.157,
      "eval_steps_per_second": 0.215,
      "step": 1320
    },
    {
      "epoch": 0.17101984011392693,
      "grad_norm": 5.064122521810584,
      "learning_rate": 1.2309280505351127e-06,
      "loss": 4.7144,
      "step": 1321
    },
    {
      "epoch": 0.17114930252128038,
      "grad_norm": 3.9690626796665454,
      "learning_rate": 1.2308587391796143e-06,
      "loss": 3.6675,
      "step": 1322
    },
    {
      "epoch": 0.17127876492863384,
      "grad_norm": 3.4272165270525634,
      "learning_rate": 1.2307893040651884e-06,
      "loss": 3.5312,
      "step": 1323
    },
    {
      "epoch": 0.17140822733598732,
      "grad_norm": 1.3454091952728904,
      "learning_rate": 1.2307197452060182e-06,
      "loss": 2.7812,
      "step": 1324
    },
    {
      "epoch": 0.17153768974334077,
      "grad_norm": 2.851743842237999,
      "learning_rate": 1.2306500626163129e-06,
      "loss": 3.2869,
      "step": 1325
    },
    {
      "epoch": 0.17153768974334077,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.2962092161178589,
      "eval_runtime": 15.2331,
      "eval_samples_per_second": 2.888,
      "eval_steps_per_second": 0.197,
      "step": 1325
    },
    {
      "epoch": 0.17166715215069425,
      "grad_norm": 2.327670258714756,
      "learning_rate": 1.2305802563103062e-06,
      "loss": 3.3921,
      "step": 1326
    },
    {
      "epoch": 0.1717966145580477,
      "grad_norm": 3.668682339811261,
      "learning_rate": 1.2305103263022577e-06,
      "loss": 3.8506,
      "step": 1327
    },
    {
      "epoch": 0.17192607696540116,
      "grad_norm": 1.4938868264866891,
      "learning_rate": 1.2304402726064517e-06,
      "loss": 2.7803,
      "step": 1328
    },
    {
      "epoch": 0.17205553937275464,
      "grad_norm": 2.9262951989712733,
      "learning_rate": 1.2303700952371986e-06,
      "loss": 3.563,
      "step": 1329
    },
    {
      "epoch": 0.1721850017801081,
      "grad_norm": 3.3946300834250827,
      "learning_rate": 1.230299794208833e-06,
      "loss": 3.4497,
      "step": 1330
    },
    {
      "epoch": 0.1721850017801081,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.299272060394287,
      "eval_runtime": 15.5694,
      "eval_samples_per_second": 2.826,
      "eval_steps_per_second": 0.193,
      "step": 1330
    },
    {
      "epoch": 0.17231446418746157,
      "grad_norm": 2.20692189896672,
      "learning_rate": 1.2302293695357157e-06,
      "loss": 2.844,
      "step": 1331
    },
    {
      "epoch": 0.17244392659481503,
      "grad_norm": 4.710959935707902,
      "learning_rate": 1.2301588212322322e-06,
      "loss": 4.3613,
      "step": 1332
    },
    {
      "epoch": 0.1725733890021685,
      "grad_norm": 4.846085251219363,
      "learning_rate": 1.2300881493127936e-06,
      "loss": 3.3613,
      "step": 1333
    },
    {
      "epoch": 0.17270285140952196,
      "grad_norm": 2.1744362773088453,
      "learning_rate": 1.2300173537918357e-06,
      "loss": 2.8889,
      "step": 1334
    },
    {
      "epoch": 0.17283231381687542,
      "grad_norm": 4.429681532794174,
      "learning_rate": 1.2299464346838202e-06,
      "loss": 3.9434,
      "step": 1335
    },
    {
      "epoch": 0.17283231381687542,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3029118776321411,
      "eval_runtime": 14.9721,
      "eval_samples_per_second": 2.939,
      "eval_steps_per_second": 0.2,
      "step": 1335
    },
    {
      "epoch": 0.1729617762242289,
      "grad_norm": 3.2691116304344794,
      "learning_rate": 1.2298753920032337e-06,
      "loss": 3.6699,
      "step": 1336
    },
    {
      "epoch": 0.17309123863158235,
      "grad_norm": 5.346699742618801,
      "learning_rate": 1.229804225764588e-06,
      "loss": 3.5596,
      "step": 1337
    },
    {
      "epoch": 0.17322070103893583,
      "grad_norm": 4.555852549866342,
      "learning_rate": 1.2297329359824206e-06,
      "loss": 3.4106,
      "step": 1338
    },
    {
      "epoch": 0.17335016344628928,
      "grad_norm": 4.213042149473917,
      "learning_rate": 1.2296615226712938e-06,
      "loss": 4.061,
      "step": 1339
    },
    {
      "epoch": 0.17347962585364274,
      "grad_norm": 3.0846887771916194,
      "learning_rate": 1.2295899858457946e-06,
      "loss": 3.6709,
      "step": 1340
    },
    {
      "epoch": 0.17347962585364274,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.302778720855713,
      "eval_runtime": 15.1779,
      "eval_samples_per_second": 2.899,
      "eval_steps_per_second": 0.198,
      "step": 1340
    },
    {
      "epoch": 0.17360908826099622,
      "grad_norm": 5.731358875225708,
      "learning_rate": 1.2295183255205368e-06,
      "loss": 4.2327,
      "step": 1341
    },
    {
      "epoch": 0.17373855066834967,
      "grad_norm": 5.4889352136054494,
      "learning_rate": 1.2294465417101575e-06,
      "loss": 4.585,
      "step": 1342
    },
    {
      "epoch": 0.17386801307570315,
      "grad_norm": 4.53134594373362,
      "learning_rate": 1.229374634429321e-06,
      "loss": 3.7031,
      "step": 1343
    },
    {
      "epoch": 0.1739974754830566,
      "grad_norm": 4.257610992869868,
      "learning_rate": 1.2293026036927146e-06,
      "loss": 3.3718,
      "step": 1344
    },
    {
      "epoch": 0.17412693789041006,
      "grad_norm": 4.8198308965127366,
      "learning_rate": 1.2292304495150535e-06,
      "loss": 4.1067,
      "step": 1345
    },
    {
      "epoch": 0.17412693789041006,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3063298463821411,
      "eval_runtime": 14.4331,
      "eval_samples_per_second": 3.049,
      "eval_steps_per_second": 0.208,
      "step": 1345
    },
    {
      "epoch": 0.17425640029776354,
      "grad_norm": 3.5659979881622537,
      "learning_rate": 1.2291581719110754e-06,
      "loss": 3.6414,
      "step": 1346
    },
    {
      "epoch": 0.174385862705117,
      "grad_norm": 4.515344822694987,
      "learning_rate": 1.2290857708955451e-06,
      "loss": 3.8911,
      "step": 1347
    },
    {
      "epoch": 0.17451532511247048,
      "grad_norm": 2.1358031247437914,
      "learning_rate": 1.229013246483252e-06,
      "loss": 2.9033,
      "step": 1348
    },
    {
      "epoch": 0.17464478751982393,
      "grad_norm": 6.95039018439194,
      "learning_rate": 1.2289405986890102e-06,
      "loss": 4.8418,
      "step": 1349
    },
    {
      "epoch": 0.17477424992717738,
      "grad_norm": 2.5381980477001598,
      "learning_rate": 1.2288678275276599e-06,
      "loss": 3.2893,
      "step": 1350
    },
    {
      "epoch": 0.17477424992717738,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.313432216644287,
      "eval_runtime": 14.6309,
      "eval_samples_per_second": 3.007,
      "eval_steps_per_second": 0.205,
      "step": 1350
    },
    {
      "epoch": 0.17490371233453086,
      "grad_norm": 3.5215403014602464,
      "learning_rate": 1.2287949330140662e-06,
      "loss": 3.416,
      "step": 1351
    },
    {
      "epoch": 0.17503317474188432,
      "grad_norm": 2.2585148190601405,
      "learning_rate": 1.2287219151631186e-06,
      "loss": 3.1743,
      "step": 1352
    },
    {
      "epoch": 0.1751626371492378,
      "grad_norm": 4.376247385684798,
      "learning_rate": 1.2286487739897334e-06,
      "loss": 4.624,
      "step": 1353
    },
    {
      "epoch": 0.17529209955659125,
      "grad_norm": 1.882267898193611,
      "learning_rate": 1.2285755095088506e-06,
      "loss": 2.9172,
      "step": 1354
    },
    {
      "epoch": 0.17542156196394473,
      "grad_norm": 7.536306206174726,
      "learning_rate": 1.2285021217354359e-06,
      "loss": 5.127,
      "step": 1355
    },
    {
      "epoch": 0.17542156196394473,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3166725635528564,
      "eval_runtime": 14.631,
      "eval_samples_per_second": 3.007,
      "eval_steps_per_second": 0.205,
      "step": 1355
    },
    {
      "epoch": 0.1755510243712982,
      "grad_norm": 6.011467643089391,
      "learning_rate": 1.2284286106844805e-06,
      "loss": 4.1162,
      "step": 1356
    },
    {
      "epoch": 0.17568048677865164,
      "grad_norm": 5.668595721896454,
      "learning_rate": 1.2283549763710005e-06,
      "loss": 3.8235,
      "step": 1357
    },
    {
      "epoch": 0.17580994918600512,
      "grad_norm": 7.105469287931377,
      "learning_rate": 1.228281218810037e-06,
      "loss": 4.7415,
      "step": 1358
    },
    {
      "epoch": 0.17593941159335857,
      "grad_norm": 3.684870119131621,
      "learning_rate": 1.2282073380166568e-06,
      "loss": 3.4263,
      "step": 1359
    },
    {
      "epoch": 0.17606887400071206,
      "grad_norm": 6.516145571408506,
      "learning_rate": 1.2281333340059512e-06,
      "loss": 4.542,
      "step": 1360
    },
    {
      "epoch": 0.17606887400071206,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.320268154144287,
      "eval_runtime": 15.377,
      "eval_samples_per_second": 2.861,
      "eval_steps_per_second": 0.195,
      "step": 1360
    },
    {
      "epoch": 0.1761983364080655,
      "grad_norm": 4.591769682088052,
      "learning_rate": 1.2280592067930371e-06,
      "loss": 3.9644,
      "step": 1361
    },
    {
      "epoch": 0.17632779881541896,
      "grad_norm": 3.502751007308499,
      "learning_rate": 1.2279849563930567e-06,
      "loss": 3.6277,
      "step": 1362
    },
    {
      "epoch": 0.17645726122277244,
      "grad_norm": 3.586476933477245,
      "learning_rate": 1.227910582821177e-06,
      "loss": 3.2766,
      "step": 1363
    },
    {
      "epoch": 0.1765867236301259,
      "grad_norm": 4.671775247889118,
      "learning_rate": 1.2278360860925904e-06,
      "loss": 3.6133,
      "step": 1364
    },
    {
      "epoch": 0.17671618603747938,
      "grad_norm": 2.7099668128122816,
      "learning_rate": 1.2277614662225142e-06,
      "loss": 3.2119,
      "step": 1365
    },
    {
      "epoch": 0.17671618603747938,
      "eval_PRM Accuracy": 0.27358490566037735,
      "eval_PRM F1": 0.1348314606741573,
      "eval_PRM F1 AUC": 0.536144578313253,
      "eval_PRM F1 Neg": 0.37398373983739835,
      "eval_PRM NPV": 0.23,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.07228915662650602,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.326615810394287,
      "eval_runtime": 14.0466,
      "eval_samples_per_second": 3.132,
      "eval_steps_per_second": 0.214,
      "step": 1365
    },
    {
      "epoch": 0.17684564844483283,
      "grad_norm": 1.986955188132743,
      "learning_rate": 1.227686723226191e-06,
      "loss": 3.3013,
      "step": 1366
    },
    {
      "epoch": 0.17697511085218628,
      "grad_norm": 3.1356637150410545,
      "learning_rate": 1.227611857118889e-06,
      "loss": 3.6294,
      "step": 1367
    },
    {
      "epoch": 0.17710457325953977,
      "grad_norm": 4.861762151968143,
      "learning_rate": 1.2275368679159002e-06,
      "loss": 3.8896,
      "step": 1368
    },
    {
      "epoch": 0.17723403566689322,
      "grad_norm": 4.958798865951923,
      "learning_rate": 1.2274617556325438e-06,
      "loss": 4.8574,
      "step": 1369
    },
    {
      "epoch": 0.1773634980742467,
      "grad_norm": 2.7922007386505765,
      "learning_rate": 1.2273865202841623e-06,
      "loss": 2.9448,
      "step": 1370
    },
    {
      "epoch": 0.1773634980742467,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3270152807235718,
      "eval_runtime": 13.8583,
      "eval_samples_per_second": 3.175,
      "eval_steps_per_second": 0.216,
      "step": 1370
    },
    {
      "epoch": 0.17749296048160015,
      "grad_norm": 3.004859369357232,
      "learning_rate": 1.2273111618861242e-06,
      "loss": 3.6426,
      "step": 1371
    },
    {
      "epoch": 0.1776224228889536,
      "grad_norm": 5.478934915785173,
      "learning_rate": 1.227235680453823e-06,
      "loss": 3.8804,
      "step": 1372
    },
    {
      "epoch": 0.1777518852963071,
      "grad_norm": 6.952092575527702,
      "learning_rate": 1.2271600760026771e-06,
      "loss": 4.3699,
      "step": 1373
    },
    {
      "epoch": 0.17788134770366054,
      "grad_norm": 3.567671522731947,
      "learning_rate": 1.2270843485481307e-06,
      "loss": 3.3667,
      "step": 1374
    },
    {
      "epoch": 0.17801081011101402,
      "grad_norm": 3.4333439439210265,
      "learning_rate": 1.2270084981056523e-06,
      "loss": 3.2964,
      "step": 1375
    },
    {
      "epoch": 0.17801081011101402,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3323420286178589,
      "eval_runtime": 15.235,
      "eval_samples_per_second": 2.888,
      "eval_steps_per_second": 0.197,
      "step": 1375
    },
    {
      "epoch": 0.17814027251836748,
      "grad_norm": 2.3634794909945134,
      "learning_rate": 1.2269325246907362e-06,
      "loss": 2.8354,
      "step": 1376
    },
    {
      "epoch": 0.17826973492572096,
      "grad_norm": 9.23330369497804,
      "learning_rate": 1.226856428318901e-06,
      "loss": 6.0547,
      "step": 1377
    },
    {
      "epoch": 0.1783991973330744,
      "grad_norm": 7.978881425745575,
      "learning_rate": 1.2267802090056916e-06,
      "loss": 4.6367,
      "step": 1378
    },
    {
      "epoch": 0.17852865974042786,
      "grad_norm": 6.601964940830656,
      "learning_rate": 1.226703866766677e-06,
      "loss": 4.6245,
      "step": 1379
    },
    {
      "epoch": 0.17865812214778135,
      "grad_norm": 2.8174317770600537,
      "learning_rate": 1.2266274016174512e-06,
      "loss": 3.7234,
      "step": 1380
    },
    {
      "epoch": 0.17865812214778135,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3362038135528564,
      "eval_runtime": 14.7693,
      "eval_samples_per_second": 2.979,
      "eval_steps_per_second": 0.203,
      "step": 1380
    },
    {
      "epoch": 0.1787875845551348,
      "grad_norm": 2.342349869371614,
      "learning_rate": 1.2265508135736344e-06,
      "loss": 3.1567,
      "step": 1381
    },
    {
      "epoch": 0.17891704696248828,
      "grad_norm": 2.59091194915207,
      "learning_rate": 1.2264741026508712e-06,
      "loss": 3.0042,
      "step": 1382
    },
    {
      "epoch": 0.17904650936984173,
      "grad_norm": 4.304851786405775,
      "learning_rate": 1.226397268864831e-06,
      "loss": 3.7949,
      "step": 1383
    },
    {
      "epoch": 0.1791759717771952,
      "grad_norm": 2.7736195632209615,
      "learning_rate": 1.2263203122312089e-06,
      "loss": 2.9407,
      "step": 1384
    },
    {
      "epoch": 0.17930543418454867,
      "grad_norm": 5.7694747836934495,
      "learning_rate": 1.2262432327657251e-06,
      "loss": 4.4531,
      "step": 1385
    },
    {
      "epoch": 0.17930543418454867,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.341264247894287,
      "eval_runtime": 14.5408,
      "eval_samples_per_second": 3.026,
      "eval_steps_per_second": 0.206,
      "step": 1385
    },
    {
      "epoch": 0.17943489659190212,
      "grad_norm": 2.4345289511602974,
      "learning_rate": 1.226166030484124e-06,
      "loss": 3.4282,
      "step": 1386
    },
    {
      "epoch": 0.1795643589992556,
      "grad_norm": 5.992520859307226,
      "learning_rate": 1.2260887054021763e-06,
      "loss": 3.6562,
      "step": 1387
    },
    {
      "epoch": 0.17969382140660906,
      "grad_norm": 1.2621959680004389,
      "learning_rate": 1.226011257535677e-06,
      "loss": 2.6984,
      "step": 1388
    },
    {
      "epoch": 0.1798232838139625,
      "grad_norm": 3.1977805600564166,
      "learning_rate": 1.2259336869004462e-06,
      "loss": 3.3037,
      "step": 1389
    },
    {
      "epoch": 0.179952746221316,
      "grad_norm": 4.075018442422728,
      "learning_rate": 1.2258559935123297e-06,
      "loss": 3.2825,
      "step": 1390
    },
    {
      "epoch": 0.179952746221316,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.341752529144287,
      "eval_runtime": 14.6531,
      "eval_samples_per_second": 3.003,
      "eval_steps_per_second": 0.205,
      "step": 1390
    },
    {
      "epoch": 0.18008220862866944,
      "grad_norm": 3.903661339265306,
      "learning_rate": 1.2257781773871977e-06,
      "loss": 3.5542,
      "step": 1391
    },
    {
      "epoch": 0.18021167103602292,
      "grad_norm": 3.790227953159591,
      "learning_rate": 1.2257002385409455e-06,
      "loss": 3.5815,
      "step": 1392
    },
    {
      "epoch": 0.18034113344337638,
      "grad_norm": 5.398416413517164,
      "learning_rate": 1.2256221769894943e-06,
      "loss": 3.5898,
      "step": 1393
    },
    {
      "epoch": 0.18047059585072983,
      "grad_norm": 4.4279447629335005,
      "learning_rate": 1.225543992748789e-06,
      "loss": 3.9104,
      "step": 1394
    },
    {
      "epoch": 0.1806000582580833,
      "grad_norm": 1.9659534356002728,
      "learning_rate": 1.2254656858348013e-06,
      "loss": 2.8894,
      "step": 1395
    },
    {
      "epoch": 0.1806000582580833,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3474787473678589,
      "eval_runtime": 14.6952,
      "eval_samples_per_second": 2.994,
      "eval_steps_per_second": 0.204,
      "step": 1395
    },
    {
      "epoch": 0.18072952066543677,
      "grad_norm": 3.3695500855419356,
      "learning_rate": 1.2253872562635258e-06,
      "loss": 3.0664,
      "step": 1396
    },
    {
      "epoch": 0.18085898307279025,
      "grad_norm": 2.923048469363371,
      "learning_rate": 1.2253087040509841e-06,
      "loss": 3.0024,
      "step": 1397
    },
    {
      "epoch": 0.1809884454801437,
      "grad_norm": 2.630209307642065,
      "learning_rate": 1.2252300292132222e-06,
      "loss": 3.0532,
      "step": 1398
    },
    {
      "epoch": 0.18111790788749718,
      "grad_norm": 5.333966673548329,
      "learning_rate": 1.2251512317663102e-06,
      "loss": 4.4951,
      "step": 1399
    },
    {
      "epoch": 0.18124737029485063,
      "grad_norm": 4.768577762469869,
      "learning_rate": 1.225072311726345e-06,
      "loss": 4.1709,
      "step": 1400
    },
    {
      "epoch": 0.18124737029485063,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.351518154144287,
      "eval_runtime": 14.5067,
      "eval_samples_per_second": 3.033,
      "eval_steps_per_second": 0.207,
      "step": 1400
    },
    {
      "epoch": 0.1813768327022041,
      "grad_norm": 2.873356644696701,
      "learning_rate": 1.2249932691094472e-06,
      "loss": 3.0024,
      "step": 1401
    },
    {
      "epoch": 0.18150629510955757,
      "grad_norm": 1.4921759905025873,
      "learning_rate": 1.2249141039317626e-06,
      "loss": 2.5623,
      "step": 1402
    },
    {
      "epoch": 0.18163575751691102,
      "grad_norm": 6.404901401988275,
      "learning_rate": 1.2248348162094628e-06,
      "loss": 4.8521,
      "step": 1403
    },
    {
      "epoch": 0.1817652199242645,
      "grad_norm": 5.340069187711158,
      "learning_rate": 1.2247554059587435e-06,
      "loss": 3.8372,
      "step": 1404
    },
    {
      "epoch": 0.18189468233161796,
      "grad_norm": 2.818863663253772,
      "learning_rate": 1.2246758731958261e-06,
      "loss": 3.5308,
      "step": 1405
    },
    {
      "epoch": 0.18189468233161796,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3575993776321411,
      "eval_runtime": 14.1089,
      "eval_samples_per_second": 3.119,
      "eval_steps_per_second": 0.213,
      "step": 1405
    },
    {
      "epoch": 0.1820241447389714,
      "grad_norm": 6.504035357386434,
      "learning_rate": 1.224596217936957e-06,
      "loss": 3.6816,
      "step": 1406
    },
    {
      "epoch": 0.1821536071463249,
      "grad_norm": 4.226594780373103,
      "learning_rate": 1.2245164401984067e-06,
      "loss": 3.3032,
      "step": 1407
    },
    {
      "epoch": 0.18228306955367835,
      "grad_norm": 6.523869931802508,
      "learning_rate": 1.2244365399964722e-06,
      "loss": 4.1499,
      "step": 1408
    },
    {
      "epoch": 0.18241253196103183,
      "grad_norm": 2.2729241748769664,
      "learning_rate": 1.224356517347474e-06,
      "loss": 3.054,
      "step": 1409
    },
    {
      "epoch": 0.18254199436838528,
      "grad_norm": 2.232002427245416,
      "learning_rate": 1.224276372267759e-06,
      "loss": 2.9849,
      "step": 1410
    },
    {
      "epoch": 0.18254199436838528,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3580876588821411,
      "eval_runtime": 14.3326,
      "eval_samples_per_second": 3.07,
      "eval_steps_per_second": 0.209,
      "step": 1410
    },
    {
      "epoch": 0.18267145677573873,
      "grad_norm": 1.8013043044712627,
      "learning_rate": 1.2241961047736982e-06,
      "loss": 3.1763,
      "step": 1411
    },
    {
      "epoch": 0.18280091918309221,
      "grad_norm": 5.0167684023817225,
      "learning_rate": 1.224115714881688e-06,
      "loss": 4.2461,
      "step": 1412
    },
    {
      "epoch": 0.18293038159044567,
      "grad_norm": 5.604938238406695,
      "learning_rate": 1.2240352026081492e-06,
      "loss": 4.6792,
      "step": 1413
    },
    {
      "epoch": 0.18305984399779915,
      "grad_norm": 2.3286796253397615,
      "learning_rate": 1.2239545679695284e-06,
      "loss": 3.5737,
      "step": 1414
    },
    {
      "epoch": 0.1831893064051526,
      "grad_norm": 3.3881593882231855,
      "learning_rate": 1.2238738109822968e-06,
      "loss": 3.2437,
      "step": 1415
    },
    {
      "epoch": 0.1831893064051526,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.357865810394287,
      "eval_runtime": 15.073,
      "eval_samples_per_second": 2.919,
      "eval_steps_per_second": 0.199,
      "step": 1415
    },
    {
      "epoch": 0.18331876881250606,
      "grad_norm": 2.8725253858671307,
      "learning_rate": 1.223792931662951e-06,
      "loss": 3.8125,
      "step": 1416
    },
    {
      "epoch": 0.18344823121985954,
      "grad_norm": 4.2082521312796075,
      "learning_rate": 1.2237119300280117e-06,
      "loss": 3.8577,
      "step": 1417
    },
    {
      "epoch": 0.183577693627213,
      "grad_norm": 3.664366756612207,
      "learning_rate": 1.2236308060940253e-06,
      "loss": 3.8291,
      "step": 1418
    },
    {
      "epoch": 0.18370715603456647,
      "grad_norm": 2.8286714352794804,
      "learning_rate": 1.223549559877563e-06,
      "loss": 3.3262,
      "step": 1419
    },
    {
      "epoch": 0.18383661844191992,
      "grad_norm": 1.8582336316662238,
      "learning_rate": 1.2234681913952213e-06,
      "loss": 2.6143,
      "step": 1420
    },
    {
      "epoch": 0.18383661844191992,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3638583421707153,
      "eval_runtime": 14.2613,
      "eval_samples_per_second": 3.085,
      "eval_steps_per_second": 0.21,
      "step": 1420
    },
    {
      "epoch": 0.1839660808492734,
      "grad_norm": 4.6996909869227075,
      "learning_rate": 1.223386700663621e-06,
      "loss": 3.7173,
      "step": 1421
    },
    {
      "epoch": 0.18409554325662686,
      "grad_norm": 4.496564894942089,
      "learning_rate": 1.2233050876994086e-06,
      "loss": 3.564,
      "step": 1422
    },
    {
      "epoch": 0.1842250056639803,
      "grad_norm": 1.581812736719712,
      "learning_rate": 1.2232233525192545e-06,
      "loss": 2.5459,
      "step": 1423
    },
    {
      "epoch": 0.1843544680713338,
      "grad_norm": 6.236373517235444,
      "learning_rate": 1.2231414951398554e-06,
      "loss": 3.665,
      "step": 1424
    },
    {
      "epoch": 0.18448393047868725,
      "grad_norm": 6.203911927277913,
      "learning_rate": 1.223059515577932e-06,
      "loss": 4.0923,
      "step": 1425
    },
    {
      "epoch": 0.18448393047868725,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3698064088821411,
      "eval_runtime": 14.1682,
      "eval_samples_per_second": 3.106,
      "eval_steps_per_second": 0.212,
      "step": 1425
    },
    {
      "epoch": 0.18461339288604073,
      "grad_norm": 3.4833382453688655,
      "learning_rate": 1.2229774138502306e-06,
      "loss": 3.0735,
      "step": 1426
    },
    {
      "epoch": 0.18474285529339418,
      "grad_norm": 4.347187809491616,
      "learning_rate": 1.222895189973522e-06,
      "loss": 3.8179,
      "step": 1427
    },
    {
      "epoch": 0.18487231770074763,
      "grad_norm": 3.476425995636057,
      "learning_rate": 1.222812843964602e-06,
      "loss": 3.5542,
      "step": 1428
    },
    {
      "epoch": 0.18500178010810112,
      "grad_norm": 4.092272962578211,
      "learning_rate": 1.2227303758402918e-06,
      "loss": 4.2007,
      "step": 1429
    },
    {
      "epoch": 0.18513124251545457,
      "grad_norm": 6.370381941248504,
      "learning_rate": 1.2226477856174366e-06,
      "loss": 4.5801,
      "step": 1430
    },
    {
      "epoch": 0.18513124251545457,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.369140625,
      "eval_runtime": 14.0312,
      "eval_samples_per_second": 3.136,
      "eval_steps_per_second": 0.214,
      "step": 1430
    },
    {
      "epoch": 0.18526070492280805,
      "grad_norm": 3.5251596519985045,
      "learning_rate": 1.2225650733129073e-06,
      "loss": 3.3674,
      "step": 1431
    },
    {
      "epoch": 0.1853901673301615,
      "grad_norm": 1.2361749130373705,
      "learning_rate": 1.2224822389435997e-06,
      "loss": 2.4565,
      "step": 1432
    },
    {
      "epoch": 0.18551962973751496,
      "grad_norm": 2.81013731681298,
      "learning_rate": 1.2223992825264347e-06,
      "loss": 3.1934,
      "step": 1433
    },
    {
      "epoch": 0.18564909214486844,
      "grad_norm": 3.483636615304478,
      "learning_rate": 1.2223162040783572e-06,
      "loss": 3.1353,
      "step": 1434
    },
    {
      "epoch": 0.1857785545522219,
      "grad_norm": 6.99999596571565,
      "learning_rate": 1.2222330036163381e-06,
      "loss": 4.8738,
      "step": 1435
    },
    {
      "epoch": 0.1857785545522219,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3704723119735718,
      "eval_runtime": 14.1881,
      "eval_samples_per_second": 3.101,
      "eval_steps_per_second": 0.211,
      "step": 1435
    },
    {
      "epoch": 0.18590801695957537,
      "grad_norm": 2.3749684960020843,
      "learning_rate": 1.2221496811573726e-06,
      "loss": 3.2864,
      "step": 1436
    },
    {
      "epoch": 0.18603747936692883,
      "grad_norm": 4.7779905519446935,
      "learning_rate": 1.222066236718481e-06,
      "loss": 3.647,
      "step": 1437
    },
    {
      "epoch": 0.18616694177428228,
      "grad_norm": 7.29843151157389,
      "learning_rate": 1.2219826703167087e-06,
      "loss": 3.9585,
      "step": 1438
    },
    {
      "epoch": 0.18629640418163576,
      "grad_norm": 2.57725108373773,
      "learning_rate": 1.2218989819691257e-06,
      "loss": 3.1404,
      "step": 1439
    },
    {
      "epoch": 0.18642586658898921,
      "grad_norm": 4.434566243366233,
      "learning_rate": 1.221815171692827e-06,
      "loss": 4.1277,
      "step": 1440
    },
    {
      "epoch": 0.18642586658898921,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3776633739471436,
      "eval_runtime": 14.7808,
      "eval_samples_per_second": 2.977,
      "eval_steps_per_second": 0.203,
      "step": 1440
    },
    {
      "epoch": 0.1865553289963427,
      "grad_norm": 3.5594124522576163,
      "learning_rate": 1.2217312395049324e-06,
      "loss": 3.4541,
      "step": 1441
    },
    {
      "epoch": 0.18668479140369615,
      "grad_norm": 6.615842610368984,
      "learning_rate": 1.2216471854225873e-06,
      "loss": 4.4895,
      "step": 1442
    },
    {
      "epoch": 0.18681425381104963,
      "grad_norm": 3.457320657983824,
      "learning_rate": 1.221563009462961e-06,
      "loss": 3.5063,
      "step": 1443
    },
    {
      "epoch": 0.18694371621840308,
      "grad_norm": 4.806707468572244,
      "learning_rate": 1.221478711643248e-06,
      "loss": 3.5095,
      "step": 1444
    },
    {
      "epoch": 0.18707317862575654,
      "grad_norm": 4.667653187215275,
      "learning_rate": 1.221394291980668e-06,
      "loss": 3.3647,
      "step": 1445
    },
    {
      "epoch": 0.18707317862575654,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3777521848678589,
      "eval_runtime": 14.2176,
      "eval_samples_per_second": 3.095,
      "eval_steps_per_second": 0.211,
      "step": 1445
    },
    {
      "epoch": 0.18720264103311002,
      "grad_norm": 2.3540701328498144,
      "learning_rate": 1.2213097504924659e-06,
      "loss": 3.0569,
      "step": 1446
    },
    {
      "epoch": 0.18733210344046347,
      "grad_norm": 3.9234725048015746,
      "learning_rate": 1.2212250871959106e-06,
      "loss": 3.4436,
      "step": 1447
    },
    {
      "epoch": 0.18746156584781695,
      "grad_norm": 6.268158860585557,
      "learning_rate": 1.221140302108296e-06,
      "loss": 4.4556,
      "step": 1448
    },
    {
      "epoch": 0.1875910282551704,
      "grad_norm": 5.735086994523188,
      "learning_rate": 1.2210553952469418e-06,
      "loss": 3.9031,
      "step": 1449
    },
    {
      "epoch": 0.18772049066252386,
      "grad_norm": 3.566209530027245,
      "learning_rate": 1.2209703666291912e-06,
      "loss": 3.6167,
      "step": 1450
    },
    {
      "epoch": 0.18772049066252386,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3832119703292847,
      "eval_runtime": 14.1238,
      "eval_samples_per_second": 3.115,
      "eval_steps_per_second": 0.212,
      "step": 1450
    },
    {
      "epoch": 0.18784995306987734,
      "grad_norm": 2.7759442862299575,
      "learning_rate": 1.2208852162724138e-06,
      "loss": 3.4282,
      "step": 1451
    },
    {
      "epoch": 0.1879794154772308,
      "grad_norm": 6.03578138417162,
      "learning_rate": 1.2207999441940028e-06,
      "loss": 5.1118,
      "step": 1452
    },
    {
      "epoch": 0.18810887788458427,
      "grad_norm": 5.045494044161035,
      "learning_rate": 1.2207145504113769e-06,
      "loss": 4.0557,
      "step": 1453
    },
    {
      "epoch": 0.18823834029193773,
      "grad_norm": 4.701816933182676,
      "learning_rate": 1.2206290349419795e-06,
      "loss": 3.6936,
      "step": 1454
    },
    {
      "epoch": 0.18836780269929118,
      "grad_norm": 3.737544974401831,
      "learning_rate": 1.220543397803279e-06,
      "loss": 3.4189,
      "step": 1455
    },
    {
      "epoch": 0.18836780269929118,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.388716220855713,
      "eval_runtime": 15.1962,
      "eval_samples_per_second": 2.895,
      "eval_steps_per_second": 0.197,
      "step": 1455
    },
    {
      "epoch": 0.18849726510664466,
      "grad_norm": 2.98202953594671,
      "learning_rate": 1.2204576390127681e-06,
      "loss": 3.0361,
      "step": 1456
    },
    {
      "epoch": 0.18862672751399812,
      "grad_norm": 4.190718268486176,
      "learning_rate": 1.2203717585879652e-06,
      "loss": 3.7644,
      "step": 1457
    },
    {
      "epoch": 0.1887561899213516,
      "grad_norm": 4.759119772432647,
      "learning_rate": 1.220285756546413e-06,
      "loss": 3.416,
      "step": 1458
    },
    {
      "epoch": 0.18888565232870505,
      "grad_norm": 9.405091538903855,
      "learning_rate": 1.2201996329056793e-06,
      "loss": 5.927,
      "step": 1459
    },
    {
      "epoch": 0.1890151147360585,
      "grad_norm": 4.871610230477146,
      "learning_rate": 1.220113387683356e-06,
      "loss": 4.231,
      "step": 1460
    },
    {
      "epoch": 0.1890151147360585,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.391157627105713,
      "eval_runtime": 14.3757,
      "eval_samples_per_second": 3.061,
      "eval_steps_per_second": 0.209,
      "step": 1460
    },
    {
      "epoch": 0.18914457714341198,
      "grad_norm": 3.093044372326429,
      "learning_rate": 1.2200270208970614e-06,
      "loss": 2.948,
      "step": 1461
    },
    {
      "epoch": 0.18927403955076544,
      "grad_norm": 9.632522186253102,
      "learning_rate": 1.2199405325644367e-06,
      "loss": 6.4829,
      "step": 1462
    },
    {
      "epoch": 0.18940350195811892,
      "grad_norm": 2.4259934682234907,
      "learning_rate": 1.2198539227031495e-06,
      "loss": 3.144,
      "step": 1463
    },
    {
      "epoch": 0.18953296436547237,
      "grad_norm": 2.0793635957219996,
      "learning_rate": 1.2197671913308914e-06,
      "loss": 2.9604,
      "step": 1464
    },
    {
      "epoch": 0.18966242677282585,
      "grad_norm": 3.162156193891948,
      "learning_rate": 1.2196803384653792e-06,
      "loss": 3.2866,
      "step": 1465
    },
    {
      "epoch": 0.18966242677282585,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.3975497484207153,
      "eval_runtime": 15.0403,
      "eval_samples_per_second": 2.925,
      "eval_steps_per_second": 0.199,
      "step": 1465
    },
    {
      "epoch": 0.1897918891801793,
      "grad_norm": 3.3374046567024465,
      "learning_rate": 1.219593364124354e-06,
      "loss": 3.667,
      "step": 1466
    },
    {
      "epoch": 0.18992135158753276,
      "grad_norm": 1.8849838522909952,
      "learning_rate": 1.2195062683255825e-06,
      "loss": 2.9072,
      "step": 1467
    },
    {
      "epoch": 0.19005081399488624,
      "grad_norm": 1.5145599328612633,
      "learning_rate": 1.2194190510868555e-06,
      "loss": 2.8418,
      "step": 1468
    },
    {
      "epoch": 0.1901802764022397,
      "grad_norm": 4.268043726959668,
      "learning_rate": 1.2193317124259894e-06,
      "loss": 3.79,
      "step": 1469
    },
    {
      "epoch": 0.19030973880959318,
      "grad_norm": 5.7488014538831225,
      "learning_rate": 1.2192442523608239e-06,
      "loss": 4.5181,
      "step": 1470
    },
    {
      "epoch": 0.19030973880959318,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.400390625,
      "eval_runtime": 14.9333,
      "eval_samples_per_second": 2.946,
      "eval_steps_per_second": 0.201,
      "step": 1470
    },
    {
      "epoch": 0.19043920121694663,
      "grad_norm": 3.2764681327677074,
      "learning_rate": 1.2191566709092253e-06,
      "loss": 3.1045,
      "step": 1471
    },
    {
      "epoch": 0.19056866362430008,
      "grad_norm": 4.4796397902105625,
      "learning_rate": 1.2190689680890837e-06,
      "loss": 3.4404,
      "step": 1472
    },
    {
      "epoch": 0.19069812603165356,
      "grad_norm": 4.800960715818861,
      "learning_rate": 1.2189811439183141e-06,
      "loss": 3.9375,
      "step": 1473
    },
    {
      "epoch": 0.19082758843900702,
      "grad_norm": 3.6856454316834553,
      "learning_rate": 1.2188931984148566e-06,
      "loss": 2.9868,
      "step": 1474
    },
    {
      "epoch": 0.1909570508463605,
      "grad_norm": 4.624893420071821,
      "learning_rate": 1.2188051315966754e-06,
      "loss": 3.5354,
      "step": 1475
    },
    {
      "epoch": 0.1909570508463605,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4031428098678589,
      "eval_runtime": 16.5659,
      "eval_samples_per_second": 2.656,
      "eval_steps_per_second": 0.181,
      "step": 1475
    },
    {
      "epoch": 0.19108651325371395,
      "grad_norm": 2.4215697306628003,
      "learning_rate": 1.2187169434817604e-06,
      "loss": 3.3665,
      "step": 1476
    },
    {
      "epoch": 0.1912159756610674,
      "grad_norm": 3.09264095899073,
      "learning_rate": 1.2186286340881253e-06,
      "loss": 3.5483,
      "step": 1477
    },
    {
      "epoch": 0.1913454380684209,
      "grad_norm": 4.027535046427222,
      "learning_rate": 1.2185402034338095e-06,
      "loss": 3.4644,
      "step": 1478
    },
    {
      "epoch": 0.19147490047577434,
      "grad_norm": 3.351719821789138,
      "learning_rate": 1.218451651536877e-06,
      "loss": 3.1255,
      "step": 1479
    },
    {
      "epoch": 0.19160436288312782,
      "grad_norm": 5.150281941211811,
      "learning_rate": 1.2183629784154155e-06,
      "loss": 3.9524,
      "step": 1480
    },
    {
      "epoch": 0.19160436288312782,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4035866260528564,
      "eval_runtime": 18.7342,
      "eval_samples_per_second": 2.349,
      "eval_steps_per_second": 0.16,
      "step": 1480
    },
    {
      "epoch": 0.19173382529048127,
      "grad_norm": 2.517345404719067,
      "learning_rate": 1.218274184087539e-06,
      "loss": 3.2493,
      "step": 1481
    },
    {
      "epoch": 0.19186328769783473,
      "grad_norm": 3.244307149627311,
      "learning_rate": 1.218185268571385e-06,
      "loss": 3.4783,
      "step": 1482
    },
    {
      "epoch": 0.1919927501051882,
      "grad_norm": 2.058493891977264,
      "learning_rate": 1.2180962318851168e-06,
      "loss": 2.584,
      "step": 1483
    },
    {
      "epoch": 0.19212221251254166,
      "grad_norm": 3.7502324472681434,
      "learning_rate": 1.2180070740469217e-06,
      "loss": 3.4126,
      "step": 1484
    },
    {
      "epoch": 0.19225167491989514,
      "grad_norm": 3.784012039523587,
      "learning_rate": 1.2179177950750118e-06,
      "loss": 3.4277,
      "step": 1485
    },
    {
      "epoch": 0.19225167491989514,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4086025953292847,
      "eval_runtime": 14.3657,
      "eval_samples_per_second": 3.063,
      "eval_steps_per_second": 0.209,
      "step": 1485
    },
    {
      "epoch": 0.1923811373272486,
      "grad_norm": 3.2291766620010343,
      "learning_rate": 1.2178283949876246e-06,
      "loss": 3.4443,
      "step": 1486
    },
    {
      "epoch": 0.19251059973460208,
      "grad_norm": 4.61459480144291,
      "learning_rate": 1.2177388738030216e-06,
      "loss": 3.6772,
      "step": 1487
    },
    {
      "epoch": 0.19264006214195553,
      "grad_norm": 3.5463107007493044,
      "learning_rate": 1.2176492315394894e-06,
      "loss": 3.6357,
      "step": 1488
    },
    {
      "epoch": 0.19276952454930898,
      "grad_norm": 4.270266157919265,
      "learning_rate": 1.217559468215339e-06,
      "loss": 3.9058,
      "step": 1489
    },
    {
      "epoch": 0.19289898695666247,
      "grad_norm": 7.575439693616707,
      "learning_rate": 1.217469583848907e-06,
      "loss": 4.4536,
      "step": 1490
    },
    {
      "epoch": 0.19289898695666247,
      "eval_PRM Accuracy": 0.2641509433962264,
      "eval_PRM F1": 0.11363636363636363,
      "eval_PRM F1 AUC": 0.5301204819277109,
      "eval_PRM F1 Neg": 0.3709677419354839,
      "eval_PRM NPV": 0.22772277227722773,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.060240963855421686,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4122425317764282,
      "eval_runtime": 16.104,
      "eval_samples_per_second": 2.732,
      "eval_steps_per_second": 0.186,
      "step": 1490
    },
    {
      "epoch": 0.19302844936401592,
      "grad_norm": 2.972792525981474,
      "learning_rate": 1.2173795784585534e-06,
      "loss": 3.1672,
      "step": 1491
    },
    {
      "epoch": 0.1931579117713694,
      "grad_norm": 2.3418243128705636,
      "learning_rate": 1.217289452062664e-06,
      "loss": 3.063,
      "step": 1492
    },
    {
      "epoch": 0.19328737417872285,
      "grad_norm": 5.567828310167252,
      "learning_rate": 1.217199204679649e-06,
      "loss": 4.2148,
      "step": 1493
    },
    {
      "epoch": 0.1934168365860763,
      "grad_norm": 4.261413175583276,
      "learning_rate": 1.2171088363279426e-06,
      "loss": 3.416,
      "step": 1494
    },
    {
      "epoch": 0.1935462989934298,
      "grad_norm": 3.3535930303205053,
      "learning_rate": 1.2170183470260053e-06,
      "loss": 3.0527,
      "step": 1495
    },
    {
      "epoch": 0.1935462989934298,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.415971279144287,
      "eval_runtime": 15.1716,
      "eval_samples_per_second": 2.9,
      "eval_steps_per_second": 0.198,
      "step": 1495
    },
    {
      "epoch": 0.19367576140078324,
      "grad_norm": 1.5289685576856569,
      "learning_rate": 1.2169277367923207e-06,
      "loss": 2.6704,
      "step": 1496
    },
    {
      "epoch": 0.19380522380813672,
      "grad_norm": 7.267548086395574,
      "learning_rate": 1.2168370056453983e-06,
      "loss": 4.4214,
      "step": 1497
    },
    {
      "epoch": 0.19393468621549018,
      "grad_norm": 4.30219798549593,
      "learning_rate": 1.2167461536037712e-06,
      "loss": 3.2427,
      "step": 1498
    },
    {
      "epoch": 0.19406414862284363,
      "grad_norm": 2.676643384513489,
      "learning_rate": 1.2166551806859984e-06,
      "loss": 3.0469,
      "step": 1499
    },
    {
      "epoch": 0.1941936110301971,
      "grad_norm": 7.036672441155513,
      "learning_rate": 1.2165640869106624e-06,
      "loss": 5.48,
      "step": 1500
    },
    {
      "epoch": 0.1941936110301971,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4214755296707153,
      "eval_runtime": 13.9305,
      "eval_samples_per_second": 3.159,
      "eval_steps_per_second": 0.215,
      "step": 1500
    },
    {
      "epoch": 0.19432307343755056,
      "grad_norm": 4.845419134564698,
      "learning_rate": 1.2164728722963713e-06,
      "loss": 3.8643,
      "step": 1501
    },
    {
      "epoch": 0.19445253584490405,
      "grad_norm": 3.995491295891669,
      "learning_rate": 1.2163815368617574e-06,
      "loss": 3.0723,
      "step": 1502
    },
    {
      "epoch": 0.1945819982522575,
      "grad_norm": 3.471203592316862,
      "learning_rate": 1.2162900806254778e-06,
      "loss": 2.9028,
      "step": 1503
    },
    {
      "epoch": 0.19471146065961095,
      "grad_norm": 3.8927115634833935,
      "learning_rate": 1.2161985036062146e-06,
      "loss": 3.6755,
      "step": 1504
    },
    {
      "epoch": 0.19484092306696443,
      "grad_norm": 3.66897107403486,
      "learning_rate": 1.2161068058226737e-06,
      "loss": 3.8296,
      "step": 1505
    },
    {
      "epoch": 0.19484092306696443,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4242275953292847,
      "eval_runtime": 14.1491,
      "eval_samples_per_second": 3.11,
      "eval_steps_per_second": 0.212,
      "step": 1505
    },
    {
      "epoch": 0.1949703854743179,
      "grad_norm": 3.945085191116555,
      "learning_rate": 1.2160149872935866e-06,
      "loss": 3.3855,
      "step": 1506
    },
    {
      "epoch": 0.19509984788167137,
      "grad_norm": 3.2532420246187286,
      "learning_rate": 1.215923048037709e-06,
      "loss": 3.2073,
      "step": 1507
    },
    {
      "epoch": 0.19522931028902482,
      "grad_norm": 2.371061160407878,
      "learning_rate": 1.2158309880738215e-06,
      "loss": 3.1289,
      "step": 1508
    },
    {
      "epoch": 0.1953587726963783,
      "grad_norm": 2.8974571435739374,
      "learning_rate": 1.2157388074207288e-06,
      "loss": 3.1313,
      "step": 1509
    },
    {
      "epoch": 0.19548823510373176,
      "grad_norm": 1.6210041807318014,
      "learning_rate": 1.2156465060972612e-06,
      "loss": 2.6797,
      "step": 1510
    },
    {
      "epoch": 0.19548823510373176,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.426313877105713,
      "eval_runtime": 15.0598,
      "eval_samples_per_second": 2.922,
      "eval_steps_per_second": 0.199,
      "step": 1510
    },
    {
      "epoch": 0.1956176975110852,
      "grad_norm": 4.338336807951537,
      "learning_rate": 1.2155540841222727e-06,
      "loss": 3.9966,
      "step": 1511
    },
    {
      "epoch": 0.1957471599184387,
      "grad_norm": 4.199510770528302,
      "learning_rate": 1.2154615415146428e-06,
      "loss": 3.0269,
      "step": 1512
    },
    {
      "epoch": 0.19587662232579214,
      "grad_norm": 3.933256709381275,
      "learning_rate": 1.2153688782932746e-06,
      "loss": 3.9829,
      "step": 1513
    },
    {
      "epoch": 0.19600608473314562,
      "grad_norm": 6.708972042680103,
      "learning_rate": 1.215276094477097e-06,
      "loss": 3.9902,
      "step": 1514
    },
    {
      "epoch": 0.19613554714049908,
      "grad_norm": 2.981069390796975,
      "learning_rate": 1.2151831900850627e-06,
      "loss": 3.3745,
      "step": 1515
    },
    {
      "epoch": 0.19613554714049908,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4276012182235718,
      "eval_runtime": 13.7181,
      "eval_samples_per_second": 3.207,
      "eval_steps_per_second": 0.219,
      "step": 1515
    },
    {
      "epoch": 0.19626500954785253,
      "grad_norm": 2.7138718303579954,
      "learning_rate": 1.2150901651361493e-06,
      "loss": 3.4946,
      "step": 1516
    },
    {
      "epoch": 0.196394471955206,
      "grad_norm": 4.580876306704034,
      "learning_rate": 1.2149970196493593e-06,
      "loss": 3.4609,
      "step": 1517
    },
    {
      "epoch": 0.19652393436255947,
      "grad_norm": 4.094165253293919,
      "learning_rate": 1.2149037536437192e-06,
      "loss": 3.4507,
      "step": 1518
    },
    {
      "epoch": 0.19665339676991295,
      "grad_norm": 4.0938811458730475,
      "learning_rate": 1.2148103671382805e-06,
      "loss": 3.7017,
      "step": 1519
    },
    {
      "epoch": 0.1967828591772664,
      "grad_norm": 6.849544131975171,
      "learning_rate": 1.2147168601521194e-06,
      "loss": 4.6694,
      "step": 1520
    },
    {
      "epoch": 0.1967828591772664,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4351917505264282,
      "eval_runtime": 14.7195,
      "eval_samples_per_second": 2.989,
      "eval_steps_per_second": 0.204,
      "step": 1520
    },
    {
      "epoch": 0.19691232158461985,
      "grad_norm": 2.862944819638343,
      "learning_rate": 1.2146232327043367e-06,
      "loss": 3.3704,
      "step": 1521
    },
    {
      "epoch": 0.19704178399197333,
      "grad_norm": 6.377635113912191,
      "learning_rate": 1.2145294848140577e-06,
      "loss": 4.2241,
      "step": 1522
    },
    {
      "epoch": 0.1971712463993268,
      "grad_norm": 7.743289927129237,
      "learning_rate": 1.214435616500432e-06,
      "loss": 4.8267,
      "step": 1523
    },
    {
      "epoch": 0.19730070880668027,
      "grad_norm": 2.641206070102616,
      "learning_rate": 1.2143416277826344e-06,
      "loss": 3.4771,
      "step": 1524
    },
    {
      "epoch": 0.19743017121403372,
      "grad_norm": 3.197824623342946,
      "learning_rate": 1.214247518679864e-06,
      "loss": 3.0527,
      "step": 1525
    },
    {
      "epoch": 0.19743017121403372,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4392756223678589,
      "eval_runtime": 14.4369,
      "eval_samples_per_second": 3.048,
      "eval_steps_per_second": 0.208,
      "step": 1525
    },
    {
      "epoch": 0.19755963362138718,
      "grad_norm": 2.902892943763722,
      "learning_rate": 1.2141532892113443e-06,
      "loss": 3.0239,
      "step": 1526
    },
    {
      "epoch": 0.19768909602874066,
      "grad_norm": 3.146768487120522,
      "learning_rate": 1.2140589393963237e-06,
      "loss": 3.2231,
      "step": 1527
    },
    {
      "epoch": 0.1978185584360941,
      "grad_norm": 7.572701064560839,
      "learning_rate": 1.213964469254075e-06,
      "loss": 4.6699,
      "step": 1528
    },
    {
      "epoch": 0.1979480208434476,
      "grad_norm": 7.8696216770995235,
      "learning_rate": 1.2138698788038958e-06,
      "loss": 4.4543,
      "step": 1529
    },
    {
      "epoch": 0.19807748325080105,
      "grad_norm": 2.1557226699300807,
      "learning_rate": 1.213775168065108e-06,
      "loss": 3.0369,
      "step": 1530
    },
    {
      "epoch": 0.19807748325080105,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.444868564605713,
      "eval_runtime": 15.4519,
      "eval_samples_per_second": 2.848,
      "eval_steps_per_second": 0.194,
      "step": 1530
    },
    {
      "epoch": 0.19820694565815453,
      "grad_norm": 2.3016394267085936,
      "learning_rate": 1.2136803370570583e-06,
      "loss": 3.012,
      "step": 1531
    },
    {
      "epoch": 0.19833640806550798,
      "grad_norm": 3.189583637555812,
      "learning_rate": 1.2135853857991175e-06,
      "loss": 3.2603,
      "step": 1532
    },
    {
      "epoch": 0.19846587047286143,
      "grad_norm": 4.261435582477583,
      "learning_rate": 1.2134903143106816e-06,
      "loss": 3.7969,
      "step": 1533
    },
    {
      "epoch": 0.19859533288021491,
      "grad_norm": 2.3626151320962174,
      "learning_rate": 1.2133951226111709e-06,
      "loss": 2.7866,
      "step": 1534
    },
    {
      "epoch": 0.19872479528756837,
      "grad_norm": 2.6919069476152626,
      "learning_rate": 1.2132998107200304e-06,
      "loss": 2.6282,
      "step": 1535
    },
    {
      "epoch": 0.19872479528756837,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.445845127105713,
      "eval_runtime": 15.099,
      "eval_samples_per_second": 2.914,
      "eval_steps_per_second": 0.199,
      "step": 1535
    },
    {
      "epoch": 0.19885425769492185,
      "grad_norm": 3.305444165670336,
      "learning_rate": 1.213204378656729e-06,
      "loss": 3.1577,
      "step": 1536
    },
    {
      "epoch": 0.1989837201022753,
      "grad_norm": 4.003643969235668,
      "learning_rate": 1.213108826440761e-06,
      "loss": 2.9624,
      "step": 1537
    },
    {
      "epoch": 0.19911318250962876,
      "grad_norm": 1.6570602475458145,
      "learning_rate": 1.2130131540916447e-06,
      "loss": 2.4124,
      "step": 1538
    },
    {
      "epoch": 0.19924264491698224,
      "grad_norm": 2.665600868005522,
      "learning_rate": 1.2129173616289233e-06,
      "loss": 3.3452,
      "step": 1539
    },
    {
      "epoch": 0.1993721073243357,
      "grad_norm": 4.039641609951077,
      "learning_rate": 1.212821449072164e-06,
      "loss": 3.7769,
      "step": 1540
    },
    {
      "epoch": 0.1993721073243357,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4509055614471436,
      "eval_runtime": 20.7699,
      "eval_samples_per_second": 2.118,
      "eval_steps_per_second": 0.144,
      "step": 1540
    },
    {
      "epoch": 0.19950156973168917,
      "grad_norm": 5.880390778230704,
      "learning_rate": 1.2127254164409593e-06,
      "loss": 3.7339,
      "step": 1541
    },
    {
      "epoch": 0.19963103213904262,
      "grad_norm": 7.626853983434958,
      "learning_rate": 1.212629263754926e-06,
      "loss": 4.4692,
      "step": 1542
    },
    {
      "epoch": 0.19976049454639608,
      "grad_norm": 1.581514636862654,
      "learning_rate": 1.2125329910337045e-06,
      "loss": 2.6382,
      "step": 1543
    },
    {
      "epoch": 0.19988995695374956,
      "grad_norm": 11.915379122547142,
      "learning_rate": 1.2124365982969607e-06,
      "loss": 6.1772,
      "step": 1544
    },
    {
      "epoch": 0.200019419361103,
      "grad_norm": 3.609418599267737,
      "learning_rate": 1.212340085564385e-06,
      "loss": 2.9995,
      "step": 1545
    },
    {
      "epoch": 0.200019419361103,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4520596265792847,
      "eval_runtime": 15.213,
      "eval_samples_per_second": 2.892,
      "eval_steps_per_second": 0.197,
      "step": 1545
    },
    {
      "epoch": 0.2001488817684565,
      "grad_norm": 5.78151162096917,
      "learning_rate": 1.212243452855692e-06,
      "loss": 4.0293,
      "step": 1546
    },
    {
      "epoch": 0.20027834417580995,
      "grad_norm": 2.937787655572086,
      "learning_rate": 1.212146700190621e-06,
      "loss": 3.1235,
      "step": 1547
    },
    {
      "epoch": 0.2004078065831634,
      "grad_norm": 8.637192704143201,
      "learning_rate": 1.2120498275889357e-06,
      "loss": 4.1572,
      "step": 1548
    },
    {
      "epoch": 0.20053726899051688,
      "grad_norm": 2.5554512529276026,
      "learning_rate": 1.2119528350704237e-06,
      "loss": 3.1335,
      "step": 1549
    },
    {
      "epoch": 0.20066673139787033,
      "grad_norm": 5.741191168988011,
      "learning_rate": 1.2118557226548984e-06,
      "loss": 4.103,
      "step": 1550
    },
    {
      "epoch": 0.20066673139787033,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4527698755264282,
      "eval_runtime": 22.2949,
      "eval_samples_per_second": 1.974,
      "eval_steps_per_second": 0.135,
      "step": 1550
    },
    {
      "epoch": 0.20079619380522382,
      "grad_norm": 3.087974490934094,
      "learning_rate": 1.2117584903621968e-06,
      "loss": 3.1123,
      "step": 1551
    },
    {
      "epoch": 0.20092565621257727,
      "grad_norm": 3.9529421240056046,
      "learning_rate": 1.2116611382121803e-06,
      "loss": 3.6387,
      "step": 1552
    },
    {
      "epoch": 0.20105511861993075,
      "grad_norm": 6.235812177458835,
      "learning_rate": 1.2115636662247355e-06,
      "loss": 4.0176,
      "step": 1553
    },
    {
      "epoch": 0.2011845810272842,
      "grad_norm": 6.410729264195401,
      "learning_rate": 1.2114660744197725e-06,
      "loss": 4.2646,
      "step": 1554
    },
    {
      "epoch": 0.20131404343463766,
      "grad_norm": 4.9644377007051705,
      "learning_rate": 1.2113683628172269e-06,
      "loss": 3.7329,
      "step": 1555
    },
    {
      "epoch": 0.20131404343463766,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4598721265792847,
      "eval_runtime": 13.6725,
      "eval_samples_per_second": 3.218,
      "eval_steps_per_second": 0.219,
      "step": 1555
    },
    {
      "epoch": 0.20144350584199114,
      "grad_norm": 2.853664447998761,
      "learning_rate": 1.2112705314370579e-06,
      "loss": 3.3838,
      "step": 1556
    },
    {
      "epoch": 0.2015729682493446,
      "grad_norm": 3.0798359641816684,
      "learning_rate": 1.2111725802992498e-06,
      "loss": 3.4993,
      "step": 1557
    },
    {
      "epoch": 0.20170243065669807,
      "grad_norm": 1.9122865891922844,
      "learning_rate": 1.2110745094238107e-06,
      "loss": 2.9854,
      "step": 1558
    },
    {
      "epoch": 0.20183189306405153,
      "grad_norm": 1.8144191675334405,
      "learning_rate": 1.2109763188307741e-06,
      "loss": 2.9585,
      "step": 1559
    },
    {
      "epoch": 0.20196135547140498,
      "grad_norm": 3.408868530623044,
      "learning_rate": 1.2108780085401971e-06,
      "loss": 3.6914,
      "step": 1560
    },
    {
      "epoch": 0.20196135547140498,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4631569385528564,
      "eval_runtime": 15.4013,
      "eval_samples_per_second": 2.857,
      "eval_steps_per_second": 0.195,
      "step": 1560
    },
    {
      "epoch": 0.20209081787875846,
      "grad_norm": 2.6121158439013947,
      "learning_rate": 1.2107795785721618e-06,
      "loss": 3.2236,
      "step": 1561
    },
    {
      "epoch": 0.20222028028611191,
      "grad_norm": 2.4932650882454106,
      "learning_rate": 1.2106810289467742e-06,
      "loss": 3.166,
      "step": 1562
    },
    {
      "epoch": 0.2023497426934654,
      "grad_norm": 2.703922425558448,
      "learning_rate": 1.2105823596841654e-06,
      "loss": 3.6685,
      "step": 1563
    },
    {
      "epoch": 0.20247920510081885,
      "grad_norm": 1.9224382757041614,
      "learning_rate": 1.2104835708044903e-06,
      "loss": 2.8481,
      "step": 1564
    },
    {
      "epoch": 0.2026086675081723,
      "grad_norm": 2.905178727106571,
      "learning_rate": 1.2103846623279285e-06,
      "loss": 2.8643,
      "step": 1565
    },
    {
      "epoch": 0.2026086675081723,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4644886255264282,
      "eval_runtime": 14.9309,
      "eval_samples_per_second": 2.947,
      "eval_steps_per_second": 0.201,
      "step": 1565
    },
    {
      "epoch": 0.20273812991552578,
      "grad_norm": 1.8911415277885153,
      "learning_rate": 1.2102856342746846e-06,
      "loss": 2.9456,
      "step": 1566
    },
    {
      "epoch": 0.20286759232287924,
      "grad_norm": 1.926688665977454,
      "learning_rate": 1.2101864866649863e-06,
      "loss": 3.0642,
      "step": 1567
    },
    {
      "epoch": 0.20299705473023272,
      "grad_norm": 4.152946729181612,
      "learning_rate": 1.2100872195190872e-06,
      "loss": 3.353,
      "step": 1568
    },
    {
      "epoch": 0.20312651713758617,
      "grad_norm": 4.275656626955722,
      "learning_rate": 1.2099878328572642e-06,
      "loss": 3.731,
      "step": 1569
    },
    {
      "epoch": 0.20325597954493962,
      "grad_norm": 1.7459558279486687,
      "learning_rate": 1.2098883266998193e-06,
      "loss": 2.6465,
      "step": 1570
    },
    {
      "epoch": 0.20325597954493962,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.468794345855713,
      "eval_runtime": 15.2123,
      "eval_samples_per_second": 2.892,
      "eval_steps_per_second": 0.197,
      "step": 1570
    },
    {
      "epoch": 0.2033854419522931,
      "grad_norm": 4.85287549461254,
      "learning_rate": 1.2097887010670783e-06,
      "loss": 3.8296,
      "step": 1571
    },
    {
      "epoch": 0.20351490435964656,
      "grad_norm": 4.527523924353318,
      "learning_rate": 1.2096889559793921e-06,
      "loss": 4.27,
      "step": 1572
    },
    {
      "epoch": 0.20364436676700004,
      "grad_norm": 4.865976890858769,
      "learning_rate": 1.2095890914571356e-06,
      "loss": 3.9246,
      "step": 1573
    },
    {
      "epoch": 0.2037738291743535,
      "grad_norm": 2.8317211998897176,
      "learning_rate": 1.209489107520708e-06,
      "loss": 2.9395,
      "step": 1574
    },
    {
      "epoch": 0.20390329158170697,
      "grad_norm": 2.3545586555200737,
      "learning_rate": 1.2093890041905332e-06,
      "loss": 3.0503,
      "step": 1575
    },
    {
      "epoch": 0.20390329158170697,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4699928760528564,
      "eval_runtime": 14.4633,
      "eval_samples_per_second": 3.042,
      "eval_steps_per_second": 0.207,
      "step": 1575
    },
    {
      "epoch": 0.20403275398906043,
      "grad_norm": 5.7779708346918035,
      "learning_rate": 1.2092887814870593e-06,
      "loss": 3.9587,
      "step": 1576
    },
    {
      "epoch": 0.20416221639641388,
      "grad_norm": 2.4822541762284733,
      "learning_rate": 1.2091884394307587e-06,
      "loss": 2.8979,
      "step": 1577
    },
    {
      "epoch": 0.20429167880376736,
      "grad_norm": 5.3614084166007006,
      "learning_rate": 1.2090879780421282e-06,
      "loss": 3.2603,
      "step": 1578
    },
    {
      "epoch": 0.20442114121112082,
      "grad_norm": 3.546560236927702,
      "learning_rate": 1.2089873973416896e-06,
      "loss": 3.2593,
      "step": 1579
    },
    {
      "epoch": 0.2045506036184743,
      "grad_norm": 2.640294338800154,
      "learning_rate": 1.2088866973499882e-06,
      "loss": 3.1023,
      "step": 1580
    },
    {
      "epoch": 0.2045506036184743,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4705699682235718,
      "eval_runtime": 14.607,
      "eval_samples_per_second": 3.012,
      "eval_steps_per_second": 0.205,
      "step": 1580
    },
    {
      "epoch": 0.20468006602582775,
      "grad_norm": 5.66685869149829,
      "learning_rate": 1.2087858780875938e-06,
      "loss": 4.0166,
      "step": 1581
    },
    {
      "epoch": 0.2048095284331812,
      "grad_norm": 2.334032769188969,
      "learning_rate": 1.2086849395751011e-06,
      "loss": 2.9531,
      "step": 1582
    },
    {
      "epoch": 0.20493899084053468,
      "grad_norm": 3.1929631814622605,
      "learning_rate": 1.208583881833129e-06,
      "loss": 3.574,
      "step": 1583
    },
    {
      "epoch": 0.20506845324788814,
      "grad_norm": 8.522138757685294,
      "learning_rate": 1.2084827048823198e-06,
      "loss": 4.2388,
      "step": 1584
    },
    {
      "epoch": 0.20519791565524162,
      "grad_norm": 2.1435488667101383,
      "learning_rate": 1.2083814087433418e-06,
      "loss": 2.689,
      "step": 1585
    },
    {
      "epoch": 0.20519791565524162,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4782493114471436,
      "eval_runtime": 14.5609,
      "eval_samples_per_second": 3.022,
      "eval_steps_per_second": 0.206,
      "step": 1585
    },
    {
      "epoch": 0.20532737806259507,
      "grad_norm": 2.892395805151098,
      "learning_rate": 1.2082799934368864e-06,
      "loss": 3.0242,
      "step": 1586
    },
    {
      "epoch": 0.20545684046994853,
      "grad_norm": 2.522406926650977,
      "learning_rate": 1.2081784589836699e-06,
      "loss": 3.1038,
      "step": 1587
    },
    {
      "epoch": 0.205586302877302,
      "grad_norm": 5.181832917055719,
      "learning_rate": 1.2080768054044324e-06,
      "loss": 4.0276,
      "step": 1588
    },
    {
      "epoch": 0.20571576528465546,
      "grad_norm": 3.348329622132548,
      "learning_rate": 1.207975032719939e-06,
      "loss": 3.0703,
      "step": 1589
    },
    {
      "epoch": 0.20584522769200894,
      "grad_norm": 2.065729560958714,
      "learning_rate": 1.2078731409509792e-06,
      "loss": 2.8737,
      "step": 1590
    },
    {
      "epoch": 0.20584522769200894,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4801136255264282,
      "eval_runtime": 14.1584,
      "eval_samples_per_second": 3.108,
      "eval_steps_per_second": 0.212,
      "step": 1590
    },
    {
      "epoch": 0.2059746900993624,
      "grad_norm": 2.6355991484725743,
      "learning_rate": 1.2077711301183656e-06,
      "loss": 2.8242,
      "step": 1591
    },
    {
      "epoch": 0.20610415250671585,
      "grad_norm": 4.223252780591359,
      "learning_rate": 1.2076690002429368e-06,
      "loss": 3.6353,
      "step": 1592
    },
    {
      "epoch": 0.20623361491406933,
      "grad_norm": 4.212169138372161,
      "learning_rate": 1.2075667513455542e-06,
      "loss": 3.6245,
      "step": 1593
    },
    {
      "epoch": 0.20636307732142278,
      "grad_norm": 6.096225964591917,
      "learning_rate": 1.2074643834471048e-06,
      "loss": 4.0625,
      "step": 1594
    },
    {
      "epoch": 0.20649253972877626,
      "grad_norm": 2.401444677656407,
      "learning_rate": 1.207361896568499e-06,
      "loss": 3.3516,
      "step": 1595
    },
    {
      "epoch": 0.20649253972877626,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4830876588821411,
      "eval_runtime": 19.1139,
      "eval_samples_per_second": 2.302,
      "eval_steps_per_second": 0.157,
      "step": 1595
    },
    {
      "epoch": 0.20662200213612972,
      "grad_norm": 4.01192822886966,
      "learning_rate": 1.2072592907306718e-06,
      "loss": 3.1951,
      "step": 1596
    },
    {
      "epoch": 0.2067514645434832,
      "grad_norm": 1.6010028669109733,
      "learning_rate": 1.2071565659545826e-06,
      "loss": 2.8215,
      "step": 1597
    },
    {
      "epoch": 0.20688092695083665,
      "grad_norm": 5.6261882001825905,
      "learning_rate": 1.2070537222612153e-06,
      "loss": 3.936,
      "step": 1598
    },
    {
      "epoch": 0.2070103893581901,
      "grad_norm": 2.675115238277584,
      "learning_rate": 1.2069507596715773e-06,
      "loss": 3.0876,
      "step": 1599
    },
    {
      "epoch": 0.2071398517655436,
      "grad_norm": 5.644124802210082,
      "learning_rate": 1.206847678206701e-06,
      "loss": 3.5344,
      "step": 1600
    },
    {
      "epoch": 0.2071398517655436,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4878817796707153,
      "eval_runtime": 14.9199,
      "eval_samples_per_second": 2.949,
      "eval_steps_per_second": 0.201,
      "step": 1600
    },
    {
      "epoch": 0.20726931417289704,
      "grad_norm": 3.059641611880596,
      "learning_rate": 1.206744477887643e-06,
      "loss": 3.1641,
      "step": 1601
    },
    {
      "epoch": 0.20739877658025052,
      "grad_norm": 1.5856627740509093,
      "learning_rate": 1.206641158735484e-06,
      "loss": 2.4263,
      "step": 1602
    },
    {
      "epoch": 0.20752823898760397,
      "grad_norm": 2.9880133530734776,
      "learning_rate": 1.2065377207713288e-06,
      "loss": 3.6694,
      "step": 1603
    },
    {
      "epoch": 0.20765770139495743,
      "grad_norm": 3.1302800518305873,
      "learning_rate": 1.2064341640163071e-06,
      "loss": 3.0247,
      "step": 1604
    },
    {
      "epoch": 0.2077871638023109,
      "grad_norm": 3.2710387355775454,
      "learning_rate": 1.2063304884915722e-06,
      "loss": 3.0269,
      "step": 1605
    },
    {
      "epoch": 0.2077871638023109,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4912109375,
      "eval_runtime": 14.9049,
      "eval_samples_per_second": 2.952,
      "eval_steps_per_second": 0.201,
      "step": 1605
    },
    {
      "epoch": 0.20791662620966436,
      "grad_norm": 4.946247369248822,
      "learning_rate": 1.206226694218302e-06,
      "loss": 3.6489,
      "step": 1606
    },
    {
      "epoch": 0.20804608861701784,
      "grad_norm": 2.5348021081816805,
      "learning_rate": 1.2061227812176986e-06,
      "loss": 3.1934,
      "step": 1607
    },
    {
      "epoch": 0.2081755510243713,
      "grad_norm": 3.2512636177454306,
      "learning_rate": 1.2060187495109884e-06,
      "loss": 3.3118,
      "step": 1608
    },
    {
      "epoch": 0.20830501343172475,
      "grad_norm": 6.142824959240283,
      "learning_rate": 1.2059145991194218e-06,
      "loss": 4.3799,
      "step": 1609
    },
    {
      "epoch": 0.20843447583907823,
      "grad_norm": 7.39309778427709,
      "learning_rate": 1.2058103300642737e-06,
      "loss": 4.1504,
      "step": 1610
    },
    {
      "epoch": 0.20843447583907823,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4938743114471436,
      "eval_runtime": 14.5684,
      "eval_samples_per_second": 3.02,
      "eval_steps_per_second": 0.206,
      "step": 1610
    },
    {
      "epoch": 0.20856393824643168,
      "grad_norm": 3.629887536139999,
      "learning_rate": 1.2057059423668433e-06,
      "loss": 3.2402,
      "step": 1611
    },
    {
      "epoch": 0.20869340065378517,
      "grad_norm": 5.163285639039155,
      "learning_rate": 1.2056014360484535e-06,
      "loss": 3.3267,
      "step": 1612
    },
    {
      "epoch": 0.20882286306113862,
      "grad_norm": 2.4021206836647178,
      "learning_rate": 1.2054968111304525e-06,
      "loss": 3.1245,
      "step": 1613
    },
    {
      "epoch": 0.20895232546849207,
      "grad_norm": 4.3005818277725485,
      "learning_rate": 1.2053920676342115e-06,
      "loss": 3.4814,
      "step": 1614
    },
    {
      "epoch": 0.20908178787584555,
      "grad_norm": 6.599324441815224,
      "learning_rate": 1.2052872055811268e-06,
      "loss": 4.0947,
      "step": 1615
    },
    {
      "epoch": 0.20908178787584555,
      "eval_PRM Accuracy": 0.25471698113207547,
      "eval_PRM F1": 0.09195402298850575,
      "eval_PRM F1 AUC": 0.5240963855421686,
      "eval_PRM F1 Neg": 0.368,
      "eval_PRM NPV": 0.22549019607843138,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.04819277108433735,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957829713821411,
      "eval_runtime": 15.0369,
      "eval_samples_per_second": 2.926,
      "eval_steps_per_second": 0.2,
      "step": 1615
    },
    {
      "epoch": 0.209211250283199,
      "grad_norm": 2.4757001552458333,
      "learning_rate": 1.2051822249926184e-06,
      "loss": 3.1318,
      "step": 1616
    },
    {
      "epoch": 0.2093407126905525,
      "grad_norm": 3.4909574273882766,
      "learning_rate": 1.2050771258901309e-06,
      "loss": 3.4346,
      "step": 1617
    },
    {
      "epoch": 0.20947017509790594,
      "grad_norm": 2.507228023621561,
      "learning_rate": 1.2049719082951326e-06,
      "loss": 2.9487,
      "step": 1618
    },
    {
      "epoch": 0.20959963750525942,
      "grad_norm": 3.015430572085104,
      "learning_rate": 1.2048665722291167e-06,
      "loss": 3.2109,
      "step": 1619
    },
    {
      "epoch": 0.20972909991261288,
      "grad_norm": 2.444443113855727,
      "learning_rate": 1.2047611177135998e-06,
      "loss": 2.8706,
      "step": 1620
    },
    {
      "epoch": 0.20972909991261288,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5016423463821411,
      "eval_runtime": 15.6167,
      "eval_samples_per_second": 2.817,
      "eval_steps_per_second": 0.192,
      "step": 1620
    },
    {
      "epoch": 0.20985856231996633,
      "grad_norm": 5.684682992040662,
      "learning_rate": 1.2046555447701236e-06,
      "loss": 3.8535,
      "step": 1621
    },
    {
      "epoch": 0.2099880247273198,
      "grad_norm": 3.5482341679820357,
      "learning_rate": 1.2045498534202532e-06,
      "loss": 3.0977,
      "step": 1622
    },
    {
      "epoch": 0.21011748713467326,
      "grad_norm": 6.18356871089719,
      "learning_rate": 1.2044440436855786e-06,
      "loss": 5.0762,
      "step": 1623
    },
    {
      "epoch": 0.21024694954202675,
      "grad_norm": 4.861517687640406,
      "learning_rate": 1.204338115587713e-06,
      "loss": 3.8105,
      "step": 1624
    },
    {
      "epoch": 0.2103764119493802,
      "grad_norm": 4.7212659396117,
      "learning_rate": 1.2042320691482947e-06,
      "loss": 3.8896,
      "step": 1625
    },
    {
      "epoch": 0.2103764119493802,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5000888109207153,
      "eval_runtime": 14.2775,
      "eval_samples_per_second": 3.082,
      "eval_steps_per_second": 0.21,
      "step": 1625
    },
    {
      "epoch": 0.21050587435673365,
      "grad_norm": 2.1688058325535056,
      "learning_rate": 1.2041259043889859e-06,
      "loss": 2.8838,
      "step": 1626
    },
    {
      "epoch": 0.21063533676408713,
      "grad_norm": 4.851804058269344,
      "learning_rate": 1.2040196213314725e-06,
      "loss": 3.7476,
      "step": 1627
    },
    {
      "epoch": 0.2107647991714406,
      "grad_norm": 4.1047585664066375,
      "learning_rate": 1.2039132199974654e-06,
      "loss": 3.533,
      "step": 1628
    },
    {
      "epoch": 0.21089426157879407,
      "grad_norm": 3.5235614311939094,
      "learning_rate": 1.203806700408699e-06,
      "loss": 2.9795,
      "step": 1629
    },
    {
      "epoch": 0.21102372398614752,
      "grad_norm": 2.9896742926939095,
      "learning_rate": 1.2037000625869322e-06,
      "loss": 3.0508,
      "step": 1630
    },
    {
      "epoch": 0.21102372398614752,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5042613744735718,
      "eval_runtime": 14.4724,
      "eval_samples_per_second": 3.04,
      "eval_steps_per_second": 0.207,
      "step": 1630
    },
    {
      "epoch": 0.21115318639350097,
      "grad_norm": 2.2886077037850994,
      "learning_rate": 1.203593306553948e-06,
      "loss": 2.9263,
      "step": 1631
    },
    {
      "epoch": 0.21128264880085446,
      "grad_norm": 3.9010393418191804,
      "learning_rate": 1.2034864323315534e-06,
      "loss": 3.1021,
      "step": 1632
    },
    {
      "epoch": 0.2114121112082079,
      "grad_norm": 9.105088685323679,
      "learning_rate": 1.2033794399415797e-06,
      "loss": 5.1152,
      "step": 1633
    },
    {
      "epoch": 0.2115415736155614,
      "grad_norm": 1.9548685583033572,
      "learning_rate": 1.2032723294058818e-06,
      "loss": 2.7971,
      "step": 1634
    },
    {
      "epoch": 0.21167103602291484,
      "grad_norm": 4.058831672758348,
      "learning_rate": 1.20316510074634e-06,
      "loss": 3.0864,
      "step": 1635
    },
    {
      "epoch": 0.21167103602291484,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5087890625,
      "eval_runtime": 14.3289,
      "eval_samples_per_second": 3.071,
      "eval_steps_per_second": 0.209,
      "step": 1635
    },
    {
      "epoch": 0.21180049843026832,
      "grad_norm": 4.21247258213656,
      "learning_rate": 1.2030577539848572e-06,
      "loss": 3.291,
      "step": 1636
    },
    {
      "epoch": 0.21192996083762178,
      "grad_norm": 2.5602033146180037,
      "learning_rate": 1.2029502891433613e-06,
      "loss": 2.9839,
      "step": 1637
    },
    {
      "epoch": 0.21205942324497523,
      "grad_norm": 2.835412945575613,
      "learning_rate": 1.2028427062438047e-06,
      "loss": 3.4473,
      "step": 1638
    },
    {
      "epoch": 0.2121888856523287,
      "grad_norm": 3.3485353118282166,
      "learning_rate": 1.202735005308163e-06,
      "loss": 3.4084,
      "step": 1639
    },
    {
      "epoch": 0.21231834805968217,
      "grad_norm": 2.1071565396713496,
      "learning_rate": 1.202627186358436e-06,
      "loss": 3.0376,
      "step": 1640
    },
    {
      "epoch": 0.21231834805968217,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5138493776321411,
      "eval_runtime": 14.8914,
      "eval_samples_per_second": 2.955,
      "eval_steps_per_second": 0.201,
      "step": 1640
    },
    {
      "epoch": 0.21244781046703565,
      "grad_norm": 2.4917125372301925,
      "learning_rate": 1.2025192494166485e-06,
      "loss": 2.8843,
      "step": 1641
    },
    {
      "epoch": 0.2125772728743891,
      "grad_norm": 5.615226008186244,
      "learning_rate": 1.2024111945048484e-06,
      "loss": 3.312,
      "step": 1642
    },
    {
      "epoch": 0.21270673528174255,
      "grad_norm": 1.8900795222263203,
      "learning_rate": 1.2023030216451083e-06,
      "loss": 2.6948,
      "step": 1643
    },
    {
      "epoch": 0.21283619768909603,
      "grad_norm": 2.2432943589139924,
      "learning_rate": 1.2021947308595246e-06,
      "loss": 2.7708,
      "step": 1644
    },
    {
      "epoch": 0.2129656600964495,
      "grad_norm": 4.01607155678928,
      "learning_rate": 1.2020863221702179e-06,
      "loss": 3.1877,
      "step": 1645
    },
    {
      "epoch": 0.2129656600964495,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5144264698028564,
      "eval_runtime": 14.6713,
      "eval_samples_per_second": 2.999,
      "eval_steps_per_second": 0.204,
      "step": 1645
    },
    {
      "epoch": 0.21309512250380297,
      "grad_norm": 4.094127544488313,
      "learning_rate": 1.201977795599333e-06,
      "loss": 3.4551,
      "step": 1646
    },
    {
      "epoch": 0.21322458491115642,
      "grad_norm": 3.8544861193389544,
      "learning_rate": 1.2018691511690384e-06,
      "loss": 3.2529,
      "step": 1647
    },
    {
      "epoch": 0.21335404731850988,
      "grad_norm": 3.2845593181976716,
      "learning_rate": 1.201760388901527e-06,
      "loss": 3.1301,
      "step": 1648
    },
    {
      "epoch": 0.21348350972586336,
      "grad_norm": 4.033912929907339,
      "learning_rate": 1.2016515088190158e-06,
      "loss": 3.3105,
      "step": 1649
    },
    {
      "epoch": 0.2136129721332168,
      "grad_norm": 5.529040187528592,
      "learning_rate": 1.201542510943746e-06,
      "loss": 3.6665,
      "step": 1650
    },
    {
      "epoch": 0.2136129721332168,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5193092823028564,
      "eval_runtime": 14.5496,
      "eval_samples_per_second": 3.024,
      "eval_steps_per_second": 0.206,
      "step": 1650
    },
    {
      "epoch": 0.2137424345405703,
      "grad_norm": 2.340119716194272,
      "learning_rate": 1.2014333952979821e-06,
      "loss": 3.1594,
      "step": 1651
    },
    {
      "epoch": 0.21387189694792375,
      "grad_norm": 4.962570979093016,
      "learning_rate": 1.2013241619040134e-06,
      "loss": 3.3188,
      "step": 1652
    },
    {
      "epoch": 0.2140013593552772,
      "grad_norm": 1.4894036679197773,
      "learning_rate": 1.2012148107841532e-06,
      "loss": 2.748,
      "step": 1653
    },
    {
      "epoch": 0.21413082176263068,
      "grad_norm": 4.650532109161744,
      "learning_rate": 1.2011053419607385e-06,
      "loss": 3.6257,
      "step": 1654
    },
    {
      "epoch": 0.21426028416998413,
      "grad_norm": 1.4126791675097967,
      "learning_rate": 1.2009957554561308e-06,
      "loss": 2.2444,
      "step": 1655
    },
    {
      "epoch": 0.21426028416998413,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5247248411178589,
      "eval_runtime": 14.211,
      "eval_samples_per_second": 3.096,
      "eval_steps_per_second": 0.211,
      "step": 1655
    },
    {
      "epoch": 0.21438974657733761,
      "grad_norm": 2.5460850375869306,
      "learning_rate": 1.2008860512927148e-06,
      "loss": 3.3462,
      "step": 1656
    },
    {
      "epoch": 0.21451920898469107,
      "grad_norm": 4.613942727621425,
      "learning_rate": 1.2007762294929004e-06,
      "loss": 3.728,
      "step": 1657
    },
    {
      "epoch": 0.21464867139204455,
      "grad_norm": 4.833038520914195,
      "learning_rate": 1.2006662900791204e-06,
      "loss": 3.0662,
      "step": 1658
    },
    {
      "epoch": 0.214778133799398,
      "grad_norm": 4.31468274739582,
      "learning_rate": 1.2005562330738326e-06,
      "loss": 3.7954,
      "step": 1659
    },
    {
      "epoch": 0.21490759620675146,
      "grad_norm": 3.441702980385447,
      "learning_rate": 1.2004460584995184e-06,
      "loss": 2.8105,
      "step": 1660
    },
    {
      "epoch": 0.21490759620675146,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5227272510528564,
      "eval_runtime": 14.1175,
      "eval_samples_per_second": 3.117,
      "eval_steps_per_second": 0.213,
      "step": 1660
    },
    {
      "epoch": 0.21503705861410494,
      "grad_norm": 4.146549975653297,
      "learning_rate": 1.2003357663786827e-06,
      "loss": 3.7661,
      "step": 1661
    },
    {
      "epoch": 0.2151665210214584,
      "grad_norm": 2.5925309897445974,
      "learning_rate": 1.2002253567338554e-06,
      "loss": 3.0469,
      "step": 1662
    },
    {
      "epoch": 0.21529598342881187,
      "grad_norm": 4.031768209484338,
      "learning_rate": 1.2001148295875894e-06,
      "loss": 3.835,
      "step": 1663
    },
    {
      "epoch": 0.21542544583616532,
      "grad_norm": 6.38961541653933,
      "learning_rate": 1.2000041849624627e-06,
      "loss": 4.2883,
      "step": 1664
    },
    {
      "epoch": 0.21555490824351878,
      "grad_norm": 3.6936392660998782,
      "learning_rate": 1.1998934228810763e-06,
      "loss": 3.3057,
      "step": 1665
    },
    {
      "epoch": 0.21555490824351878,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5295188426971436,
      "eval_runtime": 14.3642,
      "eval_samples_per_second": 3.063,
      "eval_steps_per_second": 0.209,
      "step": 1665
    },
    {
      "epoch": 0.21568437065087226,
      "grad_norm": 5.84206550953815,
      "learning_rate": 1.1997825433660557e-06,
      "loss": 3.7471,
      "step": 1666
    },
    {
      "epoch": 0.2158138330582257,
      "grad_norm": 3.6429972516641853,
      "learning_rate": 1.1996715464400504e-06,
      "loss": 3.5845,
      "step": 1667
    },
    {
      "epoch": 0.2159432954655792,
      "grad_norm": 3.3123082689141894,
      "learning_rate": 1.1995604321257335e-06,
      "loss": 3.188,
      "step": 1668
    },
    {
      "epoch": 0.21607275787293265,
      "grad_norm": 5.229457376640021,
      "learning_rate": 1.1994492004458028e-06,
      "loss": 3.0566,
      "step": 1669
    },
    {
      "epoch": 0.2162022202802861,
      "grad_norm": 4.974992338949791,
      "learning_rate": 1.1993378514229791e-06,
      "loss": 3.3535,
      "step": 1670
    },
    {
      "epoch": 0.2162022202802861,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.530229091644287,
      "eval_runtime": 14.6104,
      "eval_samples_per_second": 3.012,
      "eval_steps_per_second": 0.205,
      "step": 1670
    },
    {
      "epoch": 0.21633168268763958,
      "grad_norm": 4.905074376396741,
      "learning_rate": 1.199226385080008e-06,
      "loss": 3.427,
      "step": 1671
    },
    {
      "epoch": 0.21646114509499303,
      "grad_norm": 2.169399331763751,
      "learning_rate": 1.1991148014396587e-06,
      "loss": 3.1394,
      "step": 1672
    },
    {
      "epoch": 0.21659060750234652,
      "grad_norm": 3.729323603756549,
      "learning_rate": 1.1990031005247247e-06,
      "loss": 3.2573,
      "step": 1673
    },
    {
      "epoch": 0.21672006990969997,
      "grad_norm": 1.6181688473384985,
      "learning_rate": 1.1988912823580226e-06,
      "loss": 2.7397,
      "step": 1674
    },
    {
      "epoch": 0.21684953231705342,
      "grad_norm": 2.9622056511634325,
      "learning_rate": 1.1987793469623939e-06,
      "loss": 3.342,
      "step": 1675
    },
    {
      "epoch": 0.21684953231705342,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5323597192764282,
      "eval_runtime": 14.6107,
      "eval_samples_per_second": 3.011,
      "eval_steps_per_second": 0.205,
      "step": 1675
    },
    {
      "epoch": 0.2169789947244069,
      "grad_norm": 5.0521818109629,
      "learning_rate": 1.1986672943607037e-06,
      "loss": 3.6797,
      "step": 1676
    },
    {
      "epoch": 0.21710845713176036,
      "grad_norm": 3.2029765164963844,
      "learning_rate": 1.1985551245758408e-06,
      "loss": 3.3445,
      "step": 1677
    },
    {
      "epoch": 0.21723791953911384,
      "grad_norm": 3.4842845942844165,
      "learning_rate": 1.1984428376307184e-06,
      "loss": 3.7869,
      "step": 1678
    },
    {
      "epoch": 0.2173673819464673,
      "grad_norm": 3.708751412318944,
      "learning_rate": 1.198330433548273e-06,
      "loss": 3.0625,
      "step": 1679
    },
    {
      "epoch": 0.21749684435382077,
      "grad_norm": 1.9442744771078875,
      "learning_rate": 1.198217912351466e-06,
      "loss": 2.668,
      "step": 1680
    },
    {
      "epoch": 0.21749684435382077,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5377751588821411,
      "eval_runtime": 13.9417,
      "eval_samples_per_second": 3.156,
      "eval_steps_per_second": 0.215,
      "step": 1680
    },
    {
      "epoch": 0.21762630676117423,
      "grad_norm": 2.067723877629095,
      "learning_rate": 1.1981052740632815e-06,
      "loss": 2.719,
      "step": 1681
    },
    {
      "epoch": 0.21775576916852768,
      "grad_norm": 4.462641611864524,
      "learning_rate": 1.1979925187067287e-06,
      "loss": 3.4854,
      "step": 1682
    },
    {
      "epoch": 0.21788523157588116,
      "grad_norm": 4.811453825784267,
      "learning_rate": 1.1978796463048398e-06,
      "loss": 3.5254,
      "step": 1683
    },
    {
      "epoch": 0.21801469398323461,
      "grad_norm": 2.860584095818227,
      "learning_rate": 1.1977666568806715e-06,
      "loss": 3.0029,
      "step": 1684
    },
    {
      "epoch": 0.2181441563905881,
      "grad_norm": 3.0771830392382378,
      "learning_rate": 1.1976535504573038e-06,
      "loss": 3.1792,
      "step": 1685
    },
    {
      "epoch": 0.2181441563905881,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5383522510528564,
      "eval_runtime": 14.5011,
      "eval_samples_per_second": 3.034,
      "eval_steps_per_second": 0.207,
      "step": 1685
    },
    {
      "epoch": 0.21827361879794155,
      "grad_norm": 5.2504609356974745,
      "learning_rate": 1.1975403270578416e-06,
      "loss": 4.166,
      "step": 1686
    },
    {
      "epoch": 0.218403081205295,
      "grad_norm": 3.6410347766971634,
      "learning_rate": 1.1974269867054124e-06,
      "loss": 3.1914,
      "step": 1687
    },
    {
      "epoch": 0.21853254361264848,
      "grad_norm": 3.277556647628424,
      "learning_rate": 1.197313529423169e-06,
      "loss": 3.2913,
      "step": 1688
    },
    {
      "epoch": 0.21866200602000194,
      "grad_norm": 3.187593779756875,
      "learning_rate": 1.1971999552342867e-06,
      "loss": 3.1484,
      "step": 1689
    },
    {
      "epoch": 0.21879146842735542,
      "grad_norm": 3.6581996248932533,
      "learning_rate": 1.1970862641619657e-06,
      "loss": 3.1729,
      "step": 1690
    },
    {
      "epoch": 0.21879146842735542,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5446555614471436,
      "eval_runtime": 14.3306,
      "eval_samples_per_second": 3.07,
      "eval_steps_per_second": 0.209,
      "step": 1690
    },
    {
      "epoch": 0.21892093083470887,
      "grad_norm": 3.4757812556040912,
      "learning_rate": 1.1969724562294294e-06,
      "loss": 3.3074,
      "step": 1691
    },
    {
      "epoch": 0.21905039324206232,
      "grad_norm": 2.303171804191525,
      "learning_rate": 1.196858531459926e-06,
      "loss": 2.8193,
      "step": 1692
    },
    {
      "epoch": 0.2191798556494158,
      "grad_norm": 7.223176287536503,
      "learning_rate": 1.196744489876726e-06,
      "loss": 4.1494,
      "step": 1693
    },
    {
      "epoch": 0.21930931805676926,
      "grad_norm": 5.135747445169343,
      "learning_rate": 1.1966303315031254e-06,
      "loss": 3.8594,
      "step": 1694
    },
    {
      "epoch": 0.21943878046412274,
      "grad_norm": 2.8515472222807596,
      "learning_rate": 1.1965160563624432e-06,
      "loss": 2.7759,
      "step": 1695
    },
    {
      "epoch": 0.21943878046412274,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.548828125,
      "eval_runtime": 14.1962,
      "eval_samples_per_second": 3.099,
      "eval_steps_per_second": 0.211,
      "step": 1695
    },
    {
      "epoch": 0.2195682428714762,
      "grad_norm": 2.4805486640742993,
      "learning_rate": 1.1964016644780222e-06,
      "loss": 2.8004,
      "step": 1696
    },
    {
      "epoch": 0.21969770527882965,
      "grad_norm": 6.763167382749191,
      "learning_rate": 1.1962871558732297e-06,
      "loss": 4.3635,
      "step": 1697
    },
    {
      "epoch": 0.21982716768618313,
      "grad_norm": 3.2400450004153827,
      "learning_rate": 1.196172530571456e-06,
      "loss": 3.4238,
      "step": 1698
    },
    {
      "epoch": 0.21995663009353658,
      "grad_norm": 3.9635746979996807,
      "learning_rate": 1.1960577885961158e-06,
      "loss": 3.2061,
      "step": 1699
    },
    {
      "epoch": 0.22008609250089006,
      "grad_norm": 2.9452614763167952,
      "learning_rate": 1.1959429299706477e-06,
      "loss": 2.8008,
      "step": 1700
    },
    {
      "epoch": 0.22008609250089006,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5465642213821411,
      "eval_runtime": 13.6307,
      "eval_samples_per_second": 3.228,
      "eval_steps_per_second": 0.22,
      "step": 1700
    },
    {
      "epoch": 0.22021555490824352,
      "grad_norm": 3.497797877571419,
      "learning_rate": 1.1958279547185133e-06,
      "loss": 3.2026,
      "step": 1701
    },
    {
      "epoch": 0.220345017315597,
      "grad_norm": 2.033177315949458,
      "learning_rate": 1.1957128628631991e-06,
      "loss": 2.9541,
      "step": 1702
    },
    {
      "epoch": 0.22047447972295045,
      "grad_norm": 4.255071739556924,
      "learning_rate": 1.195597654428215e-06,
      "loss": 3.1274,
      "step": 1703
    },
    {
      "epoch": 0.2206039421303039,
      "grad_norm": 4.309098485968686,
      "learning_rate": 1.1954823294370944e-06,
      "loss": 3.3474,
      "step": 1704
    },
    {
      "epoch": 0.22073340453765739,
      "grad_norm": 3.958231854238835,
      "learning_rate": 1.1953668879133949e-06,
      "loss": 3.2529,
      "step": 1705
    },
    {
      "epoch": 0.22073340453765739,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.550248622894287,
      "eval_runtime": 14.5991,
      "eval_samples_per_second": 3.014,
      "eval_steps_per_second": 0.205,
      "step": 1705
    },
    {
      "epoch": 0.22086286694501084,
      "grad_norm": 8.126279684991312,
      "learning_rate": 1.1952513298806976e-06,
      "loss": 4.4897,
      "step": 1706
    },
    {
      "epoch": 0.22099232935236432,
      "grad_norm": 5.890598154123479,
      "learning_rate": 1.1951356553626079e-06,
      "loss": 4.1187,
      "step": 1707
    },
    {
      "epoch": 0.22112179175971777,
      "grad_norm": 5.0241134576385384,
      "learning_rate": 1.1950198643827543e-06,
      "loss": 3.6501,
      "step": 1708
    },
    {
      "epoch": 0.22125125416707123,
      "grad_norm": 2.268935548340834,
      "learning_rate": 1.1949039569647897e-06,
      "loss": 2.9326,
      "step": 1709
    },
    {
      "epoch": 0.2213807165744247,
      "grad_norm": 5.984558256083068,
      "learning_rate": 1.1947879331323905e-06,
      "loss": 3.3857,
      "step": 1710
    },
    {
      "epoch": 0.2213807165744247,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5528675317764282,
      "eval_runtime": 14.5547,
      "eval_samples_per_second": 3.023,
      "eval_steps_per_second": 0.206,
      "step": 1710
    },
    {
      "epoch": 0.22151017898177816,
      "grad_norm": 1.967178592416272,
      "learning_rate": 1.194671792909257e-06,
      "loss": 3.0688,
      "step": 1711
    },
    {
      "epoch": 0.22163964138913164,
      "grad_norm": 5.234466646323892,
      "learning_rate": 1.1945555363191127e-06,
      "loss": 3.4585,
      "step": 1712
    },
    {
      "epoch": 0.2217691037964851,
      "grad_norm": 4.72821504290605,
      "learning_rate": 1.194439163385706e-06,
      "loss": 4.0537,
      "step": 1713
    },
    {
      "epoch": 0.22189856620383855,
      "grad_norm": 9.418206175108816,
      "learning_rate": 1.1943226741328082e-06,
      "loss": 5.1987,
      "step": 1714
    },
    {
      "epoch": 0.22202802861119203,
      "grad_norm": 3.0488772639622534,
      "learning_rate": 1.1942060685842144e-06,
      "loss": 3.2327,
      "step": 1715
    },
    {
      "epoch": 0.22202802861119203,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.553755283355713,
      "eval_runtime": 18.9358,
      "eval_samples_per_second": 2.324,
      "eval_steps_per_second": 0.158,
      "step": 1715
    },
    {
      "epoch": 0.22215749101854548,
      "grad_norm": 3.1031933334475554,
      "learning_rate": 1.194089346763744e-06,
      "loss": 3.1274,
      "step": 1716
    },
    {
      "epoch": 0.22228695342589896,
      "grad_norm": 2.3409504141769717,
      "learning_rate": 1.1939725086952395e-06,
      "loss": 2.5083,
      "step": 1717
    },
    {
      "epoch": 0.22241641583325242,
      "grad_norm": 1.9560616620977689,
      "learning_rate": 1.1938555544025675e-06,
      "loss": 2.7908,
      "step": 1718
    },
    {
      "epoch": 0.22254587824060587,
      "grad_norm": 3.9541086621729105,
      "learning_rate": 1.1937384839096184e-06,
      "loss": 3.2656,
      "step": 1719
    },
    {
      "epoch": 0.22267534064795935,
      "grad_norm": 2.2271881270141907,
      "learning_rate": 1.1936212972403062e-06,
      "loss": 2.8665,
      "step": 1720
    },
    {
      "epoch": 0.22267534064795935,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5585049390792847,
      "eval_runtime": 13.9785,
      "eval_samples_per_second": 3.148,
      "eval_steps_per_second": 0.215,
      "step": 1720
    },
    {
      "epoch": 0.2228048030553128,
      "grad_norm": 5.169693777968014,
      "learning_rate": 1.1935039944185685e-06,
      "loss": 3.8398,
      "step": 1721
    },
    {
      "epoch": 0.2229342654626663,
      "grad_norm": 1.979788095273356,
      "learning_rate": 1.193386575468367e-06,
      "loss": 3.1157,
      "step": 1722
    },
    {
      "epoch": 0.22306372787001974,
      "grad_norm": 4.605638918372798,
      "learning_rate": 1.1932690404136869e-06,
      "loss": 3.1528,
      "step": 1723
    },
    {
      "epoch": 0.22319319027737322,
      "grad_norm": 2.421821502820245,
      "learning_rate": 1.1931513892785366e-06,
      "loss": 2.8501,
      "step": 1724
    },
    {
      "epoch": 0.22332265268472667,
      "grad_norm": 3.8931793816908336,
      "learning_rate": 1.1930336220869495e-06,
      "loss": 3.2036,
      "step": 1725
    },
    {
      "epoch": 0.22332265268472667,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5626775026321411,
      "eval_runtime": 14.462,
      "eval_samples_per_second": 3.042,
      "eval_steps_per_second": 0.207,
      "step": 1725
    },
    {
      "epoch": 0.22345211509208013,
      "grad_norm": 5.490700890607397,
      "learning_rate": 1.1929157388629815e-06,
      "loss": 3.8892,
      "step": 1726
    },
    {
      "epoch": 0.2235815774994336,
      "grad_norm": 4.815514048050006,
      "learning_rate": 1.1927977396307126e-06,
      "loss": 3.6133,
      "step": 1727
    },
    {
      "epoch": 0.22371103990678706,
      "grad_norm": 2.014094366630197,
      "learning_rate": 1.1926796244142466e-06,
      "loss": 2.9907,
      "step": 1728
    },
    {
      "epoch": 0.22384050231414054,
      "grad_norm": 1.47706601569829,
      "learning_rate": 1.1925613932377112e-06,
      "loss": 2.5127,
      "step": 1729
    },
    {
      "epoch": 0.223969964721494,
      "grad_norm": 3.061316440173626,
      "learning_rate": 1.1924430461252572e-06,
      "loss": 3.126,
      "step": 1730
    },
    {
      "epoch": 0.223969964721494,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5637428760528564,
      "eval_runtime": 15.4829,
      "eval_samples_per_second": 2.842,
      "eval_steps_per_second": 0.194,
      "step": 1730
    },
    {
      "epoch": 0.22409942712884745,
      "grad_norm": 4.319830308518956,
      "learning_rate": 1.1923245831010594e-06,
      "loss": 3.5161,
      "step": 1731
    },
    {
      "epoch": 0.22422888953620093,
      "grad_norm": 1.9709655852753052,
      "learning_rate": 1.1922060041893163e-06,
      "loss": 2.7952,
      "step": 1732
    },
    {
      "epoch": 0.22435835194355438,
      "grad_norm": 2.240317459449038,
      "learning_rate": 1.1920873094142504e-06,
      "loss": 2.939,
      "step": 1733
    },
    {
      "epoch": 0.22448781435090787,
      "grad_norm": 7.1995988693445545,
      "learning_rate": 1.191968498800107e-06,
      "loss": 3.9414,
      "step": 1734
    },
    {
      "epoch": 0.22461727675826132,
      "grad_norm": 2.35580583883058,
      "learning_rate": 1.1918495723711557e-06,
      "loss": 2.8203,
      "step": 1735
    },
    {
      "epoch": 0.22461727675826132,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5696910619735718,
      "eval_runtime": 14.6631,
      "eval_samples_per_second": 3.001,
      "eval_steps_per_second": 0.205,
      "step": 1735
    },
    {
      "epoch": 0.22474673916561477,
      "grad_norm": 1.6119617724320778,
      "learning_rate": 1.1917305301516898e-06,
      "loss": 2.6233,
      "step": 1736
    },
    {
      "epoch": 0.22487620157296825,
      "grad_norm": 2.5082210833575407,
      "learning_rate": 1.191611372166026e-06,
      "loss": 2.959,
      "step": 1737
    },
    {
      "epoch": 0.2250056639803217,
      "grad_norm": 3.6889685883679566,
      "learning_rate": 1.191492098438505e-06,
      "loss": 3.4849,
      "step": 1738
    },
    {
      "epoch": 0.2251351263876752,
      "grad_norm": 4.609415356121887,
      "learning_rate": 1.1913727089934905e-06,
      "loss": 3.9851,
      "step": 1739
    },
    {
      "epoch": 0.22526458879502864,
      "grad_norm": 5.308911292675775,
      "learning_rate": 1.19125320385537e-06,
      "loss": 3.5415,
      "step": 1740
    },
    {
      "epoch": 0.22526458879502864,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5680485963821411,
      "eval_runtime": 14.5759,
      "eval_samples_per_second": 3.019,
      "eval_steps_per_second": 0.206,
      "step": 1740
    },
    {
      "epoch": 0.2253940512023821,
      "grad_norm": 5.842219639744819,
      "learning_rate": 1.1911335830485556e-06,
      "loss": 3.8286,
      "step": 1741
    },
    {
      "epoch": 0.22552351360973558,
      "grad_norm": 2.5664585340326784,
      "learning_rate": 1.1910138465974815e-06,
      "loss": 3.1953,
      "step": 1742
    },
    {
      "epoch": 0.22565297601708903,
      "grad_norm": 2.172513126349862,
      "learning_rate": 1.190893994526607e-06,
      "loss": 2.7937,
      "step": 1743
    },
    {
      "epoch": 0.2257824384244425,
      "grad_norm": 4.066184345071473,
      "learning_rate": 1.1907740268604139e-06,
      "loss": 3.646,
      "step": 1744
    },
    {
      "epoch": 0.22591190083179596,
      "grad_norm": 5.395992203250564,
      "learning_rate": 1.190653943623408e-06,
      "loss": 3.8916,
      "step": 1745
    },
    {
      "epoch": 0.22591190083179596,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5708895921707153,
      "eval_runtime": 14.7459,
      "eval_samples_per_second": 2.984,
      "eval_steps_per_second": 0.203,
      "step": 1745
    },
    {
      "epoch": 0.22604136323914945,
      "grad_norm": 3.8514381132523092,
      "learning_rate": 1.190533744840119e-06,
      "loss": 3.1475,
      "step": 1746
    },
    {
      "epoch": 0.2261708256465029,
      "grad_norm": 5.996010176458452,
      "learning_rate": 1.1904134305350996e-06,
      "loss": 3.0664,
      "step": 1747
    },
    {
      "epoch": 0.22630028805385635,
      "grad_norm": 3.8790684631487284,
      "learning_rate": 1.1902930007329265e-06,
      "loss": 2.8591,
      "step": 1748
    },
    {
      "epoch": 0.22642975046120983,
      "grad_norm": 2.321835374732646,
      "learning_rate": 1.1901724554582001e-06,
      "loss": 2.7219,
      "step": 1749
    },
    {
      "epoch": 0.2265592128685633,
      "grad_norm": 4.849322392776263,
      "learning_rate": 1.1900517947355442e-06,
      "loss": 3.2275,
      "step": 1750
    },
    {
      "epoch": 0.2265592128685633,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5711559057235718,
      "eval_runtime": 14.454,
      "eval_samples_per_second": 3.044,
      "eval_steps_per_second": 0.208,
      "step": 1750
    },
    {
      "epoch": 0.22668867527591677,
      "grad_norm": 4.739930524655867,
      "learning_rate": 1.1899310185896062e-06,
      "loss": 3.6909,
      "step": 1751
    },
    {
      "epoch": 0.22681813768327022,
      "grad_norm": 2.5714961159532326,
      "learning_rate": 1.1898101270450569e-06,
      "loss": 2.7419,
      "step": 1752
    },
    {
      "epoch": 0.22694760009062367,
      "grad_norm": 1.571708821529022,
      "learning_rate": 1.1896891201265911e-06,
      "loss": 2.6799,
      "step": 1753
    },
    {
      "epoch": 0.22707706249797716,
      "grad_norm": 6.6975266751647835,
      "learning_rate": 1.1895679978589266e-06,
      "loss": 3.9404,
      "step": 1754
    },
    {
      "epoch": 0.2272065249053306,
      "grad_norm": 1.8552666812490077,
      "learning_rate": 1.1894467602668056e-06,
      "loss": 2.9622,
      "step": 1755
    },
    {
      "epoch": 0.2272065249053306,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5763493776321411,
      "eval_runtime": 15.8993,
      "eval_samples_per_second": 2.767,
      "eval_steps_per_second": 0.189,
      "step": 1755
    },
    {
      "epoch": 0.2273359873126841,
      "grad_norm": 6.798012046951795,
      "learning_rate": 1.1893254073749926e-06,
      "loss": 4.6042,
      "step": 1756
    },
    {
      "epoch": 0.22746544972003754,
      "grad_norm": 6.031227505324747,
      "learning_rate": 1.1892039392082771e-06,
      "loss": 3.5444,
      "step": 1757
    },
    {
      "epoch": 0.227594912127391,
      "grad_norm": 1.9836317094131743,
      "learning_rate": 1.189082355791471e-06,
      "loss": 2.3232,
      "step": 1758
    },
    {
      "epoch": 0.22772437453474448,
      "grad_norm": 4.378499974481291,
      "learning_rate": 1.1889606571494103e-06,
      "loss": 3.6782,
      "step": 1759
    },
    {
      "epoch": 0.22785383694209793,
      "grad_norm": 5.157970620135607,
      "learning_rate": 1.1888388433069545e-06,
      "loss": 4.2222,
      "step": 1760
    },
    {
      "epoch": 0.22785383694209793,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5777255296707153,
      "eval_runtime": 14.1102,
      "eval_samples_per_second": 3.118,
      "eval_steps_per_second": 0.213,
      "step": 1760
    },
    {
      "epoch": 0.2279832993494514,
      "grad_norm": 3.2377565781452096,
      "learning_rate": 1.1887169142889866e-06,
      "loss": 3.1804,
      "step": 1761
    },
    {
      "epoch": 0.22811276175680487,
      "grad_norm": 4.248476010929114,
      "learning_rate": 1.1885948701204125e-06,
      "loss": 3.6548,
      "step": 1762
    },
    {
      "epoch": 0.22824222416415832,
      "grad_norm": 3.9093861535477257,
      "learning_rate": 1.188472710826163e-06,
      "loss": 3.5649,
      "step": 1763
    },
    {
      "epoch": 0.2283716865715118,
      "grad_norm": 3.0200621707086452,
      "learning_rate": 1.188350436431191e-06,
      "loss": 3.3237,
      "step": 1764
    },
    {
      "epoch": 0.22850114897886525,
      "grad_norm": 2.747628671198382,
      "learning_rate": 1.188228046960474e-06,
      "loss": 2.8564,
      "step": 1765
    },
    {
      "epoch": 0.22850114897886525,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5797230005264282,
      "eval_runtime": 14.7421,
      "eval_samples_per_second": 2.985,
      "eval_steps_per_second": 0.203,
      "step": 1765
    },
    {
      "epoch": 0.22863061138621874,
      "grad_norm": 1.822798965074975,
      "learning_rate": 1.188105542439012e-06,
      "loss": 2.7983,
      "step": 1766
    },
    {
      "epoch": 0.2287600737935722,
      "grad_norm": 3.1483265131972358,
      "learning_rate": 1.1879829228918295e-06,
      "loss": 3.0876,
      "step": 1767
    },
    {
      "epoch": 0.22888953620092567,
      "grad_norm": 2.2193564040078617,
      "learning_rate": 1.1878601883439738e-06,
      "loss": 2.9304,
      "step": 1768
    },
    {
      "epoch": 0.22901899860827912,
      "grad_norm": 4.644730132949049,
      "learning_rate": 1.1877373388205162e-06,
      "loss": 4.0947,
      "step": 1769
    },
    {
      "epoch": 0.22914846101563258,
      "grad_norm": 2.395911144377658,
      "learning_rate": 1.1876143743465507e-06,
      "loss": 2.9478,
      "step": 1770
    },
    {
      "epoch": 0.22914846101563258,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5846058130264282,
      "eval_runtime": 14.7895,
      "eval_samples_per_second": 2.975,
      "eval_steps_per_second": 0.203,
      "step": 1770
    },
    {
      "epoch": 0.22927792342298606,
      "grad_norm": 1.510404962451898,
      "learning_rate": 1.1874912949471957e-06,
      "loss": 2.5127,
      "step": 1771
    },
    {
      "epoch": 0.2294073858303395,
      "grad_norm": 4.1358395535030725,
      "learning_rate": 1.1873681006475924e-06,
      "loss": 3.5691,
      "step": 1772
    },
    {
      "epoch": 0.229536848237693,
      "grad_norm": 3.268384502322116,
      "learning_rate": 1.187244791472906e-06,
      "loss": 3.0142,
      "step": 1773
    },
    {
      "epoch": 0.22966631064504645,
      "grad_norm": 6.264869445363574,
      "learning_rate": 1.1871213674483247e-06,
      "loss": 3.8196,
      "step": 1774
    },
    {
      "epoch": 0.2297957730523999,
      "grad_norm": 3.7064875790383,
      "learning_rate": 1.1869978285990604e-06,
      "loss": 3.4233,
      "step": 1775
    },
    {
      "epoch": 0.2297957730523999,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5838290452957153,
      "eval_runtime": 13.1211,
      "eval_samples_per_second": 3.353,
      "eval_steps_per_second": 0.229,
      "step": 1775
    },
    {
      "epoch": 0.22992523545975338,
      "grad_norm": 2.0280575355571004,
      "learning_rate": 1.1868741749503485e-06,
      "loss": 2.8713,
      "step": 1776
    },
    {
      "epoch": 0.23005469786710683,
      "grad_norm": 2.2507887943390634,
      "learning_rate": 1.1867504065274477e-06,
      "loss": 2.895,
      "step": 1777
    },
    {
      "epoch": 0.23018416027446031,
      "grad_norm": 1.9347822544009718,
      "learning_rate": 1.18662652335564e-06,
      "loss": 2.5469,
      "step": 1778
    },
    {
      "epoch": 0.23031362268181377,
      "grad_norm": 6.252203058900052,
      "learning_rate": 1.1865025254602317e-06,
      "loss": 3.9983,
      "step": 1779
    },
    {
      "epoch": 0.23044308508916722,
      "grad_norm": 2.7984166102703436,
      "learning_rate": 1.1863784128665512e-06,
      "loss": 3.2495,
      "step": 1780
    },
    {
      "epoch": 0.23044308508916722,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5873135328292847,
      "eval_runtime": 13.7,
      "eval_samples_per_second": 3.212,
      "eval_steps_per_second": 0.219,
      "step": 1780
    },
    {
      "epoch": 0.2305725474965207,
      "grad_norm": 4.172889916817173,
      "learning_rate": 1.186254185599951e-06,
      "loss": 3.5293,
      "step": 1781
    },
    {
      "epoch": 0.23070200990387416,
      "grad_norm": 3.0835522763063885,
      "learning_rate": 1.1861298436858075e-06,
      "loss": 3.3398,
      "step": 1782
    },
    {
      "epoch": 0.23083147231122764,
      "grad_norm": 4.9575589410589584,
      "learning_rate": 1.1860053871495199e-06,
      "loss": 3.1064,
      "step": 1783
    },
    {
      "epoch": 0.2309609347185811,
      "grad_norm": 3.6457331121335304,
      "learning_rate": 1.1858808160165108e-06,
      "loss": 2.9993,
      "step": 1784
    },
    {
      "epoch": 0.23109039712593454,
      "grad_norm": 6.046506128296669,
      "learning_rate": 1.1857561303122266e-06,
      "loss": 3.314,
      "step": 1785
    },
    {
      "epoch": 0.23109039712593454,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5865589380264282,
      "eval_runtime": 14.3035,
      "eval_samples_per_second": 3.076,
      "eval_steps_per_second": 0.21,
      "step": 1785
    },
    {
      "epoch": 0.23121985953328802,
      "grad_norm": 2.636689754406309,
      "learning_rate": 1.1856313300621367e-06,
      "loss": 2.8096,
      "step": 1786
    },
    {
      "epoch": 0.23134932194064148,
      "grad_norm": 5.578799757744635,
      "learning_rate": 1.1855064152917342e-06,
      "loss": 3.3169,
      "step": 1787
    },
    {
      "epoch": 0.23147878434799496,
      "grad_norm": 1.889160245127263,
      "learning_rate": 1.1853813860265354e-06,
      "loss": 2.4883,
      "step": 1788
    },
    {
      "epoch": 0.2316082467553484,
      "grad_norm": 3.2704907217190495,
      "learning_rate": 1.18525624229208e-06,
      "loss": 3.0698,
      "step": 1789
    },
    {
      "epoch": 0.2317377091627019,
      "grad_norm": 4.624546217255355,
      "learning_rate": 1.1851309841139313e-06,
      "loss": 3.5776,
      "step": 1790
    },
    {
      "epoch": 0.2317377091627019,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.590043544769287,
      "eval_runtime": 13.8172,
      "eval_samples_per_second": 3.184,
      "eval_steps_per_second": 0.217,
      "step": 1790
    },
    {
      "epoch": 0.23186717157005535,
      "grad_norm": 1.5267695445893605,
      "learning_rate": 1.1850056115176758e-06,
      "loss": 2.6177,
      "step": 1791
    },
    {
      "epoch": 0.2319966339774088,
      "grad_norm": 1.5428228815579181,
      "learning_rate": 1.1848801245289236e-06,
      "loss": 2.6587,
      "step": 1792
    },
    {
      "epoch": 0.23212609638476228,
      "grad_norm": 2.320113895526707,
      "learning_rate": 1.1847545231733071e-06,
      "loss": 2.8708,
      "step": 1793
    },
    {
      "epoch": 0.23225555879211573,
      "grad_norm": 2.205125100465697,
      "learning_rate": 1.184628807476484e-06,
      "loss": 3.1047,
      "step": 1794
    },
    {
      "epoch": 0.23238502119946922,
      "grad_norm": 4.439712028758625,
      "learning_rate": 1.1845029774641338e-06,
      "loss": 3.6821,
      "step": 1795
    },
    {
      "epoch": 0.23238502119946922,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.592817783355713,
      "eval_runtime": 15.1244,
      "eval_samples_per_second": 2.909,
      "eval_steps_per_second": 0.198,
      "step": 1795
    },
    {
      "epoch": 0.23251448360682267,
      "grad_norm": 4.962734232554875,
      "learning_rate": 1.1843770331619597e-06,
      "loss": 3.2542,
      "step": 1796
    },
    {
      "epoch": 0.23264394601417612,
      "grad_norm": 1.5757063984383173,
      "learning_rate": 1.1842509745956885e-06,
      "loss": 2.9131,
      "step": 1797
    },
    {
      "epoch": 0.2327734084215296,
      "grad_norm": 1.5413241371035584,
      "learning_rate": 1.1841248017910702e-06,
      "loss": 2.6279,
      "step": 1798
    },
    {
      "epoch": 0.23290287082888306,
      "grad_norm": 2.4034265576485203,
      "learning_rate": 1.1839985147738784e-06,
      "loss": 3.144,
      "step": 1799
    },
    {
      "epoch": 0.23303233323623654,
      "grad_norm": 2.212079301669578,
      "learning_rate": 1.1838721135699094e-06,
      "loss": 2.6068,
      "step": 1800
    },
    {
      "epoch": 0.23303233323623654,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5936390161514282,
      "eval_runtime": 13.3757,
      "eval_samples_per_second": 3.29,
      "eval_steps_per_second": 0.224,
      "step": 1800
    },
    {
      "epoch": 0.23316179564359,
      "grad_norm": 5.478674115184305,
      "learning_rate": 1.1837455982049835e-06,
      "loss": 3.1055,
      "step": 1801
    },
    {
      "epoch": 0.23329125805094345,
      "grad_norm": 4.347527707245364,
      "learning_rate": 1.183618968704944e-06,
      "loss": 3.356,
      "step": 1802
    },
    {
      "epoch": 0.23342072045829693,
      "grad_norm": 2.2850104345524347,
      "learning_rate": 1.1834922250956572e-06,
      "loss": 3.0022,
      "step": 1803
    },
    {
      "epoch": 0.23355018286565038,
      "grad_norm": 4.075156425449902,
      "learning_rate": 1.1833653674030134e-06,
      "loss": 3.6924,
      "step": 1804
    },
    {
      "epoch": 0.23367964527300386,
      "grad_norm": 1.7377919385412286,
      "learning_rate": 1.1832383956529258e-06,
      "loss": 2.7505,
      "step": 1805
    },
    {
      "epoch": 0.23367964527300386,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.589799404144287,
      "eval_runtime": 15.0026,
      "eval_samples_per_second": 2.933,
      "eval_steps_per_second": 0.2,
      "step": 1805
    },
    {
      "epoch": 0.23380910768035731,
      "grad_norm": 2.6073283630427344,
      "learning_rate": 1.1831113098713307e-06,
      "loss": 2.9761,
      "step": 1806
    },
    {
      "epoch": 0.23393857008771077,
      "grad_norm": 3.614322979767949,
      "learning_rate": 1.1829841100841884e-06,
      "loss": 3.074,
      "step": 1807
    },
    {
      "epoch": 0.23406803249506425,
      "grad_norm": 4.5033987054023115,
      "learning_rate": 1.1828567963174816e-06,
      "loss": 3.261,
      "step": 1808
    },
    {
      "epoch": 0.2341974949024177,
      "grad_norm": 2.2737850508785233,
      "learning_rate": 1.1827293685972172e-06,
      "loss": 2.6584,
      "step": 1809
    },
    {
      "epoch": 0.23432695730977118,
      "grad_norm": 2.8210947074723105,
      "learning_rate": 1.1826018269494244e-06,
      "loss": 2.7859,
      "step": 1810
    },
    {
      "epoch": 0.23432695730977118,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.595458984375,
      "eval_runtime": 14.9967,
      "eval_samples_per_second": 2.934,
      "eval_steps_per_second": 0.2,
      "step": 1810
    },
    {
      "epoch": 0.23445641971712464,
      "grad_norm": 3.602167179229879,
      "learning_rate": 1.1824741714001565e-06,
      "loss": 3.3247,
      "step": 1811
    },
    {
      "epoch": 0.23458588212447812,
      "grad_norm": 1.9449066011243887,
      "learning_rate": 1.1823464019754895e-06,
      "loss": 2.8586,
      "step": 1812
    },
    {
      "epoch": 0.23471534453183157,
      "grad_norm": 4.176242688853541,
      "learning_rate": 1.182218518701523e-06,
      "loss": 3.2878,
      "step": 1813
    },
    {
      "epoch": 0.23484480693918502,
      "grad_norm": 5.628199294341703,
      "learning_rate": 1.1820905216043797e-06,
      "loss": 3.6523,
      "step": 1814
    },
    {
      "epoch": 0.2349742693465385,
      "grad_norm": 3.8074938777871594,
      "learning_rate": 1.1819624107102057e-06,
      "loss": 3.4065,
      "step": 1815
    },
    {
      "epoch": 0.2349742693465385,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.594682216644287,
      "eval_runtime": 14.363,
      "eval_samples_per_second": 3.063,
      "eval_steps_per_second": 0.209,
      "step": 1815
    },
    {
      "epoch": 0.23510373175389196,
      "grad_norm": 1.5983667295130326,
      "learning_rate": 1.1818341860451703e-06,
      "loss": 2.5,
      "step": 1816
    },
    {
      "epoch": 0.23523319416124544,
      "grad_norm": 5.5634545235779065,
      "learning_rate": 1.1817058476354657e-06,
      "loss": 3.5693,
      "step": 1817
    },
    {
      "epoch": 0.2353626565685989,
      "grad_norm": 4.20888366597106,
      "learning_rate": 1.181577395507308e-06,
      "loss": 3.478,
      "step": 1818
    },
    {
      "epoch": 0.23549211897595235,
      "grad_norm": 5.3110630760512585,
      "learning_rate": 1.181448829686936e-06,
      "loss": 3.5986,
      "step": 1819
    },
    {
      "epoch": 0.23562158138330583,
      "grad_norm": 6.774465091539316,
      "learning_rate": 1.1813201502006116e-06,
      "loss": 3.5774,
      "step": 1820
    },
    {
      "epoch": 0.23562158138330583,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.596235752105713,
      "eval_runtime": 13.9516,
      "eval_samples_per_second": 3.154,
      "eval_steps_per_second": 0.215,
      "step": 1820
    },
    {
      "epoch": 0.23575104379065928,
      "grad_norm": 2.056453228202876,
      "learning_rate": 1.1811913570746205e-06,
      "loss": 2.6279,
      "step": 1821
    },
    {
      "epoch": 0.23588050619801276,
      "grad_norm": 3.2525689765267507,
      "learning_rate": 1.1810624503352713e-06,
      "loss": 3.0186,
      "step": 1822
    },
    {
      "epoch": 0.23600996860536622,
      "grad_norm": 8.605539012752132,
      "learning_rate": 1.1809334300088957e-06,
      "loss": 4.3545,
      "step": 1823
    },
    {
      "epoch": 0.23613943101271967,
      "grad_norm": 2.9217672232627896,
      "learning_rate": 1.1808042961218488e-06,
      "loss": 3.0693,
      "step": 1824
    },
    {
      "epoch": 0.23626889342007315,
      "grad_norm": 5.821165686943768,
      "learning_rate": 1.1806750487005086e-06,
      "loss": 4.2314,
      "step": 1825
    },
    {
      "epoch": 0.23626889342007315,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.599165439605713,
      "eval_runtime": 14.6408,
      "eval_samples_per_second": 3.005,
      "eval_steps_per_second": 0.205,
      "step": 1825
    },
    {
      "epoch": 0.2363983558274266,
      "grad_norm": 4.39105876666601,
      "learning_rate": 1.1805456877712769e-06,
      "loss": 3.46,
      "step": 1826
    },
    {
      "epoch": 0.23652781823478009,
      "grad_norm": 1.982333220196776,
      "learning_rate": 1.1804162133605781e-06,
      "loss": 2.709,
      "step": 1827
    },
    {
      "epoch": 0.23665728064213354,
      "grad_norm": 1.5761223394180242,
      "learning_rate": 1.1802866254948598e-06,
      "loss": 2.4607,
      "step": 1828
    },
    {
      "epoch": 0.236786743049487,
      "grad_norm": 5.235561279276746,
      "learning_rate": 1.180156924200593e-06,
      "loss": 3.4575,
      "step": 1829
    },
    {
      "epoch": 0.23691620545684047,
      "grad_norm": 4.6596904631208815,
      "learning_rate": 1.180027109504272e-06,
      "loss": 4.0371,
      "step": 1830
    },
    {
      "epoch": 0.23691620545684047,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.601606845855713,
      "eval_runtime": 14.3148,
      "eval_samples_per_second": 3.074,
      "eval_steps_per_second": 0.21,
      "step": 1830
    },
    {
      "epoch": 0.23704566786419393,
      "grad_norm": 2.825911278158945,
      "learning_rate": 1.1798971814324143e-06,
      "loss": 2.7859,
      "step": 1831
    },
    {
      "epoch": 0.2371751302715474,
      "grad_norm": 2.8793604036484863,
      "learning_rate": 1.1797671400115596e-06,
      "loss": 3.0991,
      "step": 1832
    },
    {
      "epoch": 0.23730459267890086,
      "grad_norm": 2.311542616946839,
      "learning_rate": 1.1796369852682722e-06,
      "loss": 3.1733,
      "step": 1833
    },
    {
      "epoch": 0.23743405508625434,
      "grad_norm": 1.6952525313065852,
      "learning_rate": 1.1795067172291385e-06,
      "loss": 2.6223,
      "step": 1834
    },
    {
      "epoch": 0.2375635174936078,
      "grad_norm": 2.5247398015838938,
      "learning_rate": 1.1793763359207686e-06,
      "loss": 2.9302,
      "step": 1835
    },
    {
      "epoch": 0.2375635174936078,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.605912685394287,
      "eval_runtime": 14.732,
      "eval_samples_per_second": 2.987,
      "eval_steps_per_second": 0.204,
      "step": 1835
    },
    {
      "epoch": 0.23769297990096125,
      "grad_norm": 1.9933049448504896,
      "learning_rate": 1.1792458413697951e-06,
      "loss": 2.7476,
      "step": 1836
    },
    {
      "epoch": 0.23782244230831473,
      "grad_norm": 4.7162128598490165,
      "learning_rate": 1.1791152336028746e-06,
      "loss": 3.4561,
      "step": 1837
    },
    {
      "epoch": 0.23795190471566818,
      "grad_norm": 1.943023123705872,
      "learning_rate": 1.1789845126466864e-06,
      "loss": 2.7314,
      "step": 1838
    },
    {
      "epoch": 0.23808136712302166,
      "grad_norm": 2.9772015630697406,
      "learning_rate": 1.1788536785279325e-06,
      "loss": 3.1687,
      "step": 1839
    },
    {
      "epoch": 0.23821082953037512,
      "grad_norm": 3.320013254275291,
      "learning_rate": 1.1787227312733388e-06,
      "loss": 3.2031,
      "step": 1840
    },
    {
      "epoch": 0.23821082953037512,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6009188890457153,
      "eval_runtime": 14.8977,
      "eval_samples_per_second": 2.953,
      "eval_steps_per_second": 0.201,
      "step": 1840
    },
    {
      "epoch": 0.23834029193772857,
      "grad_norm": 3.8325170390730525,
      "learning_rate": 1.1785916709096538e-06,
      "loss": 3.0173,
      "step": 1841
    },
    {
      "epoch": 0.23846975434508205,
      "grad_norm": 2.0805028534732237,
      "learning_rate": 1.178460497463649e-06,
      "loss": 2.8025,
      "step": 1842
    },
    {
      "epoch": 0.2385992167524355,
      "grad_norm": 4.461425848651309,
      "learning_rate": 1.1783292109621196e-06,
      "loss": 3.2861,
      "step": 1843
    },
    {
      "epoch": 0.238728679159789,
      "grad_norm": 2.1575720780987244,
      "learning_rate": 1.1781978114318835e-06,
      "loss": 2.9761,
      "step": 1844
    },
    {
      "epoch": 0.23885814156714244,
      "grad_norm": 4.289045229317149,
      "learning_rate": 1.1780662988997814e-06,
      "loss": 3.1362,
      "step": 1845
    },
    {
      "epoch": 0.23885814156714244,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6029163599014282,
      "eval_runtime": 14.0891,
      "eval_samples_per_second": 3.123,
      "eval_steps_per_second": 0.213,
      "step": 1845
    },
    {
      "epoch": 0.2389876039744959,
      "grad_norm": 2.2589644223106284,
      "learning_rate": 1.1779346733926778e-06,
      "loss": 2.7209,
      "step": 1846
    },
    {
      "epoch": 0.23911706638184937,
      "grad_norm": 3.9750453551208715,
      "learning_rate": 1.1778029349374597e-06,
      "loss": 3.3169,
      "step": 1847
    },
    {
      "epoch": 0.23924652878920283,
      "grad_norm": 1.7815804958225263,
      "learning_rate": 1.1776710835610371e-06,
      "loss": 2.8086,
      "step": 1848
    },
    {
      "epoch": 0.2393759911965563,
      "grad_norm": 5.351839327909766,
      "learning_rate": 1.1775391192903437e-06,
      "loss": 3.6738,
      "step": 1849
    },
    {
      "epoch": 0.23950545360390976,
      "grad_norm": 3.0449422128513954,
      "learning_rate": 1.1774070421523358e-06,
      "loss": 3.0908,
      "step": 1850
    },
    {
      "epoch": 0.23950545360390976,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6037375926971436,
      "eval_runtime": 13.198,
      "eval_samples_per_second": 3.334,
      "eval_steps_per_second": 0.227,
      "step": 1850
    },
    {
      "epoch": 0.23963491601126322,
      "grad_norm": 2.2416079880456437,
      "learning_rate": 1.1772748521739926e-06,
      "loss": 3.0264,
      "step": 1851
    },
    {
      "epoch": 0.2397643784186167,
      "grad_norm": 3.8526460107748517,
      "learning_rate": 1.1771425493823168e-06,
      "loss": 3.0195,
      "step": 1852
    },
    {
      "epoch": 0.23989384082597015,
      "grad_norm": 4.330167610957246,
      "learning_rate": 1.177010133804334e-06,
      "loss": 3.4597,
      "step": 1853
    },
    {
      "epoch": 0.24002330323332363,
      "grad_norm": 3.702924740714124,
      "learning_rate": 1.1768776054670924e-06,
      "loss": 2.8242,
      "step": 1854
    },
    {
      "epoch": 0.24015276564067708,
      "grad_norm": 2.86621389616258,
      "learning_rate": 1.1767449643976641e-06,
      "loss": 3.1323,
      "step": 1855
    },
    {
      "epoch": 0.24015276564067708,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6049138307571411,
      "eval_runtime": 14.3038,
      "eval_samples_per_second": 3.076,
      "eval_steps_per_second": 0.21,
      "step": 1855
    },
    {
      "epoch": 0.24028222804803057,
      "grad_norm": 6.8781213442945,
      "learning_rate": 1.1766122106231432e-06,
      "loss": 3.3423,
      "step": 1856
    },
    {
      "epoch": 0.24041169045538402,
      "grad_norm": 4.969605794255556,
      "learning_rate": 1.1764793441706478e-06,
      "loss": 3.6226,
      "step": 1857
    },
    {
      "epoch": 0.24054115286273747,
      "grad_norm": 1.6546195148470562,
      "learning_rate": 1.1763463650673182e-06,
      "loss": 2.7083,
      "step": 1858
    },
    {
      "epoch": 0.24067061527009095,
      "grad_norm": 6.188724141215013,
      "learning_rate": 1.1762132733403184e-06,
      "loss": 3.4285,
      "step": 1859
    },
    {
      "epoch": 0.2408000776774444,
      "grad_norm": 4.707623508261219,
      "learning_rate": 1.1760800690168349e-06,
      "loss": 3.4709,
      "step": 1860
    },
    {
      "epoch": 0.2408000776774444,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6134144067764282,
      "eval_runtime": 14.6448,
      "eval_samples_per_second": 3.004,
      "eval_steps_per_second": 0.205,
      "step": 1860
    },
    {
      "epoch": 0.2409295400847979,
      "grad_norm": 4.66981496391424,
      "learning_rate": 1.1759467521240774e-06,
      "loss": 3.5918,
      "step": 1861
    },
    {
      "epoch": 0.24105900249215134,
      "grad_norm": 2.818593930693954,
      "learning_rate": 1.1758133226892786e-06,
      "loss": 3.3066,
      "step": 1862
    },
    {
      "epoch": 0.2411884648995048,
      "grad_norm": 3.3506448891288656,
      "learning_rate": 1.1756797807396941e-06,
      "loss": 3.282,
      "step": 1863
    },
    {
      "epoch": 0.24131792730685828,
      "grad_norm": 3.9180330696979957,
      "learning_rate": 1.1755461263026028e-06,
      "loss": 3.4312,
      "step": 1864
    },
    {
      "epoch": 0.24144738971421173,
      "grad_norm": 2.7474241675121505,
      "learning_rate": 1.175412359405306e-06,
      "loss": 3.1055,
      "step": 1865
    },
    {
      "epoch": 0.24144738971421173,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6133478879928589,
      "eval_runtime": 14.4666,
      "eval_samples_per_second": 3.041,
      "eval_steps_per_second": 0.207,
      "step": 1865
    },
    {
      "epoch": 0.2415768521215652,
      "grad_norm": 1.6426711111516288,
      "learning_rate": 1.1752784800751285e-06,
      "loss": 2.4155,
      "step": 1866
    },
    {
      "epoch": 0.24170631452891866,
      "grad_norm": 1.343580295689865,
      "learning_rate": 1.1751444883394177e-06,
      "loss": 2.5276,
      "step": 1867
    },
    {
      "epoch": 0.24183577693627212,
      "grad_norm": 2.6167222392001936,
      "learning_rate": 1.1750103842255443e-06,
      "loss": 2.6816,
      "step": 1868
    },
    {
      "epoch": 0.2419652393436256,
      "grad_norm": 2.1113031892678777,
      "learning_rate": 1.1748761677609018e-06,
      "loss": 2.7498,
      "step": 1869
    },
    {
      "epoch": 0.24209470175097905,
      "grad_norm": 2.3298636724049393,
      "learning_rate": 1.1747418389729066e-06,
      "loss": 3.0835,
      "step": 1870
    },
    {
      "epoch": 0.24209470175097905,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6146794557571411,
      "eval_runtime": 14.9518,
      "eval_samples_per_second": 2.943,
      "eval_steps_per_second": 0.201,
      "step": 1870
    },
    {
      "epoch": 0.24222416415833253,
      "grad_norm": 3.7112023411479504,
      "learning_rate": 1.1746073978889977e-06,
      "loss": 3.1948,
      "step": 1871
    },
    {
      "epoch": 0.242353626565686,
      "grad_norm": 5.937930557908351,
      "learning_rate": 1.174472844536638e-06,
      "loss": 3.7815,
      "step": 1872
    },
    {
      "epoch": 0.24248308897303944,
      "grad_norm": 4.696990897913837,
      "learning_rate": 1.1743381789433126e-06,
      "loss": 3.0894,
      "step": 1873
    },
    {
      "epoch": 0.24261255138039292,
      "grad_norm": 4.737390688559019,
      "learning_rate": 1.1742034011365293e-06,
      "loss": 3.3867,
      "step": 1874
    },
    {
      "epoch": 0.24274201378774637,
      "grad_norm": 2.745308506758466,
      "learning_rate": 1.1740685111438196e-06,
      "loss": 3.1606,
      "step": 1875
    },
    {
      "epoch": 0.24274201378774637,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.615922451019287,
      "eval_runtime": 14.3167,
      "eval_samples_per_second": 3.073,
      "eval_steps_per_second": 0.21,
      "step": 1875
    },
    {
      "epoch": 0.24287147619509986,
      "grad_norm": 6.549932183153616,
      "learning_rate": 1.1739335089927373e-06,
      "loss": 3.4863,
      "step": 1876
    },
    {
      "epoch": 0.2430009386024533,
      "grad_norm": 3.359160123814084,
      "learning_rate": 1.1737983947108595e-06,
      "loss": 2.9927,
      "step": 1877
    },
    {
      "epoch": 0.2431304010098068,
      "grad_norm": 3.4392212569066025,
      "learning_rate": 1.173663168325786e-06,
      "loss": 3.3003,
      "step": 1878
    },
    {
      "epoch": 0.24325986341716024,
      "grad_norm": 4.252401229416216,
      "learning_rate": 1.1735278298651395e-06,
      "loss": 3.1992,
      "step": 1879
    },
    {
      "epoch": 0.2433893258245137,
      "grad_norm": 3.3023280095272245,
      "learning_rate": 1.1733923793565656e-06,
      "loss": 3.0347,
      "step": 1880
    },
    {
      "epoch": 0.2433893258245137,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.620893955230713,
      "eval_runtime": 14.2206,
      "eval_samples_per_second": 3.094,
      "eval_steps_per_second": 0.211,
      "step": 1880
    },
    {
      "epoch": 0.24351878823186718,
      "grad_norm": 4.540014857907009,
      "learning_rate": 1.1732568168277326e-06,
      "loss": 3.3464,
      "step": 1881
    },
    {
      "epoch": 0.24364825063922063,
      "grad_norm": 3.251888789683562,
      "learning_rate": 1.1731211423063323e-06,
      "loss": 2.8691,
      "step": 1882
    },
    {
      "epoch": 0.2437777130465741,
      "grad_norm": 4.586408463329068,
      "learning_rate": 1.1729853558200786e-06,
      "loss": 3.3435,
      "step": 1883
    },
    {
      "epoch": 0.24390717545392757,
      "grad_norm": 3.2370875789770337,
      "learning_rate": 1.172849457396709e-06,
      "loss": 2.8208,
      "step": 1884
    },
    {
      "epoch": 0.24403663786128102,
      "grad_norm": 3.095421010617364,
      "learning_rate": 1.1727134470639832e-06,
      "loss": 3.1907,
      "step": 1885
    },
    {
      "epoch": 0.24403663786128102,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6202059984207153,
      "eval_runtime": 14.3363,
      "eval_samples_per_second": 3.069,
      "eval_steps_per_second": 0.209,
      "step": 1885
    },
    {
      "epoch": 0.2441661002686345,
      "grad_norm": 3.1517139071583298,
      "learning_rate": 1.172577324849684e-06,
      "loss": 2.7622,
      "step": 1886
    },
    {
      "epoch": 0.24429556267598795,
      "grad_norm": 7.108209574349753,
      "learning_rate": 1.1724410907816178e-06,
      "loss": 3.8677,
      "step": 1887
    },
    {
      "epoch": 0.24442502508334144,
      "grad_norm": 2.061805249129852,
      "learning_rate": 1.1723047448876124e-06,
      "loss": 2.5483,
      "step": 1888
    },
    {
      "epoch": 0.2445544874906949,
      "grad_norm": 4.517874299808877,
      "learning_rate": 1.1721682871955196e-06,
      "loss": 2.8652,
      "step": 1889
    },
    {
      "epoch": 0.24468394989804834,
      "grad_norm": 3.952170357823975,
      "learning_rate": 1.1720317177332132e-06,
      "loss": 3.1326,
      "step": 1890
    },
    {
      "epoch": 0.24468394989804834,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.62841796875,
      "eval_runtime": 14.544,
      "eval_samples_per_second": 3.025,
      "eval_steps_per_second": 0.206,
      "step": 1890
    },
    {
      "epoch": 0.24481341230540182,
      "grad_norm": 3.8008526617129292,
      "learning_rate": 1.171895036528591e-06,
      "loss": 3.4922,
      "step": 1891
    },
    {
      "epoch": 0.24494287471275528,
      "grad_norm": 2.225180960314294,
      "learning_rate": 1.1717582436095725e-06,
      "loss": 2.6582,
      "step": 1892
    },
    {
      "epoch": 0.24507233712010876,
      "grad_norm": 2.322052374551846,
      "learning_rate": 1.1716213390041004e-06,
      "loss": 3.0005,
      "step": 1893
    },
    {
      "epoch": 0.2452017995274622,
      "grad_norm": 4.145224618409552,
      "learning_rate": 1.1714843227401404e-06,
      "loss": 3.8025,
      "step": 1894
    },
    {
      "epoch": 0.24533126193481566,
      "grad_norm": 2.659412652836115,
      "learning_rate": 1.1713471948456804e-06,
      "loss": 2.9658,
      "step": 1895
    },
    {
      "epoch": 0.24533126193481566,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6312366724014282,
      "eval_runtime": 15.3092,
      "eval_samples_per_second": 2.874,
      "eval_steps_per_second": 0.196,
      "step": 1895
    },
    {
      "epoch": 0.24546072434216915,
      "grad_norm": 3.10138403182918,
      "learning_rate": 1.1712099553487323e-06,
      "loss": 2.8525,
      "step": 1896
    },
    {
      "epoch": 0.2455901867495226,
      "grad_norm": 3.3659064883380463,
      "learning_rate": 1.1710726042773294e-06,
      "loss": 2.6694,
      "step": 1897
    },
    {
      "epoch": 0.24571964915687608,
      "grad_norm": 6.495870574274998,
      "learning_rate": 1.1709351416595288e-06,
      "loss": 3.8931,
      "step": 1898
    },
    {
      "epoch": 0.24584911156422953,
      "grad_norm": 3.898914673426884,
      "learning_rate": 1.1707975675234099e-06,
      "loss": 2.75,
      "step": 1899
    },
    {
      "epoch": 0.24597857397158301,
      "grad_norm": 5.218508613864783,
      "learning_rate": 1.170659881897075e-06,
      "loss": 3.2034,
      "step": 1900
    },
    {
      "epoch": 0.24597857397158301,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.634721279144287,
      "eval_runtime": 14.7638,
      "eval_samples_per_second": 2.98,
      "eval_steps_per_second": 0.203,
      "step": 1900
    },
    {
      "epoch": 0.24610803637893647,
      "grad_norm": 3.549443092089115,
      "learning_rate": 1.1705220848086491e-06,
      "loss": 3.0303,
      "step": 1901
    },
    {
      "epoch": 0.24623749878628992,
      "grad_norm": 5.924032827398414,
      "learning_rate": 1.1703841762862802e-06,
      "loss": 3.3115,
      "step": 1902
    },
    {
      "epoch": 0.2463669611936434,
      "grad_norm": 4.181701413823407,
      "learning_rate": 1.170246156358139e-06,
      "loss": 3.2271,
      "step": 1903
    },
    {
      "epoch": 0.24649642360099686,
      "grad_norm": 2.3167449836917435,
      "learning_rate": 1.1701080250524187e-06,
      "loss": 2.7825,
      "step": 1904
    },
    {
      "epoch": 0.24662588600835034,
      "grad_norm": 1.5241933551793072,
      "learning_rate": 1.1699697823973354e-06,
      "loss": 2.3547,
      "step": 1905
    },
    {
      "epoch": 0.24662588600835034,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6404030323028564,
      "eval_runtime": 15.0272,
      "eval_samples_per_second": 2.928,
      "eval_steps_per_second": 0.2,
      "step": 1905
    },
    {
      "epoch": 0.2467553484157038,
      "grad_norm": 5.322937685840939,
      "learning_rate": 1.169831428421128e-06,
      "loss": 3.0752,
      "step": 1906
    },
    {
      "epoch": 0.24688481082305724,
      "grad_norm": 5.759678630636752,
      "learning_rate": 1.169692963152058e-06,
      "loss": 3.2339,
      "step": 1907
    },
    {
      "epoch": 0.24701427323041072,
      "grad_norm": 3.206134709658607,
      "learning_rate": 1.16955438661841e-06,
      "loss": 2.9053,
      "step": 1908
    },
    {
      "epoch": 0.24714373563776418,
      "grad_norm": 3.530220628237768,
      "learning_rate": 1.1694156988484909e-06,
      "loss": 3.4482,
      "step": 1909
    },
    {
      "epoch": 0.24727319804511766,
      "grad_norm": 3.483503154963089,
      "learning_rate": 1.1692768998706306e-06,
      "loss": 3.0337,
      "step": 1910
    },
    {
      "epoch": 0.24727319804511766,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6439098119735718,
      "eval_runtime": 14.2668,
      "eval_samples_per_second": 3.084,
      "eval_steps_per_second": 0.21,
      "step": 1910
    },
    {
      "epoch": 0.2474026604524711,
      "grad_norm": 4.352787043277252,
      "learning_rate": 1.1691379897131814e-06,
      "loss": 3.0293,
      "step": 1911
    },
    {
      "epoch": 0.24753212285982457,
      "grad_norm": 2.3934729366186818,
      "learning_rate": 1.1689989684045186e-06,
      "loss": 2.9399,
      "step": 1912
    },
    {
      "epoch": 0.24766158526717805,
      "grad_norm": 2.2565153309055437,
      "learning_rate": 1.1688598359730401e-06,
      "loss": 3.0776,
      "step": 1913
    },
    {
      "epoch": 0.2477910476745315,
      "grad_norm": 2.215045688960175,
      "learning_rate": 1.1687205924471668e-06,
      "loss": 2.6819,
      "step": 1914
    },
    {
      "epoch": 0.24792051008188498,
      "grad_norm": 2.571773481207664,
      "learning_rate": 1.1685812378553416e-06,
      "loss": 2.9763,
      "step": 1915
    },
    {
      "epoch": 0.24792051008188498,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.645751953125,
      "eval_runtime": 13.988,
      "eval_samples_per_second": 3.146,
      "eval_steps_per_second": 0.214,
      "step": 1915
    },
    {
      "epoch": 0.24804997248923844,
      "grad_norm": 4.5284448031533815,
      "learning_rate": 1.1684417722260306e-06,
      "loss": 2.9907,
      "step": 1916
    },
    {
      "epoch": 0.2481794348965919,
      "grad_norm": 4.239696600057911,
      "learning_rate": 1.1683021955877226e-06,
      "loss": 3.3342,
      "step": 1917
    },
    {
      "epoch": 0.24830889730394537,
      "grad_norm": 1.4685864204533403,
      "learning_rate": 1.168162507968929e-06,
      "loss": 2.498,
      "step": 1918
    },
    {
      "epoch": 0.24843835971129882,
      "grad_norm": 3.123063655653541,
      "learning_rate": 1.1680227093981838e-06,
      "loss": 3.2268,
      "step": 1919
    },
    {
      "epoch": 0.2485678221186523,
      "grad_norm": 4.049662140273082,
      "learning_rate": 1.1678827999040434e-06,
      "loss": 3.3848,
      "step": 1920
    },
    {
      "epoch": 0.2485678221186523,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6492587327957153,
      "eval_runtime": 13.6457,
      "eval_samples_per_second": 3.224,
      "eval_steps_per_second": 0.22,
      "step": 1920
    },
    {
      "epoch": 0.24869728452600576,
      "grad_norm": 3.1032801751811925,
      "learning_rate": 1.1677427795150874e-06,
      "loss": 2.9585,
      "step": 1921
    },
    {
      "epoch": 0.24882674693335924,
      "grad_norm": 5.033863510907717,
      "learning_rate": 1.1676026482599181e-06,
      "loss": 3.1479,
      "step": 1922
    },
    {
      "epoch": 0.2489562093407127,
      "grad_norm": 2.329811089621362,
      "learning_rate": 1.1674624061671597e-06,
      "loss": 3.0049,
      "step": 1923
    },
    {
      "epoch": 0.24908567174806615,
      "grad_norm": 3.6429543909690683,
      "learning_rate": 1.1673220532654596e-06,
      "loss": 3.3096,
      "step": 1924
    },
    {
      "epoch": 0.24921513415541963,
      "grad_norm": 3.425761242416327,
      "learning_rate": 1.1671815895834878e-06,
      "loss": 2.9817,
      "step": 1925
    },
    {
      "epoch": 0.24921513415541963,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.64990234375,
      "eval_runtime": 13.9752,
      "eval_samples_per_second": 3.148,
      "eval_steps_per_second": 0.215,
      "step": 1925
    },
    {
      "epoch": 0.24934459656277308,
      "grad_norm": 2.3724580370443267,
      "learning_rate": 1.1670410151499367e-06,
      "loss": 2.9553,
      "step": 1926
    },
    {
      "epoch": 0.24947405897012656,
      "grad_norm": 3.7303337256752234,
      "learning_rate": 1.1669003299935216e-06,
      "loss": 2.9834,
      "step": 1927
    },
    {
      "epoch": 0.24960352137748001,
      "grad_norm": 1.8399631563706738,
      "learning_rate": 1.1667595341429804e-06,
      "loss": 2.5762,
      "step": 1928
    },
    {
      "epoch": 0.24973298378483347,
      "grad_norm": 2.4346348505121376,
      "learning_rate": 1.1666186276270732e-06,
      "loss": 2.8062,
      "step": 1929
    },
    {
      "epoch": 0.24986244619218695,
      "grad_norm": 3.129114112134674,
      "learning_rate": 1.1664776104745833e-06,
      "loss": 2.9463,
      "step": 1930
    },
    {
      "epoch": 0.24986244619218695,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.654296875,
      "eval_runtime": 14.3271,
      "eval_samples_per_second": 3.071,
      "eval_steps_per_second": 0.209,
      "step": 1930
    },
    {
      "epoch": 0.2499919085995404,
      "grad_norm": 4.006686316110688,
      "learning_rate": 1.1663364827143164e-06,
      "loss": 3.2124,
      "step": 1931
    },
    {
      "epoch": 0.25012137100689386,
      "grad_norm": 2.148758338978397,
      "learning_rate": 1.1661952443751003e-06,
      "loss": 3.1199,
      "step": 1932
    },
    {
      "epoch": 0.25025083341424736,
      "grad_norm": 6.715887482720353,
      "learning_rate": 1.166053895485786e-06,
      "loss": 3.7588,
      "step": 1933
    },
    {
      "epoch": 0.2503802958216008,
      "grad_norm": 1.7416847813457657,
      "learning_rate": 1.165912436075247e-06,
      "loss": 2.707,
      "step": 1934
    },
    {
      "epoch": 0.25050975822895427,
      "grad_norm": 2.9003109139957655,
      "learning_rate": 1.165770866172379e-06,
      "loss": 2.9849,
      "step": 1935
    },
    {
      "epoch": 0.25050975822895427,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6504572629928589,
      "eval_runtime": 14.3051,
      "eval_samples_per_second": 3.076,
      "eval_steps_per_second": 0.21,
      "step": 1935
    },
    {
      "epoch": 0.2506392206363077,
      "grad_norm": 4.424929347402928,
      "learning_rate": 1.1656291858061007e-06,
      "loss": 3.0566,
      "step": 1936
    },
    {
      "epoch": 0.2507686830436612,
      "grad_norm": 3.890277300436681,
      "learning_rate": 1.165487395005353e-06,
      "loss": 3.7363,
      "step": 1937
    },
    {
      "epoch": 0.2508981454510147,
      "grad_norm": 4.396164022571692,
      "learning_rate": 1.1653454937990998e-06,
      "loss": 3.1294,
      "step": 1938
    },
    {
      "epoch": 0.25102760785836814,
      "grad_norm": 2.7710371824078117,
      "learning_rate": 1.1652034822163274e-06,
      "loss": 2.6621,
      "step": 1939
    },
    {
      "epoch": 0.2511570702657216,
      "grad_norm": 4.940295711783432,
      "learning_rate": 1.1650613602860442e-06,
      "loss": 3.7119,
      "step": 1940
    },
    {
      "epoch": 0.2511570702657216,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6569157838821411,
      "eval_runtime": 14.436,
      "eval_samples_per_second": 3.048,
      "eval_steps_per_second": 0.208,
      "step": 1940
    },
    {
      "epoch": 0.25128653267307505,
      "grad_norm": 5.206647762855925,
      "learning_rate": 1.1649191280372815e-06,
      "loss": 2.8794,
      "step": 1941
    },
    {
      "epoch": 0.2514159950804285,
      "grad_norm": 3.819919695496243,
      "learning_rate": 1.1647767854990934e-06,
      "loss": 3.0488,
      "step": 1942
    },
    {
      "epoch": 0.251545457487782,
      "grad_norm": 6.460455605701821,
      "learning_rate": 1.164634332700556e-06,
      "loss": 3.4609,
      "step": 1943
    },
    {
      "epoch": 0.25167491989513546,
      "grad_norm": 2.791025876629594,
      "learning_rate": 1.1644917696707683e-06,
      "loss": 2.5996,
      "step": 1944
    },
    {
      "epoch": 0.2518043823024889,
      "grad_norm": 5.080405747007676,
      "learning_rate": 1.1643490964388516e-06,
      "loss": 2.8545,
      "step": 1945
    },
    {
      "epoch": 0.2518043823024889,
      "eval_PRM Accuracy": 0.24528301886792453,
      "eval_PRM F1": 0.06976744186046512,
      "eval_PRM F1 AUC": 0.5180722891566265,
      "eval_PRM F1 Neg": 0.36507936507936506,
      "eval_PRM NPV": 0.22330097087378642,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.03614457831325301,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6588023900985718,
      "eval_runtime": 14.35,
      "eval_samples_per_second": 3.066,
      "eval_steps_per_second": 0.209,
      "step": 1945
    },
    {
      "epoch": 0.25193384470984237,
      "grad_norm": 4.32648225151238,
      "learning_rate": 1.1642063130339499e-06,
      "loss": 3.1641,
      "step": 1946
    },
    {
      "epoch": 0.2520633071171958,
      "grad_norm": 8.471658791981463,
      "learning_rate": 1.1640634194852296e-06,
      "loss": 4.6089,
      "step": 1947
    },
    {
      "epoch": 0.25219276952454933,
      "grad_norm": 2.5457007924993973,
      "learning_rate": 1.1639204158218793e-06,
      "loss": 3.0625,
      "step": 1948
    },
    {
      "epoch": 0.2523222319319028,
      "grad_norm": 3.822627469940579,
      "learning_rate": 1.1637773020731108e-06,
      "loss": 3.0161,
      "step": 1949
    },
    {
      "epoch": 0.25245169433925624,
      "grad_norm": 5.080701908629334,
      "learning_rate": 1.1636340782681578e-06,
      "loss": 3.28,
      "step": 1950
    },
    {
      "epoch": 0.25245169433925624,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6610440015792847,
      "eval_runtime": 14.35,
      "eval_samples_per_second": 3.066,
      "eval_steps_per_second": 0.209,
      "step": 1950
    },
    {
      "epoch": 0.2525811567466097,
      "grad_norm": 2.6845066594555678,
      "learning_rate": 1.1634907444362767e-06,
      "loss": 3.0469,
      "step": 1951
    },
    {
      "epoch": 0.25271061915396315,
      "grad_norm": 3.1470433673561042,
      "learning_rate": 1.1633473006067464e-06,
      "loss": 2.9512,
      "step": 1952
    },
    {
      "epoch": 0.25284008156131665,
      "grad_norm": 2.1104458544901044,
      "learning_rate": 1.1632037468088682e-06,
      "loss": 2.8396,
      "step": 1953
    },
    {
      "epoch": 0.2529695439686701,
      "grad_norm": 3.0694203591952545,
      "learning_rate": 1.1630600830719657e-06,
      "loss": 3.1455,
      "step": 1954
    },
    {
      "epoch": 0.25309900637602356,
      "grad_norm": 3.054022978685594,
      "learning_rate": 1.1629163094253853e-06,
      "loss": 3.1145,
      "step": 1955
    },
    {
      "epoch": 0.25309900637602356,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6629971265792847,
      "eval_runtime": 14.6532,
      "eval_samples_per_second": 3.003,
      "eval_steps_per_second": 0.205,
      "step": 1955
    },
    {
      "epoch": 0.253228468783377,
      "grad_norm": 2.59793924117539,
      "learning_rate": 1.1627724258984956e-06,
      "loss": 2.6399,
      "step": 1956
    },
    {
      "epoch": 0.25335793119073047,
      "grad_norm": 2.9695030332991554,
      "learning_rate": 1.1626284325206878e-06,
      "loss": 2.9143,
      "step": 1957
    },
    {
      "epoch": 0.253487393598084,
      "grad_norm": 3.5658172072979224,
      "learning_rate": 1.1624843293213754e-06,
      "loss": 2.9331,
      "step": 1958
    },
    {
      "epoch": 0.25361685600543743,
      "grad_norm": 6.02672060693311,
      "learning_rate": 1.1623401163299944e-06,
      "loss": 3.7141,
      "step": 1959
    },
    {
      "epoch": 0.2537463184127909,
      "grad_norm": 2.5545131595119943,
      "learning_rate": 1.1621957935760033e-06,
      "loss": 2.7351,
      "step": 1960
    },
    {
      "epoch": 0.2537463184127909,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6679909229278564,
      "eval_runtime": 15.1648,
      "eval_samples_per_second": 2.901,
      "eval_steps_per_second": 0.198,
      "step": 1960
    },
    {
      "epoch": 0.25387578082014434,
      "grad_norm": 2.827112950035107,
      "learning_rate": 1.1620513610888829e-06,
      "loss": 3.3955,
      "step": 1961
    },
    {
      "epoch": 0.2540052432274978,
      "grad_norm": 3.1636385852004514,
      "learning_rate": 1.1619068188981366e-06,
      "loss": 2.9683,
      "step": 1962
    },
    {
      "epoch": 0.2541347056348513,
      "grad_norm": 3.7404378662813853,
      "learning_rate": 1.1617621670332896e-06,
      "loss": 3.314,
      "step": 1963
    },
    {
      "epoch": 0.25426416804220475,
      "grad_norm": 4.352457415208915,
      "learning_rate": 1.1616174055238907e-06,
      "loss": 3.3213,
      "step": 1964
    },
    {
      "epoch": 0.2543936304495582,
      "grad_norm": 2.1336260965762457,
      "learning_rate": 1.1614725343995096e-06,
      "loss": 2.5703,
      "step": 1965
    },
    {
      "epoch": 0.2543936304495582,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6664817333221436,
      "eval_runtime": 13.9433,
      "eval_samples_per_second": 3.156,
      "eval_steps_per_second": 0.215,
      "step": 1965
    },
    {
      "epoch": 0.25452309285691166,
      "grad_norm": 2.4535077176087063,
      "learning_rate": 1.1613275536897397e-06,
      "loss": 3.0042,
      "step": 1966
    },
    {
      "epoch": 0.2546525552642651,
      "grad_norm": 2.031682185170567,
      "learning_rate": 1.1611824634241962e-06,
      "loss": 2.9478,
      "step": 1967
    },
    {
      "epoch": 0.2547820176716186,
      "grad_norm": 3.261890774656847,
      "learning_rate": 1.1610372636325165e-06,
      "loss": 2.9756,
      "step": 1968
    },
    {
      "epoch": 0.2549114800789721,
      "grad_norm": 4.095008951501019,
      "learning_rate": 1.1608919543443608e-06,
      "loss": 2.8286,
      "step": 1969
    },
    {
      "epoch": 0.25504094248632553,
      "grad_norm": 3.7551506961258863,
      "learning_rate": 1.1607465355894114e-06,
      "loss": 3.4907,
      "step": 1970
    },
    {
      "epoch": 0.25504094248632553,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6677912473678589,
      "eval_runtime": 14.5067,
      "eval_samples_per_second": 3.033,
      "eval_steps_per_second": 0.207,
      "step": 1970
    },
    {
      "epoch": 0.255170404893679,
      "grad_norm": 5.452217070822066,
      "learning_rate": 1.160601007397373e-06,
      "loss": 3.9702,
      "step": 1971
    },
    {
      "epoch": 0.2552998673010325,
      "grad_norm": 1.9282831671714697,
      "learning_rate": 1.1604553697979725e-06,
      "loss": 2.7742,
      "step": 1972
    },
    {
      "epoch": 0.25542932970838594,
      "grad_norm": 2.7401274261259125,
      "learning_rate": 1.1603096228209599e-06,
      "loss": 2.7227,
      "step": 1973
    },
    {
      "epoch": 0.2555587921157394,
      "grad_norm": 2.872327736649136,
      "learning_rate": 1.1601637664961063e-06,
      "loss": 2.9717,
      "step": 1974
    },
    {
      "epoch": 0.25568825452309285,
      "grad_norm": 4.785932472762122,
      "learning_rate": 1.1600178008532062e-06,
      "loss": 3.0,
      "step": 1975
    },
    {
      "epoch": 0.25568825452309285,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6714088916778564,
      "eval_runtime": 15.8787,
      "eval_samples_per_second": 2.771,
      "eval_steps_per_second": 0.189,
      "step": 1975
    },
    {
      "epoch": 0.2558177169304463,
      "grad_norm": 3.7662031622478596,
      "learning_rate": 1.1598717259220762e-06,
      "loss": 2.936,
      "step": 1976
    },
    {
      "epoch": 0.2559471793377998,
      "grad_norm": 2.055785163278057,
      "learning_rate": 1.1597255417325544e-06,
      "loss": 2.6257,
      "step": 1977
    },
    {
      "epoch": 0.25607664174515327,
      "grad_norm": 2.3292145666544517,
      "learning_rate": 1.1595792483145027e-06,
      "loss": 2.823,
      "step": 1978
    },
    {
      "epoch": 0.2562061041525067,
      "grad_norm": 2.5003043444882813,
      "learning_rate": 1.1594328456978038e-06,
      "loss": 2.6716,
      "step": 1979
    },
    {
      "epoch": 0.2563355665598602,
      "grad_norm": 2.226384507862159,
      "learning_rate": 1.159286333912364e-06,
      "loss": 2.7947,
      "step": 1980
    },
    {
      "epoch": 0.2563355665598602,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.671431064605713,
      "eval_runtime": 14.8882,
      "eval_samples_per_second": 2.955,
      "eval_steps_per_second": 0.202,
      "step": 1980
    },
    {
      "epoch": 0.2564650289672136,
      "grad_norm": 2.371023306951928,
      "learning_rate": 1.1591397129881106e-06,
      "loss": 2.7693,
      "step": 1981
    },
    {
      "epoch": 0.25659449137456714,
      "grad_norm": 9.343558328232445,
      "learning_rate": 1.1589929829549947e-06,
      "loss": 4.0332,
      "step": 1982
    },
    {
      "epoch": 0.2567239537819206,
      "grad_norm": 4.712377500802992,
      "learning_rate": 1.1588461438429882e-06,
      "loss": 2.9692,
      "step": 1983
    },
    {
      "epoch": 0.25685341618927404,
      "grad_norm": 2.62681756261129,
      "learning_rate": 1.1586991956820864e-06,
      "loss": 3.2104,
      "step": 1984
    },
    {
      "epoch": 0.2569828785966275,
      "grad_norm": 4.675799683507873,
      "learning_rate": 1.1585521385023062e-06,
      "loss": 3.2573,
      "step": 1985
    },
    {
      "epoch": 0.2569828785966275,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6739612817764282,
      "eval_runtime": 14.5772,
      "eval_samples_per_second": 3.018,
      "eval_steps_per_second": 0.206,
      "step": 1985
    },
    {
      "epoch": 0.25711234100398095,
      "grad_norm": 4.361973393201989,
      "learning_rate": 1.1584049723336874e-06,
      "loss": 3.1494,
      "step": 1986
    },
    {
      "epoch": 0.25724180341133446,
      "grad_norm": 2.5864916173219825,
      "learning_rate": 1.1582576972062912e-06,
      "loss": 2.7046,
      "step": 1987
    },
    {
      "epoch": 0.2573712658186879,
      "grad_norm": 4.057478533906205,
      "learning_rate": 1.1581103131502016e-06,
      "loss": 3.0259,
      "step": 1988
    },
    {
      "epoch": 0.25750072822604136,
      "grad_norm": 2.3452153576282764,
      "learning_rate": 1.1579628201955248e-06,
      "loss": 2.9102,
      "step": 1989
    },
    {
      "epoch": 0.2576301906333948,
      "grad_norm": 4.00776815918346,
      "learning_rate": 1.1578152183723895e-06,
      "loss": 3.085,
      "step": 1990
    },
    {
      "epoch": 0.2576301906333948,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6740056276321411,
      "eval_runtime": 14.3788,
      "eval_samples_per_second": 3.06,
      "eval_steps_per_second": 0.209,
      "step": 1990
    },
    {
      "epoch": 0.25775965304074827,
      "grad_norm": 1.627012938499294,
      "learning_rate": 1.1576675077109462e-06,
      "loss": 2.7305,
      "step": 1991
    },
    {
      "epoch": 0.2578891154481018,
      "grad_norm": 2.5493138688048798,
      "learning_rate": 1.1575196882413677e-06,
      "loss": 3.1897,
      "step": 1992
    },
    {
      "epoch": 0.25801857785545523,
      "grad_norm": 2.7078753517169165,
      "learning_rate": 1.1573717599938492e-06,
      "loss": 2.6023,
      "step": 1993
    },
    {
      "epoch": 0.2581480402628087,
      "grad_norm": 4.831892795699603,
      "learning_rate": 1.157223722998608e-06,
      "loss": 3.3555,
      "step": 1994
    },
    {
      "epoch": 0.25827750267016214,
      "grad_norm": 3.2439017894440947,
      "learning_rate": 1.1570755772858837e-06,
      "loss": 3.2395,
      "step": 1995
    },
    {
      "epoch": 0.25827750267016214,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.676957607269287,
      "eval_runtime": 14.4971,
      "eval_samples_per_second": 3.035,
      "eval_steps_per_second": 0.207,
      "step": 1995
    },
    {
      "epoch": 0.2584069650775156,
      "grad_norm": 1.9346674939905308,
      "learning_rate": 1.1569273228859381e-06,
      "loss": 2.9404,
      "step": 1996
    },
    {
      "epoch": 0.2585364274848691,
      "grad_norm": 2.999529166710086,
      "learning_rate": 1.1567789598290552e-06,
      "loss": 3.043,
      "step": 1997
    },
    {
      "epoch": 0.25866588989222256,
      "grad_norm": 3.5759652702235187,
      "learning_rate": 1.156630488145541e-06,
      "loss": 3.0403,
      "step": 1998
    },
    {
      "epoch": 0.258795352299576,
      "grad_norm": 4.313763829851741,
      "learning_rate": 1.1564819078657238e-06,
      "loss": 3.0625,
      "step": 1999
    },
    {
      "epoch": 0.25892481470692946,
      "grad_norm": 2.9395046851351374,
      "learning_rate": 1.1563332190199544e-06,
      "loss": 2.6731,
      "step": 2000
    },
    {
      "epoch": 0.25892481470692946,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6755592823028564,
      "eval_runtime": 14.1564,
      "eval_samples_per_second": 3.108,
      "eval_steps_per_second": 0.212,
      "step": 2000
    },
    {
      "epoch": 0.2590542771142829,
      "grad_norm": 3.7414586362143853,
      "learning_rate": 1.1561844216386054e-06,
      "loss": 3.1995,
      "step": 2001
    },
    {
      "epoch": 0.2591837395216364,
      "grad_norm": 3.174259407155458,
      "learning_rate": 1.1560355157520717e-06,
      "loss": 3.1294,
      "step": 2002
    },
    {
      "epoch": 0.2593132019289899,
      "grad_norm": 3.281577783948224,
      "learning_rate": 1.1558865013907703e-06,
      "loss": 2.9639,
      "step": 2003
    },
    {
      "epoch": 0.25944266433634333,
      "grad_norm": 2.1271887396031803,
      "learning_rate": 1.1557373785851404e-06,
      "loss": 2.6738,
      "step": 2004
    },
    {
      "epoch": 0.2595721267436968,
      "grad_norm": 2.774438768866771,
      "learning_rate": 1.1555881473656433e-06,
      "loss": 3.1494,
      "step": 2005
    },
    {
      "epoch": 0.2595721267436968,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.681684970855713,
      "eval_runtime": 14.0066,
      "eval_samples_per_second": 3.141,
      "eval_steps_per_second": 0.214,
      "step": 2005
    },
    {
      "epoch": 0.25970158915105024,
      "grad_norm": 5.198117608582652,
      "learning_rate": 1.1554388077627625e-06,
      "loss": 3.2031,
      "step": 2006
    },
    {
      "epoch": 0.25983105155840375,
      "grad_norm": 3.78250264798494,
      "learning_rate": 1.1552893598070039e-06,
      "loss": 2.9292,
      "step": 2007
    },
    {
      "epoch": 0.2599605139657572,
      "grad_norm": 3.8546913392247575,
      "learning_rate": 1.155139803528895e-06,
      "loss": 3.1672,
      "step": 2008
    },
    {
      "epoch": 0.26008997637311065,
      "grad_norm": 4.48117237011425,
      "learning_rate": 1.1549901389589857e-06,
      "loss": 3.1758,
      "step": 2009
    },
    {
      "epoch": 0.2602194387804641,
      "grad_norm": 1.8305553827371683,
      "learning_rate": 1.1548403661278482e-06,
      "loss": 2.6951,
      "step": 2010
    },
    {
      "epoch": 0.2602194387804641,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.679643154144287,
      "eval_runtime": 13.7347,
      "eval_samples_per_second": 3.204,
      "eval_steps_per_second": 0.218,
      "step": 2010
    },
    {
      "epoch": 0.26034890118781756,
      "grad_norm": 3.109756075960622,
      "learning_rate": 1.1546904850660764e-06,
      "loss": 3.0293,
      "step": 2011
    },
    {
      "epoch": 0.26047836359517107,
      "grad_norm": 2.8726703270887923,
      "learning_rate": 1.154540495804287e-06,
      "loss": 3.1035,
      "step": 2012
    },
    {
      "epoch": 0.2606078260025245,
      "grad_norm": 2.9878880802222754,
      "learning_rate": 1.1543903983731178e-06,
      "loss": 2.7725,
      "step": 2013
    },
    {
      "epoch": 0.260737288409878,
      "grad_norm": 3.444341083421166,
      "learning_rate": 1.1542401928032295e-06,
      "loss": 2.6377,
      "step": 2014
    },
    {
      "epoch": 0.26086675081723143,
      "grad_norm": 1.7609766969107403,
      "learning_rate": 1.1540898791253047e-06,
      "loss": 2.6497,
      "step": 2015
    },
    {
      "epoch": 0.26086675081723143,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6788440942764282,
      "eval_runtime": 14.0198,
      "eval_samples_per_second": 3.138,
      "eval_steps_per_second": 0.214,
      "step": 2015
    },
    {
      "epoch": 0.26099621322458494,
      "grad_norm": 5.869701754719928,
      "learning_rate": 1.153939457370048e-06,
      "loss": 3.166,
      "step": 2016
    },
    {
      "epoch": 0.2611256756319384,
      "grad_norm": 2.2778536124349573,
      "learning_rate": 1.1537889275681857e-06,
      "loss": 2.5781,
      "step": 2017
    },
    {
      "epoch": 0.26125513803929185,
      "grad_norm": 2.468038446597042,
      "learning_rate": 1.153638289750467e-06,
      "loss": 2.6182,
      "step": 2018
    },
    {
      "epoch": 0.2613846004466453,
      "grad_norm": 5.395096398524735,
      "learning_rate": 1.1534875439476629e-06,
      "loss": 3.1719,
      "step": 2019
    },
    {
      "epoch": 0.26151406285399875,
      "grad_norm": 3.3207812998236697,
      "learning_rate": 1.1533366901905656e-06,
      "loss": 2.7502,
      "step": 2020
    },
    {
      "epoch": 0.26151406285399875,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6794655323028564,
      "eval_runtime": 13.6532,
      "eval_samples_per_second": 3.223,
      "eval_steps_per_second": 0.22,
      "step": 2020
    },
    {
      "epoch": 0.26164352526135226,
      "grad_norm": 2.427716489335077,
      "learning_rate": 1.1531857285099906e-06,
      "loss": 2.6158,
      "step": 2021
    },
    {
      "epoch": 0.2617729876687057,
      "grad_norm": 3.6106438357317048,
      "learning_rate": 1.1530346589367748e-06,
      "loss": 3.6465,
      "step": 2022
    },
    {
      "epoch": 0.26190245007605917,
      "grad_norm": 3.943559957928543,
      "learning_rate": 1.152883481501777e-06,
      "loss": 3.6538,
      "step": 2023
    },
    {
      "epoch": 0.2620319124834126,
      "grad_norm": 2.743156932662662,
      "learning_rate": 1.1527321962358785e-06,
      "loss": 2.7329,
      "step": 2024
    },
    {
      "epoch": 0.2621613748907661,
      "grad_norm": 2.6188952317713157,
      "learning_rate": 1.152580803169982e-06,
      "loss": 2.8013,
      "step": 2025
    },
    {
      "epoch": 0.2621613748907661,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6809303760528564,
      "eval_runtime": 14.3475,
      "eval_samples_per_second": 3.067,
      "eval_steps_per_second": 0.209,
      "step": 2025
    },
    {
      "epoch": 0.2622908372981196,
      "grad_norm": 2.7202565358168758,
      "learning_rate": 1.1524293023350133e-06,
      "loss": 2.9446,
      "step": 2026
    },
    {
      "epoch": 0.26242029970547304,
      "grad_norm": 2.3315902731412477,
      "learning_rate": 1.1522776937619189e-06,
      "loss": 2.9011,
      "step": 2027
    },
    {
      "epoch": 0.2625497621128265,
      "grad_norm": 2.3251635944723503,
      "learning_rate": 1.152125977481668e-06,
      "loss": 2.6204,
      "step": 2028
    },
    {
      "epoch": 0.26267922452017994,
      "grad_norm": 3.3325112007094364,
      "learning_rate": 1.151974153525252e-06,
      "loss": 2.8809,
      "step": 2029
    },
    {
      "epoch": 0.2628086869275334,
      "grad_norm": 2.199691784327611,
      "learning_rate": 1.1518222219236839e-06,
      "loss": 2.6538,
      "step": 2030
    },
    {
      "epoch": 0.2628086869275334,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6823952198028564,
      "eval_runtime": 14.006,
      "eval_samples_per_second": 3.142,
      "eval_steps_per_second": 0.214,
      "step": 2030
    },
    {
      "epoch": 0.2629381493348869,
      "grad_norm": 3.9473563499032025,
      "learning_rate": 1.1516701827079986e-06,
      "loss": 2.6489,
      "step": 2031
    },
    {
      "epoch": 0.26306761174224036,
      "grad_norm": 3.383719197891166,
      "learning_rate": 1.1515180359092534e-06,
      "loss": 3.0449,
      "step": 2032
    },
    {
      "epoch": 0.2631970741495938,
      "grad_norm": 2.9950880493165957,
      "learning_rate": 1.1513657815585276e-06,
      "loss": 2.9375,
      "step": 2033
    },
    {
      "epoch": 0.26332653655694727,
      "grad_norm": 8.46407490493878,
      "learning_rate": 1.1512134196869214e-06,
      "loss": 3.7515,
      "step": 2034
    },
    {
      "epoch": 0.2634559989643007,
      "grad_norm": 4.780627101920859,
      "learning_rate": 1.151060950325559e-06,
      "loss": 3.4456,
      "step": 2035
    },
    {
      "epoch": 0.2634559989643007,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6808860301971436,
      "eval_runtime": 13.8366,
      "eval_samples_per_second": 3.18,
      "eval_steps_per_second": 0.217,
      "step": 2035
    },
    {
      "epoch": 0.26358546137165423,
      "grad_norm": 2.79364686641326,
      "learning_rate": 1.1509083735055843e-06,
      "loss": 3.0391,
      "step": 2036
    },
    {
      "epoch": 0.2637149237790077,
      "grad_norm": 1.87933440252873,
      "learning_rate": 1.1507556892581648e-06,
      "loss": 2.8245,
      "step": 2037
    },
    {
      "epoch": 0.26384438618636114,
      "grad_norm": 1.8936710995485329,
      "learning_rate": 1.1506028976144892e-06,
      "loss": 2.6074,
      "step": 2038
    },
    {
      "epoch": 0.2639738485937146,
      "grad_norm": 1.4054084526805806,
      "learning_rate": 1.1504499986057682e-06,
      "loss": 2.4592,
      "step": 2039
    },
    {
      "epoch": 0.26410331100106804,
      "grad_norm": 2.4430917519941975,
      "learning_rate": 1.1502969922632345e-06,
      "loss": 2.929,
      "step": 2040
    },
    {
      "epoch": 0.26410331100106804,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6823064088821411,
      "eval_runtime": 14.0187,
      "eval_samples_per_second": 3.139,
      "eval_steps_per_second": 0.214,
      "step": 2040
    },
    {
      "epoch": 0.26423277340842155,
      "grad_norm": 7.223900577295312,
      "learning_rate": 1.1501438786181432e-06,
      "loss": 3.6663,
      "step": 2041
    },
    {
      "epoch": 0.264362235815775,
      "grad_norm": 2.318058792070975,
      "learning_rate": 1.1499906577017703e-06,
      "loss": 2.8945,
      "step": 2042
    },
    {
      "epoch": 0.26449169822312846,
      "grad_norm": 3.871859060919414,
      "learning_rate": 1.1498373295454142e-06,
      "loss": 3.21,
      "step": 2043
    },
    {
      "epoch": 0.2646211606304819,
      "grad_norm": 5.526379720800749,
      "learning_rate": 1.149683894180396e-06,
      "loss": 2.9705,
      "step": 2044
    },
    {
      "epoch": 0.26475062303783536,
      "grad_norm": 2.184094217105361,
      "learning_rate": 1.1495303516380575e-06,
      "loss": 2.7295,
      "step": 2045
    },
    {
      "epoch": 0.26475062303783536,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6807528734207153,
      "eval_runtime": 14.1793,
      "eval_samples_per_second": 3.103,
      "eval_steps_per_second": 0.212,
      "step": 2045
    },
    {
      "epoch": 0.2648800854451889,
      "grad_norm": 2.698760329009547,
      "learning_rate": 1.1493767019497628e-06,
      "loss": 2.5781,
      "step": 2046
    },
    {
      "epoch": 0.2650095478525423,
      "grad_norm": 2.7530255107060464,
      "learning_rate": 1.149222945146898e-06,
      "loss": 3.1382,
      "step": 2047
    },
    {
      "epoch": 0.2651390102598958,
      "grad_norm": 2.915233836831477,
      "learning_rate": 1.1490690812608714e-06,
      "loss": 2.7622,
      "step": 2048
    },
    {
      "epoch": 0.26526847266724923,
      "grad_norm": 2.3199207991923876,
      "learning_rate": 1.1489151103231125e-06,
      "loss": 2.675,
      "step": 2049
    },
    {
      "epoch": 0.2653979350746027,
      "grad_norm": 7.5563831009740365,
      "learning_rate": 1.148761032365073e-06,
      "loss": 3.752,
      "step": 2050
    },
    {
      "epoch": 0.2653979350746027,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6849032640457153,
      "eval_runtime": 14.0846,
      "eval_samples_per_second": 3.124,
      "eval_steps_per_second": 0.213,
      "step": 2050
    },
    {
      "epoch": 0.2655273974819562,
      "grad_norm": 2.4927119883817572,
      "learning_rate": 1.1486068474182264e-06,
      "loss": 2.9487,
      "step": 2051
    },
    {
      "epoch": 0.26565685988930965,
      "grad_norm": 7.5591324034845675,
      "learning_rate": 1.1484525555140682e-06,
      "loss": 3.5735,
      "step": 2052
    },
    {
      "epoch": 0.2657863222966631,
      "grad_norm": 2.0337293189836303,
      "learning_rate": 1.1482981566841156e-06,
      "loss": 2.9456,
      "step": 2053
    },
    {
      "epoch": 0.26591578470401656,
      "grad_norm": 2.389099062282753,
      "learning_rate": 1.148143650959908e-06,
      "loss": 3.1396,
      "step": 2054
    },
    {
      "epoch": 0.26604524711137,
      "grad_norm": 6.009646897002942,
      "learning_rate": 1.1479890383730059e-06,
      "loss": 3.4199,
      "step": 2055
    },
    {
      "epoch": 0.26604524711137,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6865234375,
      "eval_runtime": 15.7285,
      "eval_samples_per_second": 2.797,
      "eval_steps_per_second": 0.191,
      "step": 2055
    },
    {
      "epoch": 0.2661747095187235,
      "grad_norm": 2.126713921038673,
      "learning_rate": 1.147834318954992e-06,
      "loss": 2.551,
      "step": 2056
    },
    {
      "epoch": 0.26630417192607697,
      "grad_norm": 3.324447412588428,
      "learning_rate": 1.1476794927374713e-06,
      "loss": 2.8799,
      "step": 2057
    },
    {
      "epoch": 0.2664336343334304,
      "grad_norm": 2.3368801952968288,
      "learning_rate": 1.14752455975207e-06,
      "loss": 2.7803,
      "step": 2058
    },
    {
      "epoch": 0.2665630967407839,
      "grad_norm": 2.357680073144056,
      "learning_rate": 1.1473695200304362e-06,
      "loss": 2.6907,
      "step": 2059
    },
    {
      "epoch": 0.2666925591481374,
      "grad_norm": 3.046425147921824,
      "learning_rate": 1.1472143736042398e-06,
      "loss": 2.9746,
      "step": 2060
    },
    {
      "epoch": 0.2666925591481374,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6842595338821411,
      "eval_runtime": 14.2461,
      "eval_samples_per_second": 3.089,
      "eval_steps_per_second": 0.211,
      "step": 2060
    },
    {
      "epoch": 0.26682202155549084,
      "grad_norm": 4.973065091786923,
      "learning_rate": 1.1470591205051731e-06,
      "loss": 3.3354,
      "step": 2061
    },
    {
      "epoch": 0.2669514839628443,
      "grad_norm": 2.013444831983804,
      "learning_rate": 1.1469037607649494e-06,
      "loss": 2.6704,
      "step": 2062
    },
    {
      "epoch": 0.26708094637019775,
      "grad_norm": 5.622216107722647,
      "learning_rate": 1.1467482944153039e-06,
      "loss": 3.78,
      "step": 2063
    },
    {
      "epoch": 0.2672104087775512,
      "grad_norm": 1.798926339828797,
      "learning_rate": 1.1465927214879938e-06,
      "loss": 2.7153,
      "step": 2064
    },
    {
      "epoch": 0.2673398711849047,
      "grad_norm": 3.6711464441242017,
      "learning_rate": 1.1464370420147984e-06,
      "loss": 3.0742,
      "step": 2065
    },
    {
      "epoch": 0.2673398711849047,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.687056064605713,
      "eval_runtime": 14.2454,
      "eval_samples_per_second": 3.089,
      "eval_steps_per_second": 0.211,
      "step": 2065
    },
    {
      "epoch": 0.26746933359225816,
      "grad_norm": 5.410646675977281,
      "learning_rate": 1.146281256027518e-06,
      "loss": 3.6685,
      "step": 2066
    },
    {
      "epoch": 0.2675987959996116,
      "grad_norm": 3.9489489867754717,
      "learning_rate": 1.146125363557975e-06,
      "loss": 2.9485,
      "step": 2067
    },
    {
      "epoch": 0.26772825840696507,
      "grad_norm": 5.146815929244116,
      "learning_rate": 1.145969364638014e-06,
      "loss": 3.1311,
      "step": 2068
    },
    {
      "epoch": 0.2678577208143185,
      "grad_norm": 6.755278783157041,
      "learning_rate": 1.1458132592995005e-06,
      "loss": 3.9727,
      "step": 2069
    },
    {
      "epoch": 0.26798718322167203,
      "grad_norm": 5.54033753808382,
      "learning_rate": 1.1456570475743225e-06,
      "loss": 3.1401,
      "step": 2070
    },
    {
      "epoch": 0.26798718322167203,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6835049390792847,
      "eval_runtime": 14.487,
      "eval_samples_per_second": 3.037,
      "eval_steps_per_second": 0.207,
      "step": 2070
    },
    {
      "epoch": 0.2681166456290255,
      "grad_norm": 4.096552958120691,
      "learning_rate": 1.1455007294943895e-06,
      "loss": 3.2363,
      "step": 2071
    },
    {
      "epoch": 0.26824610803637894,
      "grad_norm": 2.5353750762707965,
      "learning_rate": 1.1453443050916322e-06,
      "loss": 2.9187,
      "step": 2072
    },
    {
      "epoch": 0.2683755704437324,
      "grad_norm": 2.4643309441062984,
      "learning_rate": 1.1451877743980036e-06,
      "loss": 2.8909,
      "step": 2073
    },
    {
      "epoch": 0.26850503285108585,
      "grad_norm": 1.348337058733433,
      "learning_rate": 1.1450311374454786e-06,
      "loss": 2.4468,
      "step": 2074
    },
    {
      "epoch": 0.26863449525843935,
      "grad_norm": 2.512610412236078,
      "learning_rate": 1.1448743942660532e-06,
      "loss": 2.7932,
      "step": 2075
    },
    {
      "epoch": 0.26863449525843935,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6842595338821411,
      "eval_runtime": 15.5582,
      "eval_samples_per_second": 2.828,
      "eval_steps_per_second": 0.193,
      "step": 2075
    },
    {
      "epoch": 0.2687639576657928,
      "grad_norm": 3.4194935079848956,
      "learning_rate": 1.1447175448917452e-06,
      "loss": 2.9795,
      "step": 2076
    },
    {
      "epoch": 0.26889342007314626,
      "grad_norm": 2.2507691559282668,
      "learning_rate": 1.1445605893545946e-06,
      "loss": 3.0302,
      "step": 2077
    },
    {
      "epoch": 0.2690228824804997,
      "grad_norm": 5.145316035641122,
      "learning_rate": 1.1444035276866625e-06,
      "loss": 3.2922,
      "step": 2078
    },
    {
      "epoch": 0.26915234488785317,
      "grad_norm": 2.8097246510674085,
      "learning_rate": 1.144246359920032e-06,
      "loss": 3.0798,
      "step": 2079
    },
    {
      "epoch": 0.2692818072952067,
      "grad_norm": 2.6726345566288794,
      "learning_rate": 1.1440890860868082e-06,
      "loss": 3.1318,
      "step": 2080
    },
    {
      "epoch": 0.2692818072952067,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6905406713485718,
      "eval_runtime": 15.2411,
      "eval_samples_per_second": 2.887,
      "eval_steps_per_second": 0.197,
      "step": 2080
    },
    {
      "epoch": 0.26941126970256013,
      "grad_norm": 5.951832984679558,
      "learning_rate": 1.1439317062191168e-06,
      "loss": 3.1069,
      "step": 2081
    },
    {
      "epoch": 0.2695407321099136,
      "grad_norm": 3.2621629390696714,
      "learning_rate": 1.1437742203491062e-06,
      "loss": 3.1311,
      "step": 2082
    },
    {
      "epoch": 0.26967019451726704,
      "grad_norm": 2.6258824367727622,
      "learning_rate": 1.1436166285089458e-06,
      "loss": 3.1406,
      "step": 2083
    },
    {
      "epoch": 0.2697996569246205,
      "grad_norm": 2.4642538273932812,
      "learning_rate": 1.1434589307308275e-06,
      "loss": 2.8,
      "step": 2084
    },
    {
      "epoch": 0.269929119331974,
      "grad_norm": 5.099022838437128,
      "learning_rate": 1.1433011270469638e-06,
      "loss": 3.3835,
      "step": 2085
    },
    {
      "epoch": 0.269929119331974,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6904518604278564,
      "eval_runtime": 14.9654,
      "eval_samples_per_second": 2.94,
      "eval_steps_per_second": 0.2,
      "step": 2085
    },
    {
      "epoch": 0.27005858173932745,
      "grad_norm": 2.6385860321112222,
      "learning_rate": 1.1431432174895894e-06,
      "loss": 2.8601,
      "step": 2086
    },
    {
      "epoch": 0.2701880441466809,
      "grad_norm": 6.821887298439188,
      "learning_rate": 1.1429852020909603e-06,
      "loss": 3.4854,
      "step": 2087
    },
    {
      "epoch": 0.27031750655403436,
      "grad_norm": 2.7629744843909694,
      "learning_rate": 1.142827080883355e-06,
      "loss": 2.7114,
      "step": 2088
    },
    {
      "epoch": 0.2704469689613878,
      "grad_norm": 4.467498101925651,
      "learning_rate": 1.1426688538990724e-06,
      "loss": 2.9343,
      "step": 2089
    },
    {
      "epoch": 0.2705764313687413,
      "grad_norm": 3.044496230423841,
      "learning_rate": 1.1425105211704338e-06,
      "loss": 3.2568,
      "step": 2090
    },
    {
      "epoch": 0.2705764313687413,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6920276880264282,
      "eval_runtime": 15.799,
      "eval_samples_per_second": 2.785,
      "eval_steps_per_second": 0.19,
      "step": 2090
    },
    {
      "epoch": 0.2707058937760948,
      "grad_norm": 1.7772418360489868,
      "learning_rate": 1.1423520827297814e-06,
      "loss": 2.6912,
      "step": 2091
    },
    {
      "epoch": 0.27083535618344823,
      "grad_norm": 3.4981336783934514,
      "learning_rate": 1.1421935386094803e-06,
      "loss": 3.1421,
      "step": 2092
    },
    {
      "epoch": 0.2709648185908017,
      "grad_norm": 2.348128872979748,
      "learning_rate": 1.1420348888419158e-06,
      "loss": 2.6309,
      "step": 2093
    },
    {
      "epoch": 0.27109428099815513,
      "grad_norm": 1.5857791684155143,
      "learning_rate": 1.1418761334594954e-06,
      "loss": 2.5281,
      "step": 2094
    },
    {
      "epoch": 0.27122374340550864,
      "grad_norm": 2.908288025149171,
      "learning_rate": 1.1417172724946482e-06,
      "loss": 3.4331,
      "step": 2095
    },
    {
      "epoch": 0.27122374340550864,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6918723583221436,
      "eval_runtime": 15.2125,
      "eval_samples_per_second": 2.892,
      "eval_steps_per_second": 0.197,
      "step": 2095
    },
    {
      "epoch": 0.2713532058128621,
      "grad_norm": 8.048230156977537,
      "learning_rate": 1.1415583059798247e-06,
      "loss": 3.7832,
      "step": 2096
    },
    {
      "epoch": 0.27148266822021555,
      "grad_norm": 4.618960932081442,
      "learning_rate": 1.1413992339474968e-06,
      "loss": 3.5208,
      "step": 2097
    },
    {
      "epoch": 0.271612130627569,
      "grad_norm": 5.446955840816096,
      "learning_rate": 1.1412400564301587e-06,
      "loss": 3.4502,
      "step": 2098
    },
    {
      "epoch": 0.27174159303492246,
      "grad_norm": 2.3579238476305817,
      "learning_rate": 1.1410807734603251e-06,
      "loss": 2.8325,
      "step": 2099
    },
    {
      "epoch": 0.27187105544227597,
      "grad_norm": 4.365697054676799,
      "learning_rate": 1.1409213850705334e-06,
      "loss": 3.1313,
      "step": 2100
    },
    {
      "epoch": 0.27187105544227597,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6954900026321411,
      "eval_runtime": 14.5369,
      "eval_samples_per_second": 3.027,
      "eval_steps_per_second": 0.206,
      "step": 2100
    },
    {
      "epoch": 0.2720005178496294,
      "grad_norm": 2.371784006098729,
      "learning_rate": 1.1407618912933414e-06,
      "loss": 2.8833,
      "step": 2101
    },
    {
      "epoch": 0.2721299802569829,
      "grad_norm": 4.746801470373952,
      "learning_rate": 1.140602292161329e-06,
      "loss": 3.0278,
      "step": 2102
    },
    {
      "epoch": 0.2722594426643363,
      "grad_norm": 2.5329249323059995,
      "learning_rate": 1.1404425877070977e-06,
      "loss": 2.7031,
      "step": 2103
    },
    {
      "epoch": 0.27238890507168984,
      "grad_norm": 2.871788693388005,
      "learning_rate": 1.1402827779632702e-06,
      "loss": 2.7466,
      "step": 2104
    },
    {
      "epoch": 0.2725183674790433,
      "grad_norm": 3.1461985929759897,
      "learning_rate": 1.1401228629624912e-06,
      "loss": 2.7627,
      "step": 2105
    },
    {
      "epoch": 0.2725183674790433,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.697509765625,
      "eval_runtime": 14.5422,
      "eval_samples_per_second": 3.026,
      "eval_steps_per_second": 0.206,
      "step": 2105
    },
    {
      "epoch": 0.27264782988639674,
      "grad_norm": 2.8704105394729185,
      "learning_rate": 1.1399628427374262e-06,
      "loss": 2.8513,
      "step": 2106
    },
    {
      "epoch": 0.2727772922937502,
      "grad_norm": 3.501116124828323,
      "learning_rate": 1.139802717320763e-06,
      "loss": 2.8237,
      "step": 2107
    },
    {
      "epoch": 0.27290675470110365,
      "grad_norm": 7.3612594472307045,
      "learning_rate": 1.1396424867452099e-06,
      "loss": 3.647,
      "step": 2108
    },
    {
      "epoch": 0.27303621710845716,
      "grad_norm": 3.401919848713317,
      "learning_rate": 1.139482151043498e-06,
      "loss": 2.9775,
      "step": 2109
    },
    {
      "epoch": 0.2731656795158106,
      "grad_norm": 3.0254158567668004,
      "learning_rate": 1.139321710248378e-06,
      "loss": 2.8589,
      "step": 2110
    },
    {
      "epoch": 0.2731656795158106,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6967995166778564,
      "eval_runtime": 15.7066,
      "eval_samples_per_second": 2.801,
      "eval_steps_per_second": 0.191,
      "step": 2110
    },
    {
      "epoch": 0.27329514192316406,
      "grad_norm": 2.956296196905673,
      "learning_rate": 1.1391611643926243e-06,
      "loss": 2.8242,
      "step": 2111
    },
    {
      "epoch": 0.2734246043305175,
      "grad_norm": 4.845459647337277,
      "learning_rate": 1.1390005135090313e-06,
      "loss": 3.3892,
      "step": 2112
    },
    {
      "epoch": 0.27355406673787097,
      "grad_norm": 2.6719008999607325,
      "learning_rate": 1.1388397576304148e-06,
      "loss": 3.292,
      "step": 2113
    },
    {
      "epoch": 0.2736835291452245,
      "grad_norm": 6.575587963808951,
      "learning_rate": 1.138678896789613e-06,
      "loss": 3.0684,
      "step": 2114
    },
    {
      "epoch": 0.27381299155257793,
      "grad_norm": 2.112497654626105,
      "learning_rate": 1.1385179310194844e-06,
      "loss": 2.6973,
      "step": 2115
    },
    {
      "epoch": 0.27381299155257793,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6988192796707153,
      "eval_runtime": 14.3833,
      "eval_samples_per_second": 3.059,
      "eval_steps_per_second": 0.209,
      "step": 2115
    },
    {
      "epoch": 0.2739424539599314,
      "grad_norm": 4.61997539323117,
      "learning_rate": 1.1383568603529101e-06,
      "loss": 2.9185,
      "step": 2116
    },
    {
      "epoch": 0.27407191636728484,
      "grad_norm": 5.3024146243248245,
      "learning_rate": 1.138195684822792e-06,
      "loss": 3.2292,
      "step": 2117
    },
    {
      "epoch": 0.2742013787746383,
      "grad_norm": 2.9069223441234873,
      "learning_rate": 1.1380344044620529e-06,
      "loss": 2.7588,
      "step": 2118
    },
    {
      "epoch": 0.2743308411819918,
      "grad_norm": 2.107055960389527,
      "learning_rate": 1.137873019303638e-06,
      "loss": 2.6982,
      "step": 2119
    },
    {
      "epoch": 0.27446030358934526,
      "grad_norm": 2.2558812451416097,
      "learning_rate": 1.1377115293805136e-06,
      "loss": 2.8921,
      "step": 2120
    },
    {
      "epoch": 0.27446030358934526,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7009055614471436,
      "eval_runtime": 13.3486,
      "eval_samples_per_second": 3.296,
      "eval_steps_per_second": 0.225,
      "step": 2120
    },
    {
      "epoch": 0.2745897659966987,
      "grad_norm": 5.964490517844797,
      "learning_rate": 1.137549934725667e-06,
      "loss": 3.783,
      "step": 2121
    },
    {
      "epoch": 0.27471922840405216,
      "grad_norm": 5.298975138485487,
      "learning_rate": 1.1373882353721074e-06,
      "loss": 3.0503,
      "step": 2122
    },
    {
      "epoch": 0.2748486908114056,
      "grad_norm": 1.7339039669352063,
      "learning_rate": 1.1372264313528652e-06,
      "loss": 2.6182,
      "step": 2123
    },
    {
      "epoch": 0.2749781532187591,
      "grad_norm": 2.515719379908353,
      "learning_rate": 1.137064522700992e-06,
      "loss": 3.1836,
      "step": 2124
    },
    {
      "epoch": 0.2751076156261126,
      "grad_norm": 3.359680611777102,
      "learning_rate": 1.1369025094495606e-06,
      "loss": 3.2922,
      "step": 2125
    },
    {
      "epoch": 0.2751076156261126,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7035911083221436,
      "eval_runtime": 14.7459,
      "eval_samples_per_second": 2.984,
      "eval_steps_per_second": 0.203,
      "step": 2125
    },
    {
      "epoch": 0.27523707803346603,
      "grad_norm": 2.5993518130739846,
      "learning_rate": 1.1367403916316663e-06,
      "loss": 3.1245,
      "step": 2126
    },
    {
      "epoch": 0.2753665404408195,
      "grad_norm": 2.409255174076651,
      "learning_rate": 1.1365781692804242e-06,
      "loss": 2.8301,
      "step": 2127
    },
    {
      "epoch": 0.27549600284817294,
      "grad_norm": 2.5872875135177296,
      "learning_rate": 1.1364158424289721e-06,
      "loss": 2.7595,
      "step": 2128
    },
    {
      "epoch": 0.27562546525552645,
      "grad_norm": 3.7293917500577956,
      "learning_rate": 1.136253411110468e-06,
      "loss": 2.7498,
      "step": 2129
    },
    {
      "epoch": 0.2757549276628799,
      "grad_norm": 2.1062540891706947,
      "learning_rate": 1.1360908753580921e-06,
      "loss": 2.8459,
      "step": 2130
    },
    {
      "epoch": 0.2757549276628799,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.702925205230713,
      "eval_runtime": 15.6268,
      "eval_samples_per_second": 2.816,
      "eval_steps_per_second": 0.192,
      "step": 2130
    },
    {
      "epoch": 0.27588439007023335,
      "grad_norm": 3.2745964864625994,
      "learning_rate": 1.1359282352050458e-06,
      "loss": 3.1592,
      "step": 2131
    },
    {
      "epoch": 0.2760138524775868,
      "grad_norm": 8.782837880746253,
      "learning_rate": 1.135765490684551e-06,
      "loss": 3.9993,
      "step": 2132
    },
    {
      "epoch": 0.27614331488494026,
      "grad_norm": 2.3521076664452263,
      "learning_rate": 1.1356026418298525e-06,
      "loss": 2.7415,
      "step": 2133
    },
    {
      "epoch": 0.27627277729229377,
      "grad_norm": 4.203858141927902,
      "learning_rate": 1.1354396886742144e-06,
      "loss": 3.0024,
      "step": 2134
    },
    {
      "epoch": 0.2764022396996472,
      "grad_norm": 8.90162122217747,
      "learning_rate": 1.1352766312509242e-06,
      "loss": 4.02,
      "step": 2135
    },
    {
      "epoch": 0.2764022396996472,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.71142578125,
      "eval_runtime": 16.0101,
      "eval_samples_per_second": 2.748,
      "eval_steps_per_second": 0.187,
      "step": 2135
    },
    {
      "epoch": 0.2765317021070007,
      "grad_norm": 2.3512780688100317,
      "learning_rate": 1.135113469593289e-06,
      "loss": 2.8206,
      "step": 2136
    },
    {
      "epoch": 0.27666116451435413,
      "grad_norm": 3.009829274006804,
      "learning_rate": 1.134950203734638e-06,
      "loss": 2.6372,
      "step": 2137
    },
    {
      "epoch": 0.2767906269217076,
      "grad_norm": 3.198183367749438,
      "learning_rate": 1.134786833708322e-06,
      "loss": 2.8228,
      "step": 2138
    },
    {
      "epoch": 0.2769200893290611,
      "grad_norm": 2.9172347484155075,
      "learning_rate": 1.134623359547712e-06,
      "loss": 2.9282,
      "step": 2139
    },
    {
      "epoch": 0.27704955173641455,
      "grad_norm": 4.423684413944046,
      "learning_rate": 1.1344597812862016e-06,
      "loss": 3.2554,
      "step": 2140
    },
    {
      "epoch": 0.27704955173641455,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7084072828292847,
      "eval_runtime": 14.5608,
      "eval_samples_per_second": 3.022,
      "eval_steps_per_second": 0.206,
      "step": 2140
    },
    {
      "epoch": 0.277179014143768,
      "grad_norm": 3.7885066030285275,
      "learning_rate": 1.1342960989572041e-06,
      "loss": 3.0969,
      "step": 2141
    },
    {
      "epoch": 0.27730847655112145,
      "grad_norm": 2.239514932877214,
      "learning_rate": 1.1341323125941559e-06,
      "loss": 2.9482,
      "step": 2142
    },
    {
      "epoch": 0.2774379389584749,
      "grad_norm": 4.071755556531883,
      "learning_rate": 1.1339684222305127e-06,
      "loss": 3.0747,
      "step": 2143
    },
    {
      "epoch": 0.2775674013658284,
      "grad_norm": 3.924183068294491,
      "learning_rate": 1.1338044278997535e-06,
      "loss": 3.0068,
      "step": 2144
    },
    {
      "epoch": 0.27769686377318187,
      "grad_norm": 2.4623776182123587,
      "learning_rate": 1.1336403296353765e-06,
      "loss": 2.6719,
      "step": 2145
    },
    {
      "epoch": 0.27769686377318187,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7112038135528564,
      "eval_runtime": 14.625,
      "eval_samples_per_second": 3.009,
      "eval_steps_per_second": 0.205,
      "step": 2145
    },
    {
      "epoch": 0.2778263261805353,
      "grad_norm": 5.22692754769894,
      "learning_rate": 1.1334761274709026e-06,
      "loss": 4.0322,
      "step": 2146
    },
    {
      "epoch": 0.2779557885878888,
      "grad_norm": 1.9911624215537151,
      "learning_rate": 1.1333118214398733e-06,
      "loss": 2.6882,
      "step": 2147
    },
    {
      "epoch": 0.2780852509952423,
      "grad_norm": 2.030006439493325,
      "learning_rate": 1.1331474115758517e-06,
      "loss": 2.6562,
      "step": 2148
    },
    {
      "epoch": 0.27821471340259574,
      "grad_norm": 2.71867645698034,
      "learning_rate": 1.1329828979124212e-06,
      "loss": 2.7881,
      "step": 2149
    },
    {
      "epoch": 0.2783441758099492,
      "grad_norm": 3.6141895546256597,
      "learning_rate": 1.1328182804831878e-06,
      "loss": 2.8708,
      "step": 2150
    },
    {
      "epoch": 0.2783441758099492,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7105823755264282,
      "eval_runtime": 14.4718,
      "eval_samples_per_second": 3.04,
      "eval_steps_per_second": 0.207,
      "step": 2150
    },
    {
      "epoch": 0.27847363821730264,
      "grad_norm": 2.63689874732753,
      "learning_rate": 1.1326535593217774e-06,
      "loss": 3.0059,
      "step": 2151
    },
    {
      "epoch": 0.2786031006246561,
      "grad_norm": 1.7058516968153965,
      "learning_rate": 1.132488734461838e-06,
      "loss": 2.3843,
      "step": 2152
    },
    {
      "epoch": 0.2787325630320096,
      "grad_norm": 2.893541070293363,
      "learning_rate": 1.132323805937038e-06,
      "loss": 2.8291,
      "step": 2153
    },
    {
      "epoch": 0.27886202543936306,
      "grad_norm": 5.965881199299705,
      "learning_rate": 1.132158773781068e-06,
      "loss": 3.4507,
      "step": 2154
    },
    {
      "epoch": 0.2789914878467165,
      "grad_norm": 3.6014909656993015,
      "learning_rate": 1.1319936380276386e-06,
      "loss": 3.0967,
      "step": 2155
    },
    {
      "epoch": 0.2789914878467165,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7100497484207153,
      "eval_runtime": 18.8346,
      "eval_samples_per_second": 2.336,
      "eval_steps_per_second": 0.159,
      "step": 2155
    },
    {
      "epoch": 0.27912095025406997,
      "grad_norm": 2.1069149059004206,
      "learning_rate": 1.1318283987104826e-06,
      "loss": 2.498,
      "step": 2156
    },
    {
      "epoch": 0.2792504126614234,
      "grad_norm": 4.101797173354059,
      "learning_rate": 1.131663055863353e-06,
      "loss": 3.4045,
      "step": 2157
    },
    {
      "epoch": 0.27937987506877693,
      "grad_norm": 2.113263306558502,
      "learning_rate": 1.1314976095200244e-06,
      "loss": 2.7646,
      "step": 2158
    },
    {
      "epoch": 0.2795093374761304,
      "grad_norm": 2.9764413417627655,
      "learning_rate": 1.1313320597142932e-06,
      "loss": 2.7197,
      "step": 2159
    },
    {
      "epoch": 0.27963879988348384,
      "grad_norm": 6.3766328601527045,
      "learning_rate": 1.1311664064799756e-06,
      "loss": 3.187,
      "step": 2160
    },
    {
      "epoch": 0.27963879988348384,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7097834348678589,
      "eval_runtime": 14.5663,
      "eval_samples_per_second": 3.021,
      "eval_steps_per_second": 0.206,
      "step": 2160
    },
    {
      "epoch": 0.2797682622908373,
      "grad_norm": 2.0292187713159056,
      "learning_rate": 1.1310006498509102e-06,
      "loss": 2.717,
      "step": 2161
    },
    {
      "epoch": 0.27989772469819074,
      "grad_norm": 2.5814056537863985,
      "learning_rate": 1.130834789860956e-06,
      "loss": 2.8288,
      "step": 2162
    },
    {
      "epoch": 0.28002718710554425,
      "grad_norm": 3.984005890631887,
      "learning_rate": 1.130668826543993e-06,
      "loss": 3.1814,
      "step": 2163
    },
    {
      "epoch": 0.2801566495128977,
      "grad_norm": 2.9525404229419685,
      "learning_rate": 1.1305027599339228e-06,
      "loss": 2.8315,
      "step": 2164
    },
    {
      "epoch": 0.28028611192025116,
      "grad_norm": 3.0591320774934507,
      "learning_rate": 1.1303365900646675e-06,
      "loss": 2.7375,
      "step": 2165
    },
    {
      "epoch": 0.28028611192025116,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7112038135528564,
      "eval_runtime": 14.2994,
      "eval_samples_per_second": 3.077,
      "eval_steps_per_second": 0.21,
      "step": 2165
    },
    {
      "epoch": 0.2804155743276046,
      "grad_norm": 1.5957569858895388,
      "learning_rate": 1.1301703169701713e-06,
      "loss": 2.6003,
      "step": 2166
    },
    {
      "epoch": 0.28054503673495806,
      "grad_norm": 1.9643281898373082,
      "learning_rate": 1.1300039406843985e-06,
      "loss": 2.8228,
      "step": 2167
    },
    {
      "epoch": 0.2806744991423116,
      "grad_norm": 3.6598082645507644,
      "learning_rate": 1.129837461241335e-06,
      "loss": 2.8616,
      "step": 2168
    },
    {
      "epoch": 0.280803961549665,
      "grad_norm": 4.983280124518397,
      "learning_rate": 1.1296708786749872e-06,
      "loss": 3.1206,
      "step": 2169
    },
    {
      "epoch": 0.2809334239570185,
      "grad_norm": 4.4437933766836615,
      "learning_rate": 1.129504193019383e-06,
      "loss": 3.0837,
      "step": 2170
    },
    {
      "epoch": 0.2809334239570185,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7062321901321411,
      "eval_runtime": 14.4183,
      "eval_samples_per_second": 3.052,
      "eval_steps_per_second": 0.208,
      "step": 2170
    },
    {
      "epoch": 0.28106288636437193,
      "grad_norm": 5.137792537324083,
      "learning_rate": 1.129337404308572e-06,
      "loss": 2.9954,
      "step": 2171
    },
    {
      "epoch": 0.2811923487717254,
      "grad_norm": 3.3956087855775547,
      "learning_rate": 1.1291705125766235e-06,
      "loss": 2.9438,
      "step": 2172
    },
    {
      "epoch": 0.2813218111790789,
      "grad_norm": 1.766554721470038,
      "learning_rate": 1.1290035178576288e-06,
      "loss": 2.8037,
      "step": 2173
    },
    {
      "epoch": 0.28145127358643235,
      "grad_norm": 2.044865806947404,
      "learning_rate": 1.1288364201856998e-06,
      "loss": 3.0144,
      "step": 2174
    },
    {
      "epoch": 0.2815807359937858,
      "grad_norm": 4.498626168572207,
      "learning_rate": 1.1286692195949696e-06,
      "loss": 3.3193,
      "step": 2175
    },
    {
      "epoch": 0.2815807359937858,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7090510129928589,
      "eval_runtime": 14.6364,
      "eval_samples_per_second": 3.006,
      "eval_steps_per_second": 0.205,
      "step": 2175
    },
    {
      "epoch": 0.28171019840113926,
      "grad_norm": 2.144516666917743,
      "learning_rate": 1.1285019161195926e-06,
      "loss": 2.7476,
      "step": 2176
    },
    {
      "epoch": 0.2818396608084927,
      "grad_norm": 8.829021740744638,
      "learning_rate": 1.1283345097937438e-06,
      "loss": 3.3428,
      "step": 2177
    },
    {
      "epoch": 0.2819691232158462,
      "grad_norm": 1.683585915021965,
      "learning_rate": 1.128167000651619e-06,
      "loss": 2.6011,
      "step": 2178
    },
    {
      "epoch": 0.28209858562319967,
      "grad_norm": 1.7929468918924358,
      "learning_rate": 1.1279993887274356e-06,
      "loss": 2.6406,
      "step": 2179
    },
    {
      "epoch": 0.2822280480305531,
      "grad_norm": 2.1813791094384816,
      "learning_rate": 1.1278316740554319e-06,
      "loss": 2.9067,
      "step": 2180
    },
    {
      "epoch": 0.2822280480305531,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7127574682235718,
      "eval_runtime": 14.0195,
      "eval_samples_per_second": 3.138,
      "eval_steps_per_second": 0.214,
      "step": 2180
    },
    {
      "epoch": 0.2823575104379066,
      "grad_norm": 4.690397057184647,
      "learning_rate": 1.1276638566698667e-06,
      "loss": 3.2859,
      "step": 2181
    },
    {
      "epoch": 0.28248697284526003,
      "grad_norm": 5.077977473556037,
      "learning_rate": 1.1274959366050203e-06,
      "loss": 3.4092,
      "step": 2182
    },
    {
      "epoch": 0.28261643525261354,
      "grad_norm": 4.632403734886668,
      "learning_rate": 1.1273279138951937e-06,
      "loss": 2.9756,
      "step": 2183
    },
    {
      "epoch": 0.282745897659967,
      "grad_norm": 2.325233264494713,
      "learning_rate": 1.127159788574709e-06,
      "loss": 2.9412,
      "step": 2184
    },
    {
      "epoch": 0.28287536006732045,
      "grad_norm": 5.290140977108026,
      "learning_rate": 1.1269915606779091e-06,
      "loss": 3.4653,
      "step": 2185
    },
    {
      "epoch": 0.28287536006732045,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.708984375,
      "eval_runtime": 14.488,
      "eval_samples_per_second": 3.037,
      "eval_steps_per_second": 0.207,
      "step": 2185
    },
    {
      "epoch": 0.2830048224746739,
      "grad_norm": 1.7586007089312548,
      "learning_rate": 1.126823230239158e-06,
      "loss": 2.5349,
      "step": 2186
    },
    {
      "epoch": 0.28313428488202735,
      "grad_norm": 2.1762952956245547,
      "learning_rate": 1.1266547972928407e-06,
      "loss": 2.7998,
      "step": 2187
    },
    {
      "epoch": 0.28326374728938086,
      "grad_norm": 1.8217932197025986,
      "learning_rate": 1.126486261873363e-06,
      "loss": 2.9033,
      "step": 2188
    },
    {
      "epoch": 0.2833932096967343,
      "grad_norm": 2.1877432405932264,
      "learning_rate": 1.1263176240151516e-06,
      "loss": 2.54,
      "step": 2189
    },
    {
      "epoch": 0.28352267210408777,
      "grad_norm": 2.557256721108407,
      "learning_rate": 1.1261488837526543e-06,
      "loss": 2.688,
      "step": 2190
    },
    {
      "epoch": 0.28352267210408777,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7084072828292847,
      "eval_runtime": 14.6506,
      "eval_samples_per_second": 3.003,
      "eval_steps_per_second": 0.205,
      "step": 2190
    },
    {
      "epoch": 0.2836521345114412,
      "grad_norm": 2.0520267301152955,
      "learning_rate": 1.1259800411203395e-06,
      "loss": 2.6309,
      "step": 2191
    },
    {
      "epoch": 0.28378159691879473,
      "grad_norm": 3.058716618496961,
      "learning_rate": 1.1258110961526973e-06,
      "loss": 2.6948,
      "step": 2192
    },
    {
      "epoch": 0.2839110593261482,
      "grad_norm": 3.378395717359709,
      "learning_rate": 1.1256420488842373e-06,
      "loss": 3.1841,
      "step": 2193
    },
    {
      "epoch": 0.28404052173350164,
      "grad_norm": 2.875080689847274,
      "learning_rate": 1.1254728993494914e-06,
      "loss": 2.925,
      "step": 2194
    },
    {
      "epoch": 0.2841699841408551,
      "grad_norm": 2.840536060028901,
      "learning_rate": 1.1253036475830116e-06,
      "loss": 2.8147,
      "step": 2195
    },
    {
      "epoch": 0.2841699841408551,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7090065479278564,
      "eval_runtime": 13.7959,
      "eval_samples_per_second": 3.189,
      "eval_steps_per_second": 0.217,
      "step": 2195
    },
    {
      "epoch": 0.28429944654820855,
      "grad_norm": 2.804921306109501,
      "learning_rate": 1.1251342936193713e-06,
      "loss": 2.8738,
      "step": 2196
    },
    {
      "epoch": 0.28442890895556205,
      "grad_norm": 2.3753824637981005,
      "learning_rate": 1.124964837493164e-06,
      "loss": 2.8962,
      "step": 2197
    },
    {
      "epoch": 0.2845583713629155,
      "grad_norm": 2.6462816839800647,
      "learning_rate": 1.124795279239005e-06,
      "loss": 2.8684,
      "step": 2198
    },
    {
      "epoch": 0.28468783377026896,
      "grad_norm": 3.0179004137056396,
      "learning_rate": 1.12462561889153e-06,
      "loss": 3.0945,
      "step": 2199
    },
    {
      "epoch": 0.2848172961776224,
      "grad_norm": 2.4404753685891007,
      "learning_rate": 1.1244558564853951e-06,
      "loss": 2.5583,
      "step": 2200
    },
    {
      "epoch": 0.2848172961776224,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7096946239471436,
      "eval_runtime": 14.7399,
      "eval_samples_per_second": 2.985,
      "eval_steps_per_second": 0.204,
      "step": 2200
    },
    {
      "epoch": 0.28494675858497587,
      "grad_norm": 1.3181340072316032,
      "learning_rate": 1.1242859920552782e-06,
      "loss": 2.3582,
      "step": 2201
    },
    {
      "epoch": 0.2850762209923294,
      "grad_norm": 2.572755352280697,
      "learning_rate": 1.1241160256358771e-06,
      "loss": 2.626,
      "step": 2202
    },
    {
      "epoch": 0.28520568339968283,
      "grad_norm": 2.1431148659891854,
      "learning_rate": 1.1239459572619114e-06,
      "loss": 2.9226,
      "step": 2203
    },
    {
      "epoch": 0.2853351458070363,
      "grad_norm": 1.755824724041719,
      "learning_rate": 1.1237757869681207e-06,
      "loss": 2.8125,
      "step": 2204
    },
    {
      "epoch": 0.28546460821438974,
      "grad_norm": 2.3796147900861357,
      "learning_rate": 1.1236055147892655e-06,
      "loss": 2.9813,
      "step": 2205
    },
    {
      "epoch": 0.28546460821438974,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7118252515792847,
      "eval_runtime": 13.9685,
      "eval_samples_per_second": 3.15,
      "eval_steps_per_second": 0.215,
      "step": 2205
    },
    {
      "epoch": 0.2855940706217432,
      "grad_norm": 2.6189704283882884,
      "learning_rate": 1.123435140760128e-06,
      "loss": 3.0723,
      "step": 2206
    },
    {
      "epoch": 0.2857235330290967,
      "grad_norm": 2.1538075409259925,
      "learning_rate": 1.12326466491551e-06,
      "loss": 2.8909,
      "step": 2207
    },
    {
      "epoch": 0.28585299543645015,
      "grad_norm": 5.4951468246839195,
      "learning_rate": 1.123094087290235e-06,
      "loss": 3.0295,
      "step": 2208
    },
    {
      "epoch": 0.2859824578438036,
      "grad_norm": 3.380280735939302,
      "learning_rate": 1.1229234079191465e-06,
      "loss": 2.8865,
      "step": 2209
    },
    {
      "epoch": 0.28611192025115706,
      "grad_norm": 4.557573953496405,
      "learning_rate": 1.1227526268371097e-06,
      "loss": 2.7681,
      "step": 2210
    },
    {
      "epoch": 0.28611192025115706,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.705322265625,
      "eval_runtime": 14.0208,
      "eval_samples_per_second": 3.138,
      "eval_steps_per_second": 0.214,
      "step": 2210
    },
    {
      "epoch": 0.2862413826585105,
      "grad_norm": 2.252841511868517,
      "learning_rate": 1.1225817440790099e-06,
      "loss": 2.7974,
      "step": 2211
    },
    {
      "epoch": 0.286370845065864,
      "grad_norm": 1.9598246311021126,
      "learning_rate": 1.1224107596797533e-06,
      "loss": 2.5537,
      "step": 2212
    },
    {
      "epoch": 0.2865003074732175,
      "grad_norm": 3.0782150864525466,
      "learning_rate": 1.1222396736742667e-06,
      "loss": 3.0103,
      "step": 2213
    },
    {
      "epoch": 0.28662976988057093,
      "grad_norm": 2.8642008163212114,
      "learning_rate": 1.1220684860974984e-06,
      "loss": 2.6191,
      "step": 2214
    },
    {
      "epoch": 0.2867592322879244,
      "grad_norm": 1.9046792639731418,
      "learning_rate": 1.1218971969844168e-06,
      "loss": 2.5088,
      "step": 2215
    },
    {
      "epoch": 0.2867592322879244,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.71044921875,
      "eval_runtime": 14.2214,
      "eval_samples_per_second": 3.094,
      "eval_steps_per_second": 0.211,
      "step": 2215
    },
    {
      "epoch": 0.28688869469527783,
      "grad_norm": 2.3047614889168737,
      "learning_rate": 1.1217258063700109e-06,
      "loss": 2.9648,
      "step": 2216
    },
    {
      "epoch": 0.28701815710263134,
      "grad_norm": 2.0683882833066893,
      "learning_rate": 1.121554314289291e-06,
      "loss": 2.7445,
      "step": 2217
    },
    {
      "epoch": 0.2871476195099848,
      "grad_norm": 1.6373505253723921,
      "learning_rate": 1.1213827207772876e-06,
      "loss": 2.7129,
      "step": 2218
    },
    {
      "epoch": 0.28727708191733825,
      "grad_norm": 2.2017342222320244,
      "learning_rate": 1.1212110258690522e-06,
      "loss": 2.6328,
      "step": 2219
    },
    {
      "epoch": 0.2874065443246917,
      "grad_norm": 4.138346815214576,
      "learning_rate": 1.1210392295996572e-06,
      "loss": 3.5071,
      "step": 2220
    },
    {
      "epoch": 0.2874065443246917,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.706099033355713,
      "eval_runtime": 13.7416,
      "eval_samples_per_second": 3.202,
      "eval_steps_per_second": 0.218,
      "step": 2220
    },
    {
      "epoch": 0.28753600673204516,
      "grad_norm": 2.0699961736404457,
      "learning_rate": 1.1208673320041954e-06,
      "loss": 2.7375,
      "step": 2221
    },
    {
      "epoch": 0.28766546913939867,
      "grad_norm": 2.132947347687674,
      "learning_rate": 1.12069533311778e-06,
      "loss": 2.7334,
      "step": 2222
    },
    {
      "epoch": 0.2877949315467521,
      "grad_norm": 3.079188286394741,
      "learning_rate": 1.1205232329755455e-06,
      "loss": 2.8538,
      "step": 2223
    },
    {
      "epoch": 0.2879243939541056,
      "grad_norm": 3.7583789646619277,
      "learning_rate": 1.120351031612647e-06,
      "loss": 2.7488,
      "step": 2224
    },
    {
      "epoch": 0.288053856361459,
      "grad_norm": 3.195703457543387,
      "learning_rate": 1.1201787290642603e-06,
      "loss": 2.6689,
      "step": 2225
    },
    {
      "epoch": 0.288053856361459,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7103382349014282,
      "eval_runtime": 14.2532,
      "eval_samples_per_second": 3.087,
      "eval_steps_per_second": 0.21,
      "step": 2225
    },
    {
      "epoch": 0.2881833187688125,
      "grad_norm": 4.278129489842916,
      "learning_rate": 1.120006325365581e-06,
      "loss": 2.8003,
      "step": 2226
    },
    {
      "epoch": 0.288312781176166,
      "grad_norm": 2.6893671025774553,
      "learning_rate": 1.1198338205518264e-06,
      "loss": 2.9006,
      "step": 2227
    },
    {
      "epoch": 0.28844224358351944,
      "grad_norm": 1.7955960037653107,
      "learning_rate": 1.1196612146582341e-06,
      "loss": 2.6873,
      "step": 2228
    },
    {
      "epoch": 0.2885717059908729,
      "grad_norm": 3.5214441972193504,
      "learning_rate": 1.1194885077200625e-06,
      "loss": 3.0398,
      "step": 2229
    },
    {
      "epoch": 0.28870116839822635,
      "grad_norm": 3.3684742701925687,
      "learning_rate": 1.1193156997725905e-06,
      "loss": 2.9468,
      "step": 2230
    },
    {
      "epoch": 0.28870116839822635,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.705277919769287,
      "eval_runtime": 13.6533,
      "eval_samples_per_second": 3.223,
      "eval_steps_per_second": 0.22,
      "step": 2230
    },
    {
      "epoch": 0.2888306308055798,
      "grad_norm": 4.425861298149365,
      "learning_rate": 1.1191427908511171e-06,
      "loss": 3.2285,
      "step": 2231
    },
    {
      "epoch": 0.2889600932129333,
      "grad_norm": 3.7181983296949146,
      "learning_rate": 1.1189697809909632e-06,
      "loss": 2.7056,
      "step": 2232
    },
    {
      "epoch": 0.28908955562028676,
      "grad_norm": 7.126669767382223,
      "learning_rate": 1.1187966702274692e-06,
      "loss": 3.3049,
      "step": 2233
    },
    {
      "epoch": 0.2892190180276402,
      "grad_norm": 5.403212436625998,
      "learning_rate": 1.1186234585959962e-06,
      "loss": 2.7183,
      "step": 2234
    },
    {
      "epoch": 0.28934848043499367,
      "grad_norm": 2.1687910630315694,
      "learning_rate": 1.1184501461319268e-06,
      "loss": 2.7959,
      "step": 2235
    },
    {
      "epoch": 0.28934848043499367,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7081853151321411,
      "eval_runtime": 14.6207,
      "eval_samples_per_second": 3.009,
      "eval_steps_per_second": 0.205,
      "step": 2235
    },
    {
      "epoch": 0.2894779428423472,
      "grad_norm": 5.694975752507845,
      "learning_rate": 1.1182767328706633e-06,
      "loss": 3.2422,
      "step": 2236
    },
    {
      "epoch": 0.28960740524970063,
      "grad_norm": 3.3886008328371173,
      "learning_rate": 1.1181032188476288e-06,
      "loss": 3.0723,
      "step": 2237
    },
    {
      "epoch": 0.2897368676570541,
      "grad_norm": 2.798038835489442,
      "learning_rate": 1.117929604098267e-06,
      "loss": 2.7288,
      "step": 2238
    },
    {
      "epoch": 0.28986633006440754,
      "grad_norm": 2.8702299878043536,
      "learning_rate": 1.1177558886580423e-06,
      "loss": 2.9917,
      "step": 2239
    },
    {
      "epoch": 0.289995792471761,
      "grad_norm": 4.485289812994543,
      "learning_rate": 1.1175820725624397e-06,
      "loss": 2.9668,
      "step": 2240
    },
    {
      "epoch": 0.289995792471761,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7096058130264282,
      "eval_runtime": 14.228,
      "eval_samples_per_second": 3.092,
      "eval_steps_per_second": 0.211,
      "step": 2240
    },
    {
      "epoch": 0.2901252548791145,
      "grad_norm": 3.33508683390436,
      "learning_rate": 1.1174081558469647e-06,
      "loss": 2.8782,
      "step": 2241
    },
    {
      "epoch": 0.29025471728646796,
      "grad_norm": 2.277414213004642,
      "learning_rate": 1.117234138547143e-06,
      "loss": 2.7119,
      "step": 2242
    },
    {
      "epoch": 0.2903841796938214,
      "grad_norm": 1.7408559160921249,
      "learning_rate": 1.1170600206985217e-06,
      "loss": 2.5103,
      "step": 2243
    },
    {
      "epoch": 0.29051364210117486,
      "grad_norm": 1.8986201651011303,
      "learning_rate": 1.1168858023366672e-06,
      "loss": 2.5569,
      "step": 2244
    },
    {
      "epoch": 0.2906431045085283,
      "grad_norm": 4.292140335695194,
      "learning_rate": 1.1167114834971676e-06,
      "loss": 2.958,
      "step": 2245
    },
    {
      "epoch": 0.2906431045085283,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7058327198028564,
      "eval_runtime": 14.5886,
      "eval_samples_per_second": 3.016,
      "eval_steps_per_second": 0.206,
      "step": 2245
    },
    {
      "epoch": 0.2907725669158818,
      "grad_norm": 4.806996848236093,
      "learning_rate": 1.116537064215631e-06,
      "loss": 2.9956,
      "step": 2246
    },
    {
      "epoch": 0.2909020293232353,
      "grad_norm": 2.9439820261049516,
      "learning_rate": 1.116362544527686e-06,
      "loss": 2.9871,
      "step": 2247
    },
    {
      "epoch": 0.29103149173058873,
      "grad_norm": 2.579305957742088,
      "learning_rate": 1.1161879244689818e-06,
      "loss": 2.7568,
      "step": 2248
    },
    {
      "epoch": 0.2911609541379422,
      "grad_norm": 8.68177069753059,
      "learning_rate": 1.1160132040751882e-06,
      "loss": 3.366,
      "step": 2249
    },
    {
      "epoch": 0.29129041654529564,
      "grad_norm": 2.560707001929491,
      "learning_rate": 1.1158383833819953e-06,
      "loss": 2.4294,
      "step": 2250
    },
    {
      "epoch": 0.29129041654529564,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.713090419769287,
      "eval_runtime": 13.8762,
      "eval_samples_per_second": 3.171,
      "eval_steps_per_second": 0.216,
      "step": 2250
    },
    {
      "epoch": 0.29141987895264915,
      "grad_norm": 6.28404934998927,
      "learning_rate": 1.1156634624251135e-06,
      "loss": 3.0762,
      "step": 2251
    },
    {
      "epoch": 0.2915493413600026,
      "grad_norm": 4.00146180175229,
      "learning_rate": 1.1154884412402747e-06,
      "loss": 3.4414,
      "step": 2252
    },
    {
      "epoch": 0.29167880376735605,
      "grad_norm": 2.1571934650948927,
      "learning_rate": 1.1153133198632296e-06,
      "loss": 2.8652,
      "step": 2253
    },
    {
      "epoch": 0.2918082661747095,
      "grad_norm": 2.6638967026898395,
      "learning_rate": 1.1151380983297508e-06,
      "loss": 2.5972,
      "step": 2254
    },
    {
      "epoch": 0.29193772858206296,
      "grad_norm": 4.952341860227406,
      "learning_rate": 1.114962776675631e-06,
      "loss": 3.4187,
      "step": 2255
    },
    {
      "epoch": 0.29193772858206296,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.713578701019287,
      "eval_runtime": 14.5247,
      "eval_samples_per_second": 3.029,
      "eval_steps_per_second": 0.207,
      "step": 2255
    },
    {
      "epoch": 0.29206719098941647,
      "grad_norm": 3.129056514406557,
      "learning_rate": 1.1147873549366827e-06,
      "loss": 2.8794,
      "step": 2256
    },
    {
      "epoch": 0.2921966533967699,
      "grad_norm": 2.1370040412469504,
      "learning_rate": 1.1146118331487398e-06,
      "loss": 2.637,
      "step": 2257
    },
    {
      "epoch": 0.2923261158041234,
      "grad_norm": 2.457779404919745,
      "learning_rate": 1.1144362113476558e-06,
      "loss": 2.5955,
      "step": 2258
    },
    {
      "epoch": 0.29245557821147683,
      "grad_norm": 4.973510975174887,
      "learning_rate": 1.1142604895693055e-06,
      "loss": 3.0322,
      "step": 2259
    },
    {
      "epoch": 0.2925850406188303,
      "grad_norm": 1.8751211881924508,
      "learning_rate": 1.114084667849583e-06,
      "loss": 2.533,
      "step": 2260
    },
    {
      "epoch": 0.2925850406188303,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7122913599014282,
      "eval_runtime": 15.8396,
      "eval_samples_per_second": 2.778,
      "eval_steps_per_second": 0.189,
      "step": 2260
    },
    {
      "epoch": 0.2927145030261838,
      "grad_norm": 2.0114337638364734,
      "learning_rate": 1.1139087462244038e-06,
      "loss": 2.6738,
      "step": 2261
    },
    {
      "epoch": 0.29284396543353725,
      "grad_norm": 5.29225363324954,
      "learning_rate": 1.113732724729704e-06,
      "loss": 2.9424,
      "step": 2262
    },
    {
      "epoch": 0.2929734278408907,
      "grad_norm": 1.9882829970955298,
      "learning_rate": 1.1135566034014384e-06,
      "loss": 2.9143,
      "step": 2263
    },
    {
      "epoch": 0.29310289024824415,
      "grad_norm": 2.542988863606374,
      "learning_rate": 1.1133803822755843e-06,
      "loss": 2.6362,
      "step": 2264
    },
    {
      "epoch": 0.2932323526555976,
      "grad_norm": 2.3777462071011293,
      "learning_rate": 1.1132040613881378e-06,
      "loss": 2.7874,
      "step": 2265
    },
    {
      "epoch": 0.2932323526555976,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7107599973678589,
      "eval_runtime": 14.3715,
      "eval_samples_per_second": 3.062,
      "eval_steps_per_second": 0.209,
      "step": 2265
    },
    {
      "epoch": 0.2933618150629511,
      "grad_norm": 1.618431288280825,
      "learning_rate": 1.1130276407751167e-06,
      "loss": 2.3726,
      "step": 2266
    },
    {
      "epoch": 0.29349127747030457,
      "grad_norm": 1.7050746325844315,
      "learning_rate": 1.1128511204725579e-06,
      "loss": 2.5781,
      "step": 2267
    },
    {
      "epoch": 0.293620739877658,
      "grad_norm": 1.5178288970628233,
      "learning_rate": 1.1126745005165194e-06,
      "loss": 2.4829,
      "step": 2268
    },
    {
      "epoch": 0.2937502022850115,
      "grad_norm": 4.677534963476821,
      "learning_rate": 1.1124977809430794e-06,
      "loss": 2.4578,
      "step": 2269
    },
    {
      "epoch": 0.29387966469236493,
      "grad_norm": 2.0305203104429923,
      "learning_rate": 1.1123209617883368e-06,
      "loss": 2.5623,
      "step": 2270
    },
    {
      "epoch": 0.29387966469236493,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7080078125,
      "eval_runtime": 14.8327,
      "eval_samples_per_second": 2.966,
      "eval_steps_per_second": 0.202,
      "step": 2270
    },
    {
      "epoch": 0.29400912709971844,
      "grad_norm": 3.9270302038057214,
      "learning_rate": 1.11214404308841e-06,
      "loss": 2.7964,
      "step": 2271
    },
    {
      "epoch": 0.2941385895070719,
      "grad_norm": 2.462702196807948,
      "learning_rate": 1.1119670248794384e-06,
      "loss": 2.502,
      "step": 2272
    },
    {
      "epoch": 0.29426805191442534,
      "grad_norm": 4.34227279463719,
      "learning_rate": 1.1117899071975819e-06,
      "loss": 2.9658,
      "step": 2273
    },
    {
      "epoch": 0.2943975143217788,
      "grad_norm": 4.6700635772428125,
      "learning_rate": 1.1116126900790197e-06,
      "loss": 2.9985,
      "step": 2274
    },
    {
      "epoch": 0.29452697672913225,
      "grad_norm": 10.694848827573995,
      "learning_rate": 1.1114353735599525e-06,
      "loss": 3.7954,
      "step": 2275
    },
    {
      "epoch": 0.29452697672913225,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7079411745071411,
      "eval_runtime": 13.7701,
      "eval_samples_per_second": 3.195,
      "eval_steps_per_second": 0.218,
      "step": 2275
    },
    {
      "epoch": 0.29465643913648576,
      "grad_norm": 6.199938665661265,
      "learning_rate": 1.111257957676601e-06,
      "loss": 3.0825,
      "step": 2276
    },
    {
      "epoch": 0.2947859015438392,
      "grad_norm": 2.3000652490116313,
      "learning_rate": 1.111080442465205e-06,
      "loss": 2.6416,
      "step": 2277
    },
    {
      "epoch": 0.29491536395119267,
      "grad_norm": 1.608724350549297,
      "learning_rate": 1.1109028279620266e-06,
      "loss": 2.7773,
      "step": 2278
    },
    {
      "epoch": 0.2950448263585461,
      "grad_norm": 3.5004184752607883,
      "learning_rate": 1.1107251142033468e-06,
      "loss": 2.8203,
      "step": 2279
    },
    {
      "epoch": 0.29517428876589963,
      "grad_norm": 6.078583931008407,
      "learning_rate": 1.110547301225467e-06,
      "loss": 3.416,
      "step": 2280
    },
    {
      "epoch": 0.29517428876589963,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.713667392730713,
      "eval_runtime": 14.9524,
      "eval_samples_per_second": 2.943,
      "eval_steps_per_second": 0.201,
      "step": 2280
    },
    {
      "epoch": 0.2953037511732531,
      "grad_norm": 2.5075926605183962,
      "learning_rate": 1.1103693890647097e-06,
      "loss": 2.6797,
      "step": 2281
    },
    {
      "epoch": 0.29543321358060654,
      "grad_norm": 2.725853349756007,
      "learning_rate": 1.1101913777574164e-06,
      "loss": 2.7324,
      "step": 2282
    },
    {
      "epoch": 0.29556267598796,
      "grad_norm": 2.8218512901821917,
      "learning_rate": 1.1100132673399498e-06,
      "loss": 3.1111,
      "step": 2283
    },
    {
      "epoch": 0.29569213839531344,
      "grad_norm": 5.088712382362241,
      "learning_rate": 1.1098350578486927e-06,
      "loss": 2.8916,
      "step": 2284
    },
    {
      "epoch": 0.29582160080266695,
      "grad_norm": 5.494502510637724,
      "learning_rate": 1.1096567493200477e-06,
      "loss": 3.0034,
      "step": 2285
    },
    {
      "epoch": 0.29582160080266695,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7105602025985718,
      "eval_runtime": 13.5554,
      "eval_samples_per_second": 3.246,
      "eval_steps_per_second": 0.221,
      "step": 2285
    },
    {
      "epoch": 0.2959510632100204,
      "grad_norm": 3.1714050045330002,
      "learning_rate": 1.1094783417904385e-06,
      "loss": 3.0225,
      "step": 2286
    },
    {
      "epoch": 0.29608052561737386,
      "grad_norm": 2.990935943544762,
      "learning_rate": 1.1092998352963078e-06,
      "loss": 2.9847,
      "step": 2287
    },
    {
      "epoch": 0.2962099880247273,
      "grad_norm": 3.1417296172973757,
      "learning_rate": 1.1091212298741195e-06,
      "loss": 2.8726,
      "step": 2288
    },
    {
      "epoch": 0.29633945043208076,
      "grad_norm": 1.759831550793735,
      "learning_rate": 1.108942525560357e-06,
      "loss": 2.6489,
      "step": 2289
    },
    {
      "epoch": 0.2964689128394343,
      "grad_norm": 2.8246089030720825,
      "learning_rate": 1.1087637223915249e-06,
      "loss": 2.6543,
      "step": 2290
    },
    {
      "epoch": 0.2964689128394343,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7157536745071411,
      "eval_runtime": 15.0643,
      "eval_samples_per_second": 2.921,
      "eval_steps_per_second": 0.199,
      "step": 2290
    },
    {
      "epoch": 0.2965983752467877,
      "grad_norm": 4.684436599881485,
      "learning_rate": 1.1085848204041468e-06,
      "loss": 2.7144,
      "step": 2291
    },
    {
      "epoch": 0.2967278376541412,
      "grad_norm": 5.325197371726781,
      "learning_rate": 1.1084058196347676e-06,
      "loss": 3.3655,
      "step": 2292
    },
    {
      "epoch": 0.29685730006149463,
      "grad_norm": 3.5037919949092777,
      "learning_rate": 1.1082267201199515e-06,
      "loss": 2.814,
      "step": 2293
    },
    {
      "epoch": 0.2969867624688481,
      "grad_norm": 3.2014843591185596,
      "learning_rate": 1.1080475218962832e-06,
      "loss": 2.9329,
      "step": 2294
    },
    {
      "epoch": 0.2971162248762016,
      "grad_norm": 7.271137073874824,
      "learning_rate": 1.1078682250003677e-06,
      "loss": 3.5168,
      "step": 2295
    },
    {
      "epoch": 0.2971162248762016,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7149547338485718,
      "eval_runtime": 14.2664,
      "eval_samples_per_second": 3.084,
      "eval_steps_per_second": 0.21,
      "step": 2295
    },
    {
      "epoch": 0.29724568728355505,
      "grad_norm": 4.901814231369322,
      "learning_rate": 1.1076888294688298e-06,
      "loss": 3.1787,
      "step": 2296
    },
    {
      "epoch": 0.2973751496909085,
      "grad_norm": 2.219257235597538,
      "learning_rate": 1.1075093353383148e-06,
      "loss": 2.8198,
      "step": 2297
    },
    {
      "epoch": 0.29750461209826196,
      "grad_norm": 1.357931993492483,
      "learning_rate": 1.1073297426454883e-06,
      "loss": 2.3682,
      "step": 2298
    },
    {
      "epoch": 0.2976340745056154,
      "grad_norm": 5.267575887649589,
      "learning_rate": 1.1071500514270354e-06,
      "loss": 3.2883,
      "step": 2299
    },
    {
      "epoch": 0.2977635369129689,
      "grad_norm": 3.817558876719962,
      "learning_rate": 1.1069702617196616e-06,
      "loss": 3.0457,
      "step": 2300
    },
    {
      "epoch": 0.2977635369129689,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7156871557235718,
      "eval_runtime": 13.8965,
      "eval_samples_per_second": 3.166,
      "eval_steps_per_second": 0.216,
      "step": 2300
    },
    {
      "epoch": 0.29789299932032237,
      "grad_norm": 2.2686835609527787,
      "learning_rate": 1.1067903735600928e-06,
      "loss": 3.019,
      "step": 2301
    },
    {
      "epoch": 0.2980224617276758,
      "grad_norm": 8.127838148779066,
      "learning_rate": 1.1066103869850749e-06,
      "loss": 3.4963,
      "step": 2302
    },
    {
      "epoch": 0.2981519241350293,
      "grad_norm": 7.454411254477891,
      "learning_rate": 1.1064303020313736e-06,
      "loss": 3.6494,
      "step": 2303
    },
    {
      "epoch": 0.29828138654238273,
      "grad_norm": 3.620502155925507,
      "learning_rate": 1.1062501187357747e-06,
      "loss": 2.8923,
      "step": 2304
    },
    {
      "epoch": 0.29841084894973624,
      "grad_norm": 7.703118869295989,
      "learning_rate": 1.1060698371350849e-06,
      "loss": 3.0095,
      "step": 2305
    },
    {
      "epoch": 0.29841084894973624,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.720414638519287,
      "eval_runtime": 13.8658,
      "eval_samples_per_second": 3.173,
      "eval_steps_per_second": 0.216,
      "step": 2305
    },
    {
      "epoch": 0.2985403113570897,
      "grad_norm": 5.672908820019964,
      "learning_rate": 1.10588945726613e-06,
      "loss": 3.0894,
      "step": 2306
    },
    {
      "epoch": 0.29866977376444315,
      "grad_norm": 1.7377480439703714,
      "learning_rate": 1.105708979165756e-06,
      "loss": 2.5754,
      "step": 2307
    },
    {
      "epoch": 0.2987992361717966,
      "grad_norm": 2.755588309548811,
      "learning_rate": 1.1055284028708297e-06,
      "loss": 2.4216,
      "step": 2308
    },
    {
      "epoch": 0.29892869857915005,
      "grad_norm": 4.3430421940125346,
      "learning_rate": 1.1053477284182373e-06,
      "loss": 2.8965,
      "step": 2309
    },
    {
      "epoch": 0.29905816098650356,
      "grad_norm": 2.9417969979345906,
      "learning_rate": 1.105166955844885e-06,
      "loss": 2.8525,
      "step": 2310
    },
    {
      "epoch": 0.29905816098650356,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.719682216644287,
      "eval_runtime": 14.4237,
      "eval_samples_per_second": 3.051,
      "eval_steps_per_second": 0.208,
      "step": 2310
    },
    {
      "epoch": 0.299187623393857,
      "grad_norm": 5.464198809873469,
      "learning_rate": 1.1049860851876996e-06,
      "loss": 3.3276,
      "step": 2311
    },
    {
      "epoch": 0.29931708580121047,
      "grad_norm": 4.174417484442815,
      "learning_rate": 1.1048051164836271e-06,
      "loss": 3.0596,
      "step": 2312
    },
    {
      "epoch": 0.2994465482085639,
      "grad_norm": 2.583972012195666,
      "learning_rate": 1.1046240497696345e-06,
      "loss": 2.8989,
      "step": 2313
    },
    {
      "epoch": 0.2995760106159174,
      "grad_norm": 4.851229023177933,
      "learning_rate": 1.1044428850827083e-06,
      "loss": 3.2666,
      "step": 2314
    },
    {
      "epoch": 0.2997054730232709,
      "grad_norm": 4.0543917094532,
      "learning_rate": 1.104261622459855e-06,
      "loss": 3.0627,
      "step": 2315
    },
    {
      "epoch": 0.2997054730232709,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.723921298980713,
      "eval_runtime": 14.797,
      "eval_samples_per_second": 2.974,
      "eval_steps_per_second": 0.203,
      "step": 2315
    },
    {
      "epoch": 0.29983493543062434,
      "grad_norm": 4.332318062558142,
      "learning_rate": 1.104080261938101e-06,
      "loss": 2.9961,
      "step": 2316
    },
    {
      "epoch": 0.2999643978379778,
      "grad_norm": 2.3053382259492112,
      "learning_rate": 1.103898803554493e-06,
      "loss": 2.739,
      "step": 2317
    },
    {
      "epoch": 0.30009386024533125,
      "grad_norm": 5.479569282279513,
      "learning_rate": 1.1037172473460976e-06,
      "loss": 2.949,
      "step": 2318
    },
    {
      "epoch": 0.30022332265268475,
      "grad_norm": 4.024732523908445,
      "learning_rate": 1.1035355933500012e-06,
      "loss": 2.7861,
      "step": 2319
    },
    {
      "epoch": 0.3003527850600382,
      "grad_norm": 2.3241983284303758,
      "learning_rate": 1.1033538416033105e-06,
      "loss": 2.7703,
      "step": 2320
    },
    {
      "epoch": 0.3003527850600382,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7218571901321411,
      "eval_runtime": 14.0487,
      "eval_samples_per_second": 3.132,
      "eval_steps_per_second": 0.214,
      "step": 2320
    },
    {
      "epoch": 0.30048224746739166,
      "grad_norm": 2.8018145010136335,
      "learning_rate": 1.1031719921431517e-06,
      "loss": 2.9968,
      "step": 2321
    },
    {
      "epoch": 0.3006117098747451,
      "grad_norm": 2.7994496703703327,
      "learning_rate": 1.1029900450066715e-06,
      "loss": 2.4685,
      "step": 2322
    },
    {
      "epoch": 0.30074117228209857,
      "grad_norm": 2.359341875876882,
      "learning_rate": 1.1028080002310361e-06,
      "loss": 2.4744,
      "step": 2323
    },
    {
      "epoch": 0.3008706346894521,
      "grad_norm": 3.3150985146941876,
      "learning_rate": 1.1026258578534322e-06,
      "loss": 2.825,
      "step": 2324
    },
    {
      "epoch": 0.30100009709680553,
      "grad_norm": 2.311207815437649,
      "learning_rate": 1.1024436179110654e-06,
      "loss": 2.8203,
      "step": 2325
    },
    {
      "epoch": 0.30100009709680553,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.726806640625,
      "eval_runtime": 14.0322,
      "eval_samples_per_second": 3.136,
      "eval_steps_per_second": 0.214,
      "step": 2325
    },
    {
      "epoch": 0.301129559504159,
      "grad_norm": 2.7065193663620923,
      "learning_rate": 1.1022612804411625e-06,
      "loss": 2.9934,
      "step": 2326
    },
    {
      "epoch": 0.30125902191151244,
      "grad_norm": 6.4200675593452265,
      "learning_rate": 1.1020788454809695e-06,
      "loss": 3.2385,
      "step": 2327
    },
    {
      "epoch": 0.3013884843188659,
      "grad_norm": 3.918861280770376,
      "learning_rate": 1.1018963130677523e-06,
      "loss": 2.6145,
      "step": 2328
    },
    {
      "epoch": 0.3015179467262194,
      "grad_norm": 2.73809457857486,
      "learning_rate": 1.1017136832387967e-06,
      "loss": 2.6553,
      "step": 2329
    },
    {
      "epoch": 0.30164740913357285,
      "grad_norm": 1.8389206163537233,
      "learning_rate": 1.101530956031409e-06,
      "loss": 2.6775,
      "step": 2330
    },
    {
      "epoch": 0.30164740913357285,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7268732786178589,
      "eval_runtime": 15.0474,
      "eval_samples_per_second": 2.924,
      "eval_steps_per_second": 0.199,
      "step": 2330
    },
    {
      "epoch": 0.3017768715409263,
      "grad_norm": 2.3396551197934907,
      "learning_rate": 1.1013481314829146e-06,
      "loss": 2.7395,
      "step": 2331
    },
    {
      "epoch": 0.30190633394827976,
      "grad_norm": 2.202878194051444,
      "learning_rate": 1.1011652096306593e-06,
      "loss": 2.4819,
      "step": 2332
    },
    {
      "epoch": 0.3020357963556332,
      "grad_norm": 2.1879117401839703,
      "learning_rate": 1.1009821905120087e-06,
      "loss": 2.7915,
      "step": 2333
    },
    {
      "epoch": 0.3021652587629867,
      "grad_norm": 2.233376883365686,
      "learning_rate": 1.1007990741643476e-06,
      "loss": 2.6038,
      "step": 2334
    },
    {
      "epoch": 0.3022947211703402,
      "grad_norm": 4.103359338062117,
      "learning_rate": 1.1006158606250816e-06,
      "loss": 3.2329,
      "step": 2335
    },
    {
      "epoch": 0.3022947211703402,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7268732786178589,
      "eval_runtime": 14.1474,
      "eval_samples_per_second": 3.11,
      "eval_steps_per_second": 0.212,
      "step": 2335
    },
    {
      "epoch": 0.30242418357769363,
      "grad_norm": 3.1153083607544505,
      "learning_rate": 1.100432549931636e-06,
      "loss": 2.5474,
      "step": 2336
    },
    {
      "epoch": 0.3025536459850471,
      "grad_norm": 2.2457332628017412,
      "learning_rate": 1.1002491421214552e-06,
      "loss": 2.6084,
      "step": 2337
    },
    {
      "epoch": 0.30268310839240054,
      "grad_norm": 3.770406804964287,
      "learning_rate": 1.1000656372320045e-06,
      "loss": 2.7412,
      "step": 2338
    },
    {
      "epoch": 0.30281257079975404,
      "grad_norm": 3.290597648196265,
      "learning_rate": 1.0998820353007683e-06,
      "loss": 3.0625,
      "step": 2339
    },
    {
      "epoch": 0.3029420332071075,
      "grad_norm": 3.3024932641343496,
      "learning_rate": 1.099698336365251e-06,
      "loss": 2.8193,
      "step": 2340
    },
    {
      "epoch": 0.3029420332071075,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7282493114471436,
      "eval_runtime": 14.838,
      "eval_samples_per_second": 2.965,
      "eval_steps_per_second": 0.202,
      "step": 2340
    },
    {
      "epoch": 0.30307149561446095,
      "grad_norm": 2.673703705595916,
      "learning_rate": 1.0995145404629764e-06,
      "loss": 2.813,
      "step": 2341
    },
    {
      "epoch": 0.3032009580218144,
      "grad_norm": 1.9614944555464808,
      "learning_rate": 1.0993306476314892e-06,
      "loss": 2.8359,
      "step": 2342
    },
    {
      "epoch": 0.30333042042916786,
      "grad_norm": 4.077729405270378,
      "learning_rate": 1.099146657908353e-06,
      "loss": 2.5928,
      "step": 2343
    },
    {
      "epoch": 0.30345988283652137,
      "grad_norm": 2.163937911996922,
      "learning_rate": 1.0989625713311514e-06,
      "loss": 2.6296,
      "step": 2344
    },
    {
      "epoch": 0.3035893452438748,
      "grad_norm": 1.7445607961316325,
      "learning_rate": 1.0987783879374878e-06,
      "loss": 2.6089,
      "step": 2345
    },
    {
      "epoch": 0.3035893452438748,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7274724245071411,
      "eval_runtime": 14.3576,
      "eval_samples_per_second": 3.065,
      "eval_steps_per_second": 0.209,
      "step": 2345
    },
    {
      "epoch": 0.3037188076512283,
      "grad_norm": 3.306514351501887,
      "learning_rate": 1.0985941077649854e-06,
      "loss": 2.8125,
      "step": 2346
    },
    {
      "epoch": 0.3038482700585817,
      "grad_norm": 1.5654794429116161,
      "learning_rate": 1.098409730851287e-06,
      "loss": 2.5405,
      "step": 2347
    },
    {
      "epoch": 0.3039777324659352,
      "grad_norm": 3.028447916983368,
      "learning_rate": 1.098225257234056e-06,
      "loss": 2.9534,
      "step": 2348
    },
    {
      "epoch": 0.3041071948732887,
      "grad_norm": 2.555158984805008,
      "learning_rate": 1.0980406869509737e-06,
      "loss": 2.5488,
      "step": 2349
    },
    {
      "epoch": 0.30423665728064214,
      "grad_norm": 2.1549844505405695,
      "learning_rate": 1.0978560200397434e-06,
      "loss": 2.7473,
      "step": 2350
    },
    {
      "epoch": 0.30423665728064214,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.728227138519287,
      "eval_runtime": 14.7034,
      "eval_samples_per_second": 2.993,
      "eval_steps_per_second": 0.204,
      "step": 2350
    },
    {
      "epoch": 0.3043661196879956,
      "grad_norm": 4.344356650267298,
      "learning_rate": 1.0976712565380868e-06,
      "loss": 2.7275,
      "step": 2351
    },
    {
      "epoch": 0.30449558209534905,
      "grad_norm": 1.9759526783379688,
      "learning_rate": 1.0974863964837452e-06,
      "loss": 2.7452,
      "step": 2352
    },
    {
      "epoch": 0.3046250445027025,
      "grad_norm": 2.918339179192121,
      "learning_rate": 1.0973014399144804e-06,
      "loss": 2.8164,
      "step": 2353
    },
    {
      "epoch": 0.304754506910056,
      "grad_norm": 1.8913136195210352,
      "learning_rate": 1.0971163868680732e-06,
      "loss": 2.567,
      "step": 2354
    },
    {
      "epoch": 0.30488396931740946,
      "grad_norm": 3.5293196141146024,
      "learning_rate": 1.0969312373823246e-06,
      "loss": 2.5986,
      "step": 2355
    },
    {
      "epoch": 0.30488396931740946,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7251864671707153,
      "eval_runtime": 13.9786,
      "eval_samples_per_second": 3.148,
      "eval_steps_per_second": 0.215,
      "step": 2355
    },
    {
      "epoch": 0.3050134317247629,
      "grad_norm": 1.3062140494467265,
      "learning_rate": 1.0967459914950555e-06,
      "loss": 2.3357,
      "step": 2356
    },
    {
      "epoch": 0.30514289413211637,
      "grad_norm": 2.7197663773225043,
      "learning_rate": 1.0965606492441055e-06,
      "loss": 3.1416,
      "step": 2357
    },
    {
      "epoch": 0.3052723565394698,
      "grad_norm": 3.2655455323575144,
      "learning_rate": 1.0963752106673348e-06,
      "loss": 2.874,
      "step": 2358
    },
    {
      "epoch": 0.30540181894682333,
      "grad_norm": 3.49802099192436,
      "learning_rate": 1.0961896758026231e-06,
      "loss": 2.9258,
      "step": 2359
    },
    {
      "epoch": 0.3055312813541768,
      "grad_norm": 1.9628910595002096,
      "learning_rate": 1.0960040446878691e-06,
      "loss": 2.7124,
      "step": 2360
    },
    {
      "epoch": 0.3055312813541768,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7253639698028564,
      "eval_runtime": 14.1944,
      "eval_samples_per_second": 3.1,
      "eval_steps_per_second": 0.211,
      "step": 2360
    },
    {
      "epoch": 0.30566074376153024,
      "grad_norm": 2.4587835627703334,
      "learning_rate": 1.0958183173609927e-06,
      "loss": 2.7305,
      "step": 2361
    },
    {
      "epoch": 0.3057902061688837,
      "grad_norm": 1.6832969713945503,
      "learning_rate": 1.0956324938599317e-06,
      "loss": 2.4036,
      "step": 2362
    },
    {
      "epoch": 0.3059196685762372,
      "grad_norm": 4.404962521109919,
      "learning_rate": 1.0954465742226446e-06,
      "loss": 2.76,
      "step": 2363
    },
    {
      "epoch": 0.30604913098359066,
      "grad_norm": 2.374721501109396,
      "learning_rate": 1.0952605584871092e-06,
      "loss": 2.5376,
      "step": 2364
    },
    {
      "epoch": 0.3061785933909441,
      "grad_norm": 3.664580468008269,
      "learning_rate": 1.095074446691323e-06,
      "loss": 2.5071,
      "step": 2365
    },
    {
      "epoch": 0.3061785933909441,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.723832607269287,
      "eval_runtime": 14.6734,
      "eval_samples_per_second": 2.999,
      "eval_steps_per_second": 0.204,
      "step": 2365
    },
    {
      "epoch": 0.30630805579829756,
      "grad_norm": 2.9028991689676436,
      "learning_rate": 1.0948882388733028e-06,
      "loss": 2.6248,
      "step": 2366
    },
    {
      "epoch": 0.306437518205651,
      "grad_norm": 1.7629813529037595,
      "learning_rate": 1.094701935071086e-06,
      "loss": 2.5582,
      "step": 2367
    },
    {
      "epoch": 0.3065669806130045,
      "grad_norm": 2.092818393140363,
      "learning_rate": 1.0945155353227284e-06,
      "loss": 2.6292,
      "step": 2368
    },
    {
      "epoch": 0.306696443020358,
      "grad_norm": 6.977125321100761,
      "learning_rate": 1.094329039666306e-06,
      "loss": 2.5737,
      "step": 2369
    },
    {
      "epoch": 0.30682590542771143,
      "grad_norm": 8.652030826571032,
      "learning_rate": 1.0941424481399145e-06,
      "loss": 3.5762,
      "step": 2370
    },
    {
      "epoch": 0.30682590542771143,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7288485765457153,
      "eval_runtime": 15.0853,
      "eval_samples_per_second": 2.917,
      "eval_steps_per_second": 0.199,
      "step": 2370
    },
    {
      "epoch": 0.3069553678350649,
      "grad_norm": 2.345543682802869,
      "learning_rate": 1.0939557607816687e-06,
      "loss": 2.759,
      "step": 2371
    },
    {
      "epoch": 0.30708483024241834,
      "grad_norm": 3.5004531029562096,
      "learning_rate": 1.0937689776297036e-06,
      "loss": 2.8242,
      "step": 2372
    },
    {
      "epoch": 0.30721429264977185,
      "grad_norm": 3.0975953965554592,
      "learning_rate": 1.0935820987221734e-06,
      "loss": 2.8235,
      "step": 2373
    },
    {
      "epoch": 0.3073437550571253,
      "grad_norm": 3.548805030095774,
      "learning_rate": 1.0933951240972517e-06,
      "loss": 3.1929,
      "step": 2374
    },
    {
      "epoch": 0.30747321746447875,
      "grad_norm": 2.3444064758342638,
      "learning_rate": 1.0932080537931319e-06,
      "loss": 2.7817,
      "step": 2375
    },
    {
      "epoch": 0.30747321746447875,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.728759765625,
      "eval_runtime": 14.9551,
      "eval_samples_per_second": 2.942,
      "eval_steps_per_second": 0.201,
      "step": 2375
    },
    {
      "epoch": 0.3076026798718322,
      "grad_norm": 1.9652062446980842,
      "learning_rate": 1.093020887848027e-06,
      "loss": 2.5242,
      "step": 2376
    },
    {
      "epoch": 0.30773214227918566,
      "grad_norm": 4.292845713711578,
      "learning_rate": 1.0928336263001694e-06,
      "loss": 2.8331,
      "step": 2377
    },
    {
      "epoch": 0.30786160468653917,
      "grad_norm": 2.1080984858369374,
      "learning_rate": 1.0926462691878113e-06,
      "loss": 2.7549,
      "step": 2378
    },
    {
      "epoch": 0.3079910670938926,
      "grad_norm": 2.9578317131351026,
      "learning_rate": 1.0924588165492236e-06,
      "loss": 2.9641,
      "step": 2379
    },
    {
      "epoch": 0.3081205295012461,
      "grad_norm": 4.041543951211056,
      "learning_rate": 1.0922712684226977e-06,
      "loss": 3.1533,
      "step": 2380
    },
    {
      "epoch": 0.3081205295012461,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7217463254928589,
      "eval_runtime": 14.1731,
      "eval_samples_per_second": 3.104,
      "eval_steps_per_second": 0.212,
      "step": 2380
    },
    {
      "epoch": 0.30824999190859953,
      "grad_norm": 3.7349426704360096,
      "learning_rate": 1.0920836248465444e-06,
      "loss": 2.8679,
      "step": 2381
    },
    {
      "epoch": 0.308379454315953,
      "grad_norm": 2.3819826233324326,
      "learning_rate": 1.091895885859093e-06,
      "loss": 2.7231,
      "step": 2382
    },
    {
      "epoch": 0.3085089167233065,
      "grad_norm": 1.5817564064631358,
      "learning_rate": 1.0917080514986936e-06,
      "loss": 2.5522,
      "step": 2383
    },
    {
      "epoch": 0.30863837913065995,
      "grad_norm": 3.1587163866305343,
      "learning_rate": 1.091520121803715e-06,
      "loss": 3.1121,
      "step": 2384
    },
    {
      "epoch": 0.3087678415380134,
      "grad_norm": 4.826990616632971,
      "learning_rate": 1.0913320968125454e-06,
      "loss": 2.856,
      "step": 2385
    },
    {
      "epoch": 0.3087678415380134,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7281161546707153,
      "eval_runtime": 13.8847,
      "eval_samples_per_second": 3.169,
      "eval_steps_per_second": 0.216,
      "step": 2385
    },
    {
      "epoch": 0.30889730394536685,
      "grad_norm": 2.2901727560733858,
      "learning_rate": 1.0911439765635927e-06,
      "loss": 2.6177,
      "step": 2386
    },
    {
      "epoch": 0.3090267663527203,
      "grad_norm": 2.7665876493584816,
      "learning_rate": 1.090955761095285e-06,
      "loss": 2.4453,
      "step": 2387
    },
    {
      "epoch": 0.3091562287600738,
      "grad_norm": 2.071003556745559,
      "learning_rate": 1.0907674504460684e-06,
      "loss": 2.666,
      "step": 2388
    },
    {
      "epoch": 0.30928569116742727,
      "grad_norm": 5.602577345978535,
      "learning_rate": 1.0905790446544094e-06,
      "loss": 2.7539,
      "step": 2389
    },
    {
      "epoch": 0.3094151535747807,
      "grad_norm": 2.509865921768195,
      "learning_rate": 1.0903905437587934e-06,
      "loss": 2.9897,
      "step": 2390
    },
    {
      "epoch": 0.3094151535747807,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.723100185394287,
      "eval_runtime": 14.0773,
      "eval_samples_per_second": 3.126,
      "eval_steps_per_second": 0.213,
      "step": 2390
    },
    {
      "epoch": 0.3095446159821342,
      "grad_norm": 4.721741217295041,
      "learning_rate": 1.090201947797726e-06,
      "loss": 2.9995,
      "step": 2391
    },
    {
      "epoch": 0.30967407838948763,
      "grad_norm": 3.1456676592206936,
      "learning_rate": 1.0900132568097315e-06,
      "loss": 2.9624,
      "step": 2392
    },
    {
      "epoch": 0.30980354079684114,
      "grad_norm": 1.4610708989058652,
      "learning_rate": 1.089824470833354e-06,
      "loss": 2.561,
      "step": 2393
    },
    {
      "epoch": 0.3099330032041946,
      "grad_norm": 2.6181156043233225,
      "learning_rate": 1.0896355899071568e-06,
      "loss": 2.5471,
      "step": 2394
    },
    {
      "epoch": 0.31006246561154804,
      "grad_norm": 5.9134217414205485,
      "learning_rate": 1.0894466140697225e-06,
      "loss": 2.7651,
      "step": 2395
    },
    {
      "epoch": 0.31006246561154804,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7242985963821411,
      "eval_runtime": 15.146,
      "eval_samples_per_second": 2.905,
      "eval_steps_per_second": 0.198,
      "step": 2395
    },
    {
      "epoch": 0.3101919280189015,
      "grad_norm": 8.412361559878502,
      "learning_rate": 1.0892575433596533e-06,
      "loss": 3.1011,
      "step": 2396
    },
    {
      "epoch": 0.31032139042625495,
      "grad_norm": 5.419321444894031,
      "learning_rate": 1.0890683778155708e-06,
      "loss": 2.8979,
      "step": 2397
    },
    {
      "epoch": 0.31045085283360846,
      "grad_norm": 8.117119304665449,
      "learning_rate": 1.0888791174761162e-06,
      "loss": 3.2632,
      "step": 2398
    },
    {
      "epoch": 0.3105803152409619,
      "grad_norm": 5.778830542819061,
      "learning_rate": 1.088689762379949e-06,
      "loss": 3.231,
      "step": 2399
    },
    {
      "epoch": 0.31070977764831537,
      "grad_norm": 3.503842564948514,
      "learning_rate": 1.0885003125657497e-06,
      "loss": 2.8774,
      "step": 2400
    },
    {
      "epoch": 0.31070977764831537,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7264958620071411,
      "eval_runtime": 14.483,
      "eval_samples_per_second": 3.038,
      "eval_steps_per_second": 0.207,
      "step": 2400
    },
    {
      "epoch": 0.3108392400556688,
      "grad_norm": 2.609445130844578,
      "learning_rate": 1.0883107680722167e-06,
      "loss": 2.793,
      "step": 2401
    },
    {
      "epoch": 0.3109687024630223,
      "grad_norm": 3.201048399294415,
      "learning_rate": 1.0881211289380682e-06,
      "loss": 2.7988,
      "step": 2402
    },
    {
      "epoch": 0.3110981648703758,
      "grad_norm": 2.9433106975568752,
      "learning_rate": 1.0879313952020424e-06,
      "loss": 2.7368,
      "step": 2403
    },
    {
      "epoch": 0.31122762727772924,
      "grad_norm": 3.287459843621182,
      "learning_rate": 1.087741566902896e-06,
      "loss": 2.4546,
      "step": 2404
    },
    {
      "epoch": 0.3113570896850827,
      "grad_norm": 2.545160737961092,
      "learning_rate": 1.0875516440794048e-06,
      "loss": 2.5977,
      "step": 2405
    },
    {
      "epoch": 0.3113570896850827,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7250310182571411,
      "eval_runtime": 14.8953,
      "eval_samples_per_second": 2.954,
      "eval_steps_per_second": 0.201,
      "step": 2405
    },
    {
      "epoch": 0.31148655209243614,
      "grad_norm": 4.7298583669312215,
      "learning_rate": 1.0873616267703651e-06,
      "loss": 2.7993,
      "step": 2406
    },
    {
      "epoch": 0.31161601449978965,
      "grad_norm": 1.6697275418840263,
      "learning_rate": 1.0871715150145915e-06,
      "loss": 2.6411,
      "step": 2407
    },
    {
      "epoch": 0.3117454769071431,
      "grad_norm": 4.145696219088108,
      "learning_rate": 1.086981308850918e-06,
      "loss": 3.0166,
      "step": 2408
    },
    {
      "epoch": 0.31187493931449656,
      "grad_norm": 4.1341049890440935,
      "learning_rate": 1.0867910083181984e-06,
      "loss": 3.1152,
      "step": 2409
    },
    {
      "epoch": 0.31200440172185,
      "grad_norm": 1.7365756726224395,
      "learning_rate": 1.0866006134553053e-06,
      "loss": 2.5522,
      "step": 2410
    },
    {
      "epoch": 0.31200440172185,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7258079051971436,
      "eval_runtime": 14.1523,
      "eval_samples_per_second": 3.109,
      "eval_steps_per_second": 0.212,
      "step": 2410
    },
    {
      "epoch": 0.31213386412920346,
      "grad_norm": 3.8991441737204475,
      "learning_rate": 1.0864101243011306e-06,
      "loss": 2.5142,
      "step": 2411
    },
    {
      "epoch": 0.312263326536557,
      "grad_norm": 3.2662779745530073,
      "learning_rate": 1.0862195408945856e-06,
      "loss": 2.8677,
      "step": 2412
    },
    {
      "epoch": 0.3123927889439104,
      "grad_norm": 4.277262769897312,
      "learning_rate": 1.0860288632746007e-06,
      "loss": 2.8499,
      "step": 2413
    },
    {
      "epoch": 0.3125222513512639,
      "grad_norm": 5.01450381901903,
      "learning_rate": 1.085838091480126e-06,
      "loss": 2.8115,
      "step": 2414
    },
    {
      "epoch": 0.31265171375861733,
      "grad_norm": 5.140632361962117,
      "learning_rate": 1.0856472255501307e-06,
      "loss": 2.7261,
      "step": 2415
    },
    {
      "epoch": 0.31265171375861733,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7264958620071411,
      "eval_runtime": 15.1146,
      "eval_samples_per_second": 2.911,
      "eval_steps_per_second": 0.198,
      "step": 2415
    },
    {
      "epoch": 0.3127811761659708,
      "grad_norm": 1.5762636180296328,
      "learning_rate": 1.0854562655236022e-06,
      "loss": 2.6777,
      "step": 2416
    },
    {
      "epoch": 0.3129106385733243,
      "grad_norm": 2.811376902037237,
      "learning_rate": 1.0852652114395488e-06,
      "loss": 2.6733,
      "step": 2417
    },
    {
      "epoch": 0.31304010098067775,
      "grad_norm": 2.764862726756628,
      "learning_rate": 1.0850740633369965e-06,
      "loss": 2.3457,
      "step": 2418
    },
    {
      "epoch": 0.3131695633880312,
      "grad_norm": 6.20786143692676,
      "learning_rate": 1.0848828212549917e-06,
      "loss": 3.0454,
      "step": 2419
    },
    {
      "epoch": 0.31329902579538466,
      "grad_norm": 2.931228851104307,
      "learning_rate": 1.0846914852325994e-06,
      "loss": 2.9658,
      "step": 2420
    },
    {
      "epoch": 0.31329902579538466,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.730668544769287,
      "eval_runtime": 15.5422,
      "eval_samples_per_second": 2.831,
      "eval_steps_per_second": 0.193,
      "step": 2420
    },
    {
      "epoch": 0.3134284882027381,
      "grad_norm": 4.246217231741376,
      "learning_rate": 1.0845000553089032e-06,
      "loss": 2.8145,
      "step": 2421
    },
    {
      "epoch": 0.3135579506100916,
      "grad_norm": 3.7715143370034263,
      "learning_rate": 1.0843085315230076e-06,
      "loss": 3.0273,
      "step": 2422
    },
    {
      "epoch": 0.31368741301744507,
      "grad_norm": 1.8352538755017078,
      "learning_rate": 1.0841169139140345e-06,
      "loss": 2.696,
      "step": 2423
    },
    {
      "epoch": 0.3138168754247985,
      "grad_norm": 3.951555566734122,
      "learning_rate": 1.0839252025211259e-06,
      "loss": 2.5566,
      "step": 2424
    },
    {
      "epoch": 0.313946337832152,
      "grad_norm": 5.255888607410781,
      "learning_rate": 1.0837333973834423e-06,
      "loss": 2.8184,
      "step": 2425
    },
    {
      "epoch": 0.313946337832152,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7328213453292847,
      "eval_runtime": 13.8188,
      "eval_samples_per_second": 3.184,
      "eval_steps_per_second": 0.217,
      "step": 2425
    },
    {
      "epoch": 0.31407580023950543,
      "grad_norm": 4.27767564354849,
      "learning_rate": 1.0835414985401646e-06,
      "loss": 2.6543,
      "step": 2426
    },
    {
      "epoch": 0.31420526264685894,
      "grad_norm": 2.0119119746602645,
      "learning_rate": 1.0833495060304916e-06,
      "loss": 2.5239,
      "step": 2427
    },
    {
      "epoch": 0.3143347250542124,
      "grad_norm": 2.0564351305921904,
      "learning_rate": 1.0831574198936414e-06,
      "loss": 2.7466,
      "step": 2428
    },
    {
      "epoch": 0.31446418746156585,
      "grad_norm": 3.328244710080324,
      "learning_rate": 1.082965240168852e-06,
      "loss": 2.8818,
      "step": 2429
    },
    {
      "epoch": 0.3145936498689193,
      "grad_norm": 5.202876420862465,
      "learning_rate": 1.08277296689538e-06,
      "loss": 2.9172,
      "step": 2430
    },
    {
      "epoch": 0.3145936498689193,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7305575609207153,
      "eval_runtime": 14.3207,
      "eval_samples_per_second": 3.072,
      "eval_steps_per_second": 0.209,
      "step": 2430
    },
    {
      "epoch": 0.31472311227627275,
      "grad_norm": 2.3954730238284814,
      "learning_rate": 1.0825806001125003e-06,
      "loss": 2.6597,
      "step": 2431
    },
    {
      "epoch": 0.31485257468362626,
      "grad_norm": 2.45988008419959,
      "learning_rate": 1.0823881398595085e-06,
      "loss": 2.7405,
      "step": 2432
    },
    {
      "epoch": 0.3149820370909797,
      "grad_norm": 1.784575661810735,
      "learning_rate": 1.0821955861757185e-06,
      "loss": 2.6685,
      "step": 2433
    },
    {
      "epoch": 0.31511149949833317,
      "grad_norm": 2.581102977328374,
      "learning_rate": 1.082002939100463e-06,
      "loss": 2.7251,
      "step": 2434
    },
    {
      "epoch": 0.3152409619056866,
      "grad_norm": 3.8718914848734465,
      "learning_rate": 1.0818101986730942e-06,
      "loss": 2.7039,
      "step": 2435
    },
    {
      "epoch": 0.3152409619056866,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7306907176971436,
      "eval_runtime": 14.7789,
      "eval_samples_per_second": 2.977,
      "eval_steps_per_second": 0.203,
      "step": 2435
    },
    {
      "epoch": 0.3153704243130401,
      "grad_norm": 2.655205853014172,
      "learning_rate": 1.0816173649329831e-06,
      "loss": 2.6807,
      "step": 2436
    },
    {
      "epoch": 0.3154998867203936,
      "grad_norm": 2.6546314761607563,
      "learning_rate": 1.0814244379195199e-06,
      "loss": 2.7903,
      "step": 2437
    },
    {
      "epoch": 0.31562934912774704,
      "grad_norm": 2.115226926621325,
      "learning_rate": 1.0812314176721141e-06,
      "loss": 2.7529,
      "step": 2438
    },
    {
      "epoch": 0.3157588115351005,
      "grad_norm": 2.36549653962922,
      "learning_rate": 1.0810383042301938e-06,
      "loss": 2.4944,
      "step": 2439
    },
    {
      "epoch": 0.31588827394245395,
      "grad_norm": 3.4032755258886045,
      "learning_rate": 1.0808450976332064e-06,
      "loss": 2.947,
      "step": 2440
    },
    {
      "epoch": 0.31588827394245395,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7308460474014282,
      "eval_runtime": 14.0494,
      "eval_samples_per_second": 3.132,
      "eval_steps_per_second": 0.214,
      "step": 2440
    },
    {
      "epoch": 0.3160177363498074,
      "grad_norm": 3.160928109932279,
      "learning_rate": 1.0806517979206183e-06,
      "loss": 2.6792,
      "step": 2441
    },
    {
      "epoch": 0.3161471987571609,
      "grad_norm": 3.9320513444446266,
      "learning_rate": 1.0804584051319146e-06,
      "loss": 2.8491,
      "step": 2442
    },
    {
      "epoch": 0.31627666116451436,
      "grad_norm": 3.0947554246258244,
      "learning_rate": 1.0802649193066003e-06,
      "loss": 2.7412,
      "step": 2443
    },
    {
      "epoch": 0.3164061235718678,
      "grad_norm": 3.229832254575071,
      "learning_rate": 1.0800713404841982e-06,
      "loss": 2.5986,
      "step": 2444
    },
    {
      "epoch": 0.31653558597922127,
      "grad_norm": 3.2882254814348744,
      "learning_rate": 1.079877668704251e-06,
      "loss": 2.7722,
      "step": 2445
    },
    {
      "epoch": 0.31653558597922127,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7264736890792847,
      "eval_runtime": 14.1039,
      "eval_samples_per_second": 3.12,
      "eval_steps_per_second": 0.213,
      "step": 2445
    },
    {
      "epoch": 0.3166650483865747,
      "grad_norm": 8.512871943307267,
      "learning_rate": 1.07968390400632e-06,
      "loss": 2.9233,
      "step": 2446
    },
    {
      "epoch": 0.31679451079392823,
      "grad_norm": 1.8824532694724163,
      "learning_rate": 1.0794900464299856e-06,
      "loss": 2.5317,
      "step": 2447
    },
    {
      "epoch": 0.3169239732012817,
      "grad_norm": 2.31123638509395,
      "learning_rate": 1.0792960960148472e-06,
      "loss": 2.7749,
      "step": 2448
    },
    {
      "epoch": 0.31705343560863514,
      "grad_norm": 1.9713933039519702,
      "learning_rate": 1.0791020528005228e-06,
      "loss": 2.7087,
      "step": 2449
    },
    {
      "epoch": 0.3171828980159886,
      "grad_norm": 6.951210945555584,
      "learning_rate": 1.07890791682665e-06,
      "loss": 2.9065,
      "step": 2450
    },
    {
      "epoch": 0.3171828980159886,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.725053310394287,
      "eval_runtime": 14.3067,
      "eval_samples_per_second": 3.075,
      "eval_steps_per_second": 0.21,
      "step": 2450
    },
    {
      "epoch": 0.3173123604233421,
      "grad_norm": 1.5610967085634204,
      "learning_rate": 1.078713688132885e-06,
      "loss": 2.6248,
      "step": 2451
    },
    {
      "epoch": 0.31744182283069555,
      "grad_norm": 2.6455580630439464,
      "learning_rate": 1.0785193667589029e-06,
      "loss": 2.7341,
      "step": 2452
    },
    {
      "epoch": 0.317571285238049,
      "grad_norm": 1.6280969815206268,
      "learning_rate": 1.0783249527443973e-06,
      "loss": 2.5791,
      "step": 2453
    },
    {
      "epoch": 0.31770074764540246,
      "grad_norm": 4.436350882045652,
      "learning_rate": 1.0781304461290821e-06,
      "loss": 2.603,
      "step": 2454
    },
    {
      "epoch": 0.3178302100527559,
      "grad_norm": 3.1764506320770693,
      "learning_rate": 1.0779358469526886e-06,
      "loss": 2.8035,
      "step": 2455
    },
    {
      "epoch": 0.3178302100527559,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.722367763519287,
      "eval_runtime": 13.7875,
      "eval_samples_per_second": 3.191,
      "eval_steps_per_second": 0.218,
      "step": 2455
    },
    {
      "epoch": 0.3179596724601094,
      "grad_norm": 2.553147022982123,
      "learning_rate": 1.0777411552549675e-06,
      "loss": 2.7024,
      "step": 2456
    },
    {
      "epoch": 0.3180891348674629,
      "grad_norm": 4.184804910513455,
      "learning_rate": 1.077546371075689e-06,
      "loss": 3.0112,
      "step": 2457
    },
    {
      "epoch": 0.31821859727481633,
      "grad_norm": 2.8942752790324002,
      "learning_rate": 1.0773514944546414e-06,
      "loss": 3.0034,
      "step": 2458
    },
    {
      "epoch": 0.3183480596821698,
      "grad_norm": 2.7530188522035974,
      "learning_rate": 1.0771565254316327e-06,
      "loss": 2.8999,
      "step": 2459
    },
    {
      "epoch": 0.31847752208952324,
      "grad_norm": 2.6219202515606526,
      "learning_rate": 1.0769614640464883e-06,
      "loss": 2.8696,
      "step": 2460
    },
    {
      "epoch": 0.31847752208952324,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7179287672042847,
      "eval_runtime": 14.4399,
      "eval_samples_per_second": 3.047,
      "eval_steps_per_second": 0.208,
      "step": 2460
    },
    {
      "epoch": 0.31860698449687674,
      "grad_norm": 2.84148124067926,
      "learning_rate": 1.0767663103390546e-06,
      "loss": 2.7959,
      "step": 2461
    },
    {
      "epoch": 0.3187364469042302,
      "grad_norm": 2.756660157319629,
      "learning_rate": 1.0765710643491947e-06,
      "loss": 3.083,
      "step": 2462
    },
    {
      "epoch": 0.31886590931158365,
      "grad_norm": 4.201171957519662,
      "learning_rate": 1.0763757261167923e-06,
      "loss": 2.6294,
      "step": 2463
    },
    {
      "epoch": 0.3189953717189371,
      "grad_norm": 4.321268425030079,
      "learning_rate": 1.0761802956817486e-06,
      "loss": 2.6978,
      "step": 2464
    },
    {
      "epoch": 0.31912483412629056,
      "grad_norm": 1.7481726935604223,
      "learning_rate": 1.0759847730839847e-06,
      "loss": 2.865,
      "step": 2465
    },
    {
      "epoch": 0.31912483412629056,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7188388109207153,
      "eval_runtime": 16.101,
      "eval_samples_per_second": 2.733,
      "eval_steps_per_second": 0.186,
      "step": 2465
    },
    {
      "epoch": 0.31925429653364407,
      "grad_norm": 2.9924227571415556,
      "learning_rate": 1.0757891583634398e-06,
      "loss": 2.6526,
      "step": 2466
    },
    {
      "epoch": 0.3193837589409975,
      "grad_norm": 2.5318284906218587,
      "learning_rate": 1.0755934515600721e-06,
      "loss": 2.667,
      "step": 2467
    },
    {
      "epoch": 0.319513221348351,
      "grad_norm": 3.454384497407765,
      "learning_rate": 1.075397652713859e-06,
      "loss": 3.1702,
      "step": 2468
    },
    {
      "epoch": 0.3196426837557044,
      "grad_norm": 3.5162367439322177,
      "learning_rate": 1.0752017618647959e-06,
      "loss": 2.6421,
      "step": 2469
    },
    {
      "epoch": 0.3197721461630579,
      "grad_norm": 3.327768218711653,
      "learning_rate": 1.075005779052898e-06,
      "loss": 2.9227,
      "step": 2470
    },
    {
      "epoch": 0.3197721461630579,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7175514698028564,
      "eval_runtime": 15.8133,
      "eval_samples_per_second": 2.782,
      "eval_steps_per_second": 0.19,
      "step": 2470
    },
    {
      "epoch": 0.3199016085704114,
      "grad_norm": 2.0077234479844415,
      "learning_rate": 1.0748097043181984e-06,
      "loss": 2.7979,
      "step": 2471
    },
    {
      "epoch": 0.32003107097776484,
      "grad_norm": 2.907360607367345,
      "learning_rate": 1.0746135377007496e-06,
      "loss": 2.728,
      "step": 2472
    },
    {
      "epoch": 0.3201605333851183,
      "grad_norm": 6.426058586786103,
      "learning_rate": 1.0744172792406225e-06,
      "loss": 2.6758,
      "step": 2473
    },
    {
      "epoch": 0.32028999579247175,
      "grad_norm": 1.2931018024081558,
      "learning_rate": 1.0742209289779069e-06,
      "loss": 2.449,
      "step": 2474
    },
    {
      "epoch": 0.3204194581998252,
      "grad_norm": 3.1610311723919224,
      "learning_rate": 1.074024486952711e-06,
      "loss": 2.7855,
      "step": 2475
    },
    {
      "epoch": 0.3204194581998252,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.72021484375,
      "eval_runtime": 15.1815,
      "eval_samples_per_second": 2.898,
      "eval_steps_per_second": 0.198,
      "step": 2475
    },
    {
      "epoch": 0.3205489206071787,
      "grad_norm": 2.648317789620508,
      "learning_rate": 1.0738279532051625e-06,
      "loss": 2.8459,
      "step": 2476
    },
    {
      "epoch": 0.32067838301453216,
      "grad_norm": 1.5009025415287334,
      "learning_rate": 1.073631327775407e-06,
      "loss": 2.439,
      "step": 2477
    },
    {
      "epoch": 0.3208078454218856,
      "grad_norm": 1.6609496433491577,
      "learning_rate": 1.0734346107036097e-06,
      "loss": 2.4409,
      "step": 2478
    },
    {
      "epoch": 0.32093730782923907,
      "grad_norm": 2.290237376431311,
      "learning_rate": 1.073237802029954e-06,
      "loss": 2.7837,
      "step": 2479
    },
    {
      "epoch": 0.3210667702365925,
      "grad_norm": 3.1314784364313724,
      "learning_rate": 1.0730409017946414e-06,
      "loss": 2.6189,
      "step": 2480
    },
    {
      "epoch": 0.3210667702365925,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.716020107269287,
      "eval_runtime": 14.5865,
      "eval_samples_per_second": 3.016,
      "eval_steps_per_second": 0.206,
      "step": 2480
    },
    {
      "epoch": 0.32119623264394603,
      "grad_norm": 2.480958092689093,
      "learning_rate": 1.0728439100378935e-06,
      "loss": 2.6064,
      "step": 2481
    },
    {
      "epoch": 0.3213256950512995,
      "grad_norm": 1.485956326766334,
      "learning_rate": 1.0726468267999495e-06,
      "loss": 2.5059,
      "step": 2482
    },
    {
      "epoch": 0.32145515745865294,
      "grad_norm": 3.9416117523117613,
      "learning_rate": 1.072449652121068e-06,
      "loss": 3.1322,
      "step": 2483
    },
    {
      "epoch": 0.3215846198660064,
      "grad_norm": 2.118533380022602,
      "learning_rate": 1.0722523860415251e-06,
      "loss": 2.7395,
      "step": 2484
    },
    {
      "epoch": 0.32171408227335985,
      "grad_norm": 7.2211385103012224,
      "learning_rate": 1.0720550286016174e-06,
      "loss": 3.2327,
      "step": 2485
    },
    {
      "epoch": 0.32171408227335985,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.714555263519287,
      "eval_runtime": 14.4227,
      "eval_samples_per_second": 3.051,
      "eval_steps_per_second": 0.208,
      "step": 2485
    },
    {
      "epoch": 0.32184354468071336,
      "grad_norm": 2.9886870997262838,
      "learning_rate": 1.0718575798416585e-06,
      "loss": 2.5952,
      "step": 2486
    },
    {
      "epoch": 0.3219730070880668,
      "grad_norm": 6.80834374440038,
      "learning_rate": 1.0716600398019818e-06,
      "loss": 2.417,
      "step": 2487
    },
    {
      "epoch": 0.32210246949542026,
      "grad_norm": 6.216963909338738,
      "learning_rate": 1.0714624085229383e-06,
      "loss": 2.6121,
      "step": 2488
    },
    {
      "epoch": 0.3222319319027737,
      "grad_norm": 3.567677258886948,
      "learning_rate": 1.0712646860448985e-06,
      "loss": 3.1226,
      "step": 2489
    },
    {
      "epoch": 0.32236139431012717,
      "grad_norm": 4.148947984312593,
      "learning_rate": 1.0710668724082516e-06,
      "loss": 2.6794,
      "step": 2490
    },
    {
      "epoch": 0.32236139431012717,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7180397510528564,
      "eval_runtime": 14.2145,
      "eval_samples_per_second": 3.095,
      "eval_steps_per_second": 0.211,
      "step": 2490
    },
    {
      "epoch": 0.3224908567174807,
      "grad_norm": 2.409469581957139,
      "learning_rate": 1.0708689676534044e-06,
      "loss": 2.7512,
      "step": 2491
    },
    {
      "epoch": 0.32262031912483413,
      "grad_norm": 1.989219831788488,
      "learning_rate": 1.070670971820783e-06,
      "loss": 2.6553,
      "step": 2492
    },
    {
      "epoch": 0.3227497815321876,
      "grad_norm": 1.2342416306836181,
      "learning_rate": 1.0704728849508324e-06,
      "loss": 2.2351,
      "step": 2493
    },
    {
      "epoch": 0.32287924393954104,
      "grad_norm": 2.6907820017073503,
      "learning_rate": 1.070274707084016e-06,
      "loss": 2.7109,
      "step": 2494
    },
    {
      "epoch": 0.32300870634689455,
      "grad_norm": 3.650577051923038,
      "learning_rate": 1.0700764382608148e-06,
      "loss": 2.9163,
      "step": 2495
    },
    {
      "epoch": 0.32300870634689455,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7165749073028564,
      "eval_runtime": 14.7022,
      "eval_samples_per_second": 2.993,
      "eval_steps_per_second": 0.204,
      "step": 2495
    },
    {
      "epoch": 0.323138168754248,
      "grad_norm": 6.1785950075176235,
      "learning_rate": 1.0698780785217302e-06,
      "loss": 3.3462,
      "step": 2496
    },
    {
      "epoch": 0.32326763116160145,
      "grad_norm": 2.5343607882108694,
      "learning_rate": 1.0696796279072807e-06,
      "loss": 2.4346,
      "step": 2497
    },
    {
      "epoch": 0.3233970935689549,
      "grad_norm": 1.846750528713725,
      "learning_rate": 1.0694810864580041e-06,
      "loss": 2.501,
      "step": 2498
    },
    {
      "epoch": 0.32352655597630836,
      "grad_norm": 2.0810508054385037,
      "learning_rate": 1.0692824542144562e-06,
      "loss": 2.564,
      "step": 2499
    },
    {
      "epoch": 0.32365601838366187,
      "grad_norm": 2.5691675830070513,
      "learning_rate": 1.0690837312172117e-06,
      "loss": 2.5383,
      "step": 2500
    },
    {
      "epoch": 0.32365601838366187,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7138671875,
      "eval_runtime": 14.264,
      "eval_samples_per_second": 3.085,
      "eval_steps_per_second": 0.21,
      "step": 2500
    },
    {
      "epoch": 0.3237854807910153,
      "grad_norm": 2.481378628997151,
      "learning_rate": 1.0688849175068642e-06,
      "loss": 2.6182,
      "step": 2501
    },
    {
      "epoch": 0.3239149431983688,
      "grad_norm": 2.1029337448683956,
      "learning_rate": 1.0686860131240251e-06,
      "loss": 2.7554,
      "step": 2502
    },
    {
      "epoch": 0.32404440560572223,
      "grad_norm": 9.161150601212466,
      "learning_rate": 1.0684870181093246e-06,
      "loss": 3.8335,
      "step": 2503
    },
    {
      "epoch": 0.3241738680130757,
      "grad_norm": 4.310980258284213,
      "learning_rate": 1.0682879325034119e-06,
      "loss": 2.9722,
      "step": 2504
    },
    {
      "epoch": 0.3243033304204292,
      "grad_norm": 7.151629142739274,
      "learning_rate": 1.0680887563469537e-06,
      "loss": 2.79,
      "step": 2505
    },
    {
      "epoch": 0.3243033304204292,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7144886255264282,
      "eval_runtime": 21.4504,
      "eval_samples_per_second": 2.051,
      "eval_steps_per_second": 0.14,
      "step": 2505
    },
    {
      "epoch": 0.32443279282778265,
      "grad_norm": 2.2164747783602596,
      "learning_rate": 1.0678894896806357e-06,
      "loss": 2.6233,
      "step": 2506
    },
    {
      "epoch": 0.3245622552351361,
      "grad_norm": 1.7307157437209966,
      "learning_rate": 1.0676901325451628e-06,
      "loss": 2.7095,
      "step": 2507
    },
    {
      "epoch": 0.32469171764248955,
      "grad_norm": 2.823082635305362,
      "learning_rate": 1.0674906849812574e-06,
      "loss": 2.8804,
      "step": 2508
    },
    {
      "epoch": 0.324821180049843,
      "grad_norm": 3.657282187888237,
      "learning_rate": 1.0672911470296609e-06,
      "loss": 2.8359,
      "step": 2509
    },
    {
      "epoch": 0.3249506424571965,
      "grad_norm": 2.0672970379658873,
      "learning_rate": 1.0670915187311326e-06,
      "loss": 2.8318,
      "step": 2510
    },
    {
      "epoch": 0.3249506424571965,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7151988744735718,
      "eval_runtime": 14.2028,
      "eval_samples_per_second": 3.098,
      "eval_steps_per_second": 0.211,
      "step": 2510
    },
    {
      "epoch": 0.32508010486454997,
      "grad_norm": 2.505969159834915,
      "learning_rate": 1.0668918001264508e-06,
      "loss": 3.0093,
      "step": 2511
    },
    {
      "epoch": 0.3252095672719034,
      "grad_norm": 3.286925230960295,
      "learning_rate": 1.0666919912564119e-06,
      "loss": 2.8894,
      "step": 2512
    },
    {
      "epoch": 0.3253390296792569,
      "grad_norm": 4.099638565541644,
      "learning_rate": 1.0664920921618313e-06,
      "loss": 2.9404,
      "step": 2513
    },
    {
      "epoch": 0.32546849208661033,
      "grad_norm": 1.760262411384234,
      "learning_rate": 1.0662921028835424e-06,
      "loss": 2.5403,
      "step": 2514
    },
    {
      "epoch": 0.32559795449396384,
      "grad_norm": 3.8364946154879727,
      "learning_rate": 1.0660920234623968e-06,
      "loss": 2.8704,
      "step": 2515
    },
    {
      "epoch": 0.32559795449396384,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.714399814605713,
      "eval_runtime": 14.7459,
      "eval_samples_per_second": 2.984,
      "eval_steps_per_second": 0.203,
      "step": 2515
    },
    {
      "epoch": 0.3257274169013173,
      "grad_norm": 2.6276134600347665,
      "learning_rate": 1.0658918539392645e-06,
      "loss": 2.5798,
      "step": 2516
    },
    {
      "epoch": 0.32585687930867074,
      "grad_norm": 1.735659873738537,
      "learning_rate": 1.065691594355035e-06,
      "loss": 2.611,
      "step": 2517
    },
    {
      "epoch": 0.3259863417160242,
      "grad_norm": 2.1113989104365047,
      "learning_rate": 1.065491244750615e-06,
      "loss": 2.8362,
      "step": 2518
    },
    {
      "epoch": 0.32611580412337765,
      "grad_norm": 2.452792140237132,
      "learning_rate": 1.0652908051669296e-06,
      "loss": 2.4827,
      "step": 2519
    },
    {
      "epoch": 0.32624526653073116,
      "grad_norm": 2.3767173997955653,
      "learning_rate": 1.065090275644923e-06,
      "loss": 2.5935,
      "step": 2520
    },
    {
      "epoch": 0.32624526653073116,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7179954051971436,
      "eval_runtime": 14.1747,
      "eval_samples_per_second": 3.104,
      "eval_steps_per_second": 0.212,
      "step": 2520
    },
    {
      "epoch": 0.3263747289380846,
      "grad_norm": 2.0999032594110827,
      "learning_rate": 1.0648896562255576e-06,
      "loss": 2.894,
      "step": 2521
    },
    {
      "epoch": 0.32650419134543807,
      "grad_norm": 1.5953578490442484,
      "learning_rate": 1.0646889469498137e-06,
      "loss": 2.8232,
      "step": 2522
    },
    {
      "epoch": 0.3266336537527915,
      "grad_norm": 2.4790876234481383,
      "learning_rate": 1.06448814785869e-06,
      "loss": 2.6841,
      "step": 2523
    },
    {
      "epoch": 0.326763116160145,
      "grad_norm": 3.859717858383115,
      "learning_rate": 1.0642872589932044e-06,
      "loss": 2.6362,
      "step": 2524
    },
    {
      "epoch": 0.3268925785674985,
      "grad_norm": 2.2439064342288795,
      "learning_rate": 1.0640862803943922e-06,
      "loss": 2.8138,
      "step": 2525
    },
    {
      "epoch": 0.3268925785674985,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7114924192428589,
      "eval_runtime": 14.7113,
      "eval_samples_per_second": 2.991,
      "eval_steps_per_second": 0.204,
      "step": 2525
    },
    {
      "epoch": 0.32702204097485194,
      "grad_norm": 2.9644071113280783,
      "learning_rate": 1.0638852121033071e-06,
      "loss": 2.8468,
      "step": 2526
    },
    {
      "epoch": 0.3271515033822054,
      "grad_norm": 4.798662640337993,
      "learning_rate": 1.0636840541610216e-06,
      "loss": 2.8193,
      "step": 2527
    },
    {
      "epoch": 0.32728096578955884,
      "grad_norm": 1.8512132346857104,
      "learning_rate": 1.0634828066086266e-06,
      "loss": 2.5137,
      "step": 2528
    },
    {
      "epoch": 0.3274104281969123,
      "grad_norm": 2.43011310685653,
      "learning_rate": 1.0632814694872306e-06,
      "loss": 2.5033,
      "step": 2529
    },
    {
      "epoch": 0.3275398906042658,
      "grad_norm": 2.1162293285413107,
      "learning_rate": 1.0630800428379609e-06,
      "loss": 2.7549,
      "step": 2530
    },
    {
      "epoch": 0.3275398906042658,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7129794359207153,
      "eval_runtime": 14.2024,
      "eval_samples_per_second": 3.098,
      "eval_steps_per_second": 0.211,
      "step": 2530
    },
    {
      "epoch": 0.32766935301161926,
      "grad_norm": 2.866204210153781,
      "learning_rate": 1.0628785267019631e-06,
      "loss": 2.8296,
      "step": 2531
    },
    {
      "epoch": 0.3277988154189727,
      "grad_norm": 7.048963034175104,
      "learning_rate": 1.0626769211204007e-06,
      "loss": 3.2961,
      "step": 2532
    },
    {
      "epoch": 0.32792827782632616,
      "grad_norm": 3.116704526548702,
      "learning_rate": 1.062475226134456e-06,
      "loss": 2.7109,
      "step": 2533
    },
    {
      "epoch": 0.3280577402336796,
      "grad_norm": 2.945880722895081,
      "learning_rate": 1.062273441785329e-06,
      "loss": 2.7776,
      "step": 2534
    },
    {
      "epoch": 0.3281872026410331,
      "grad_norm": 4.774742623417359,
      "learning_rate": 1.0620715681142382e-06,
      "loss": 3.3157,
      "step": 2535
    },
    {
      "epoch": 0.3281872026410331,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7116920948028564,
      "eval_runtime": 14.274,
      "eval_samples_per_second": 3.083,
      "eval_steps_per_second": 0.21,
      "step": 2535
    },
    {
      "epoch": 0.3283166650483866,
      "grad_norm": 12.183511694896266,
      "learning_rate": 1.0618696051624208e-06,
      "loss": 3.6406,
      "step": 2536
    },
    {
      "epoch": 0.32844612745574003,
      "grad_norm": 1.8943073734233922,
      "learning_rate": 1.0616675529711317e-06,
      "loss": 2.5791,
      "step": 2537
    },
    {
      "epoch": 0.3285755898630935,
      "grad_norm": 2.6902875372080386,
      "learning_rate": 1.0614654115816439e-06,
      "loss": 2.7078,
      "step": 2538
    },
    {
      "epoch": 0.328705052270447,
      "grad_norm": 3.8938437205833343,
      "learning_rate": 1.0612631810352491e-06,
      "loss": 2.7634,
      "step": 2539
    },
    {
      "epoch": 0.32883451467780045,
      "grad_norm": 4.6834832274348885,
      "learning_rate": 1.0610608613732568e-06,
      "loss": 2.5994,
      "step": 2540
    },
    {
      "epoch": 0.32883451467780045,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7151100635528564,
      "eval_runtime": 14.4515,
      "eval_samples_per_second": 3.045,
      "eval_steps_per_second": 0.208,
      "step": 2540
    },
    {
      "epoch": 0.3289639770851539,
      "grad_norm": 1.8705268773501096,
      "learning_rate": 1.060858452636995e-06,
      "loss": 2.6608,
      "step": 2541
    },
    {
      "epoch": 0.32909343949250736,
      "grad_norm": 1.5657623022903622,
      "learning_rate": 1.06065595486781e-06,
      "loss": 2.5117,
      "step": 2542
    },
    {
      "epoch": 0.3292229018998608,
      "grad_norm": 1.9806770784366814,
      "learning_rate": 1.0604533681070657e-06,
      "loss": 2.6636,
      "step": 2543
    },
    {
      "epoch": 0.3293523643072143,
      "grad_norm": 3.3274996869229763,
      "learning_rate": 1.0602506923961447e-06,
      "loss": 2.7339,
      "step": 2544
    },
    {
      "epoch": 0.32948182671456777,
      "grad_norm": 2.2561489568209305,
      "learning_rate": 1.060047927776448e-06,
      "loss": 2.7913,
      "step": 2545
    },
    {
      "epoch": 0.32948182671456777,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7157981395721436,
      "eval_runtime": 14.7355,
      "eval_samples_per_second": 2.986,
      "eval_steps_per_second": 0.204,
      "step": 2545
    },
    {
      "epoch": 0.3296112891219212,
      "grad_norm": 1.887133622590239,
      "learning_rate": 1.0598450742893937e-06,
      "loss": 2.8413,
      "step": 2546
    },
    {
      "epoch": 0.3297407515292747,
      "grad_norm": 1.6186431866264301,
      "learning_rate": 1.0596421319764192e-06,
      "loss": 2.5122,
      "step": 2547
    },
    {
      "epoch": 0.32987021393662813,
      "grad_norm": 3.9729680448090408,
      "learning_rate": 1.0594391008789791e-06,
      "loss": 3.2952,
      "step": 2548
    },
    {
      "epoch": 0.32999967634398164,
      "grad_norm": 5.515309859946067,
      "learning_rate": 1.0592359810385475e-06,
      "loss": 3.3301,
      "step": 2549
    },
    {
      "epoch": 0.3301291387513351,
      "grad_norm": 6.131034232086031,
      "learning_rate": 1.059032772496615e-06,
      "loss": 2.9854,
      "step": 2550
    },
    {
      "epoch": 0.3301291387513351,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7115145921707153,
      "eval_runtime": 13.4007,
      "eval_samples_per_second": 3.283,
      "eval_steps_per_second": 0.224,
      "step": 2550
    },
    {
      "epoch": 0.33025860115868855,
      "grad_norm": 3.772467972650206,
      "learning_rate": 1.0588294752946912e-06,
      "loss": 2.8049,
      "step": 2551
    },
    {
      "epoch": 0.330388063566042,
      "grad_norm": 2.261533079776973,
      "learning_rate": 1.0586260894743037e-06,
      "loss": 2.6274,
      "step": 2552
    },
    {
      "epoch": 0.33051752597339545,
      "grad_norm": 1.819366922636373,
      "learning_rate": 1.0584226150769983e-06,
      "loss": 2.5884,
      "step": 2553
    },
    {
      "epoch": 0.33064698838074896,
      "grad_norm": 2.074689928427377,
      "learning_rate": 1.0582190521443388e-06,
      "loss": 2.3972,
      "step": 2554
    },
    {
      "epoch": 0.3307764507881024,
      "grad_norm": 3.3102970858262384,
      "learning_rate": 1.0580154007179068e-06,
      "loss": 2.8782,
      "step": 2555
    },
    {
      "epoch": 0.3307764507881024,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7094504833221436,
      "eval_runtime": 13.644,
      "eval_samples_per_second": 3.225,
      "eval_steps_per_second": 0.22,
      "step": 2555
    },
    {
      "epoch": 0.33090591319545587,
      "grad_norm": 2.891603549055289,
      "learning_rate": 1.0578116608393025e-06,
      "loss": 2.5237,
      "step": 2556
    },
    {
      "epoch": 0.3310353756028093,
      "grad_norm": 5.775247090996414,
      "learning_rate": 1.0576078325501437e-06,
      "loss": 2.7983,
      "step": 2557
    },
    {
      "epoch": 0.3311648380101628,
      "grad_norm": 2.594827979279055,
      "learning_rate": 1.0574039158920665e-06,
      "loss": 2.6567,
      "step": 2558
    },
    {
      "epoch": 0.3312943004175163,
      "grad_norm": 4.2059805640075405,
      "learning_rate": 1.0571999109067252e-06,
      "loss": 2.7373,
      "step": 2559
    },
    {
      "epoch": 0.33142376282486974,
      "grad_norm": 2.194684530127607,
      "learning_rate": 1.0569958176357916e-06,
      "loss": 2.45,
      "step": 2560
    },
    {
      "epoch": 0.33142376282486974,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7145329713821411,
      "eval_runtime": 15.2063,
      "eval_samples_per_second": 2.894,
      "eval_steps_per_second": 0.197,
      "step": 2560
    },
    {
      "epoch": 0.3315532252322232,
      "grad_norm": 2.0551486877861813,
      "learning_rate": 1.0567916361209562e-06,
      "loss": 2.9585,
      "step": 2561
    },
    {
      "epoch": 0.33168268763957665,
      "grad_norm": 3.0576880647182128,
      "learning_rate": 1.0565873664039272e-06,
      "loss": 2.915,
      "step": 2562
    },
    {
      "epoch": 0.3318121500469301,
      "grad_norm": 1.6618363833424774,
      "learning_rate": 1.0563830085264307e-06,
      "loss": 2.491,
      "step": 2563
    },
    {
      "epoch": 0.3319416124542836,
      "grad_norm": 2.242516572982805,
      "learning_rate": 1.056178562530211e-06,
      "loss": 2.5446,
      "step": 2564
    },
    {
      "epoch": 0.33207107486163706,
      "grad_norm": 1.6886334607956923,
      "learning_rate": 1.0559740284570301e-06,
      "loss": 2.5901,
      "step": 2565
    },
    {
      "epoch": 0.33207107486163706,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7080078125,
      "eval_runtime": 14.1946,
      "eval_samples_per_second": 3.1,
      "eval_steps_per_second": 0.211,
      "step": 2565
    },
    {
      "epoch": 0.3322005372689905,
      "grad_norm": 2.9808641489517624,
      "learning_rate": 1.0557694063486685e-06,
      "loss": 2.8701,
      "step": 2566
    },
    {
      "epoch": 0.33232999967634397,
      "grad_norm": 1.7861623716182986,
      "learning_rate": 1.0555646962469247e-06,
      "loss": 2.667,
      "step": 2567
    },
    {
      "epoch": 0.3324594620836974,
      "grad_norm": 2.60652426416351,
      "learning_rate": 1.0553598981936143e-06,
      "loss": 2.6201,
      "step": 2568
    },
    {
      "epoch": 0.33258892449105093,
      "grad_norm": 1.5322282912622005,
      "learning_rate": 1.0551550122305714e-06,
      "loss": 2.4189,
      "step": 2569
    },
    {
      "epoch": 0.3327183868984044,
      "grad_norm": 2.2602851489110876,
      "learning_rate": 1.0549500383996487e-06,
      "loss": 3.0356,
      "step": 2570
    },
    {
      "epoch": 0.3327183868984044,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7066761255264282,
      "eval_runtime": 14.9728,
      "eval_samples_per_second": 2.939,
      "eval_steps_per_second": 0.2,
      "step": 2570
    },
    {
      "epoch": 0.33284784930575784,
      "grad_norm": 3.201181629676007,
      "learning_rate": 1.054744976742716e-06,
      "loss": 3.0901,
      "step": 2571
    },
    {
      "epoch": 0.3329773117131113,
      "grad_norm": 1.8098001825815357,
      "learning_rate": 1.0545398273016612e-06,
      "loss": 2.6633,
      "step": 2572
    },
    {
      "epoch": 0.33310677412046474,
      "grad_norm": 1.60975394734999,
      "learning_rate": 1.0543345901183903e-06,
      "loss": 2.5728,
      "step": 2573
    },
    {
      "epoch": 0.33323623652781825,
      "grad_norm": 2.216419063714736,
      "learning_rate": 1.0541292652348273e-06,
      "loss": 2.7195,
      "step": 2574
    },
    {
      "epoch": 0.3333656989351717,
      "grad_norm": 2.299470226558847,
      "learning_rate": 1.0539238526929135e-06,
      "loss": 2.4348,
      "step": 2575
    },
    {
      "epoch": 0.3333656989351717,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7066539525985718,
      "eval_runtime": 15.2649,
      "eval_samples_per_second": 2.882,
      "eval_steps_per_second": 0.197,
      "step": 2575
    },
    {
      "epoch": 0.33349516134252516,
      "grad_norm": 1.8527712121918218,
      "learning_rate": 1.0537183525346093e-06,
      "loss": 2.5061,
      "step": 2576
    },
    {
      "epoch": 0.3336246237498786,
      "grad_norm": 2.3838701760593164,
      "learning_rate": 1.0535127648018915e-06,
      "loss": 2.5959,
      "step": 2577
    },
    {
      "epoch": 0.33375408615723207,
      "grad_norm": 2.242244066746959,
      "learning_rate": 1.0533070895367562e-06,
      "loss": 2.4413,
      "step": 2578
    },
    {
      "epoch": 0.3338835485645856,
      "grad_norm": 3.104114420841631,
      "learning_rate": 1.0531013267812162e-06,
      "loss": 2.6984,
      "step": 2579
    },
    {
      "epoch": 0.33401301097193903,
      "grad_norm": 2.3356753120654736,
      "learning_rate": 1.0528954765773032e-06,
      "loss": 2.4277,
      "step": 2580
    },
    {
      "epoch": 0.33401301097193903,
      "eval_PRM Accuracy": 0.22641509433962265,
      "eval_PRM F1": 0.023809523809523808,
      "eval_PRM F1 AUC": 0.5060240963855421,
      "eval_PRM F1 Neg": 0.359375,
      "eval_PRM NPV": 0.21904761904761905,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.012048192771084338,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7060546875,
      "eval_runtime": 14.541,
      "eval_samples_per_second": 3.026,
      "eval_steps_per_second": 0.206,
      "step": 2580
    },
    {
      "epoch": 0.3341424733792925,
      "grad_norm": 1.7649654060736761,
      "learning_rate": 1.052689538967066e-06,
      "loss": 2.6802,
      "step": 2581
    },
    {
      "epoch": 0.33427193578664594,
      "grad_norm": 7.250748495019686,
      "learning_rate": 1.0524835139925715e-06,
      "loss": 3.5027,
      "step": 2582
    },
    {
      "epoch": 0.33440139819399944,
      "grad_norm": 3.818864789122838,
      "learning_rate": 1.0522774016959048e-06,
      "loss": 2.6765,
      "step": 2583
    },
    {
      "epoch": 0.3345308606013529,
      "grad_norm": 2.1650385201773656,
      "learning_rate": 1.0520712021191682e-06,
      "loss": 2.4958,
      "step": 2584
    },
    {
      "epoch": 0.33466032300870635,
      "grad_norm": 1.5332268958923907,
      "learning_rate": 1.0518649153044822e-06,
      "loss": 2.5698,
      "step": 2585
    },
    {
      "epoch": 0.33466032300870635,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.700972080230713,
      "eval_runtime": 19.4365,
      "eval_samples_per_second": 2.264,
      "eval_steps_per_second": 0.154,
      "step": 2585
    },
    {
      "epoch": 0.3347897854160598,
      "grad_norm": 5.596014398610464,
      "learning_rate": 1.0516585412939851e-06,
      "loss": 2.7388,
      "step": 2586
    },
    {
      "epoch": 0.33491924782341326,
      "grad_norm": 4.790651496509038,
      "learning_rate": 1.0514520801298328e-06,
      "loss": 2.5227,
      "step": 2587
    },
    {
      "epoch": 0.33504871023076677,
      "grad_norm": 1.5280812110782658,
      "learning_rate": 1.0512455318541996e-06,
      "loss": 2.4468,
      "step": 2588
    },
    {
      "epoch": 0.3351781726381202,
      "grad_norm": 1.8383410894966703,
      "learning_rate": 1.0510388965092767e-06,
      "loss": 2.7007,
      "step": 2589
    },
    {
      "epoch": 0.3353076350454737,
      "grad_norm": 3.15220381997441,
      "learning_rate": 1.0508321741372738e-06,
      "loss": 2.447,
      "step": 2590
    },
    {
      "epoch": 0.3353076350454737,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7074307203292847,
      "eval_runtime": 15.0432,
      "eval_samples_per_second": 2.925,
      "eval_steps_per_second": 0.199,
      "step": 2590
    },
    {
      "epoch": 0.3354370974528271,
      "grad_norm": 8.649867462195461,
      "learning_rate": 1.0506253647804183e-06,
      "loss": 3.1772,
      "step": 2591
    },
    {
      "epoch": 0.3355665598601806,
      "grad_norm": 3.5187212252695512,
      "learning_rate": 1.0504184684809548e-06,
      "loss": 2.8203,
      "step": 2592
    },
    {
      "epoch": 0.3356960222675341,
      "grad_norm": 3.517012507058948,
      "learning_rate": 1.0502114852811463e-06,
      "loss": 2.6792,
      "step": 2593
    },
    {
      "epoch": 0.33582548467488754,
      "grad_norm": 1.8325109205582257,
      "learning_rate": 1.0500044152232734e-06,
      "loss": 2.4644,
      "step": 2594
    },
    {
      "epoch": 0.335954947082241,
      "grad_norm": 8.076194111303268,
      "learning_rate": 1.0497972583496341e-06,
      "loss": 2.7915,
      "step": 2595
    },
    {
      "epoch": 0.335954947082241,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6988636255264282,
      "eval_runtime": 14.1777,
      "eval_samples_per_second": 3.103,
      "eval_steps_per_second": 0.212,
      "step": 2595
    },
    {
      "epoch": 0.33608440948959445,
      "grad_norm": 5.061235217847475,
      "learning_rate": 1.0495900147025449e-06,
      "loss": 2.624,
      "step": 2596
    },
    {
      "epoch": 0.3362138718969479,
      "grad_norm": 2.3499978010332083,
      "learning_rate": 1.0493826843243388e-06,
      "loss": 2.4851,
      "step": 2597
    },
    {
      "epoch": 0.3363433343043014,
      "grad_norm": 1.8463737649360012,
      "learning_rate": 1.0491752672573676e-06,
      "loss": 2.8057,
      "step": 2598
    },
    {
      "epoch": 0.33647279671165486,
      "grad_norm": 3.100055065218298,
      "learning_rate": 1.0489677635440004e-06,
      "loss": 2.8335,
      "step": 2599
    },
    {
      "epoch": 0.3366022591190083,
      "grad_norm": 2.8126375049506955,
      "learning_rate": 1.0487601732266242e-06,
      "loss": 2.9911,
      "step": 2600
    },
    {
      "epoch": 0.3366022591190083,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.701615810394287,
      "eval_runtime": 14.9005,
      "eval_samples_per_second": 2.953,
      "eval_steps_per_second": 0.201,
      "step": 2600
    },
    {
      "epoch": 0.33673172152636177,
      "grad_norm": 2.6393499229911552,
      "learning_rate": 1.0485524963476433e-06,
      "loss": 2.7493,
      "step": 2601
    },
    {
      "epoch": 0.3368611839337152,
      "grad_norm": 2.996586152663859,
      "learning_rate": 1.04834473294948e-06,
      "loss": 2.4272,
      "step": 2602
    },
    {
      "epoch": 0.33699064634106873,
      "grad_norm": 3.303399548666325,
      "learning_rate": 1.0481368830745743e-06,
      "loss": 2.7395,
      "step": 2603
    },
    {
      "epoch": 0.3371201087484222,
      "grad_norm": 4.107136549027995,
      "learning_rate": 1.0479289467653836e-06,
      "loss": 2.9109,
      "step": 2604
    },
    {
      "epoch": 0.33724957115577564,
      "grad_norm": 2.8623640167485,
      "learning_rate": 1.047720924064383e-06,
      "loss": 2.6252,
      "step": 2605
    },
    {
      "epoch": 0.33724957115577564,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.700883388519287,
      "eval_runtime": 14.4136,
      "eval_samples_per_second": 3.053,
      "eval_steps_per_second": 0.208,
      "step": 2605
    },
    {
      "epoch": 0.3373790335631291,
      "grad_norm": 3.1138147749905123,
      "learning_rate": 1.0475128150140658e-06,
      "loss": 2.4238,
      "step": 2606
    },
    {
      "epoch": 0.33750849597048255,
      "grad_norm": 8.823845814907507,
      "learning_rate": 1.0473046196569419e-06,
      "loss": 3.0791,
      "step": 2607
    },
    {
      "epoch": 0.33763795837783606,
      "grad_norm": 2.3002487856353655,
      "learning_rate": 1.0470963380355398e-06,
      "loss": 2.5059,
      "step": 2608
    },
    {
      "epoch": 0.3377674207851895,
      "grad_norm": 2.062908385782388,
      "learning_rate": 1.046887970192405e-06,
      "loss": 2.7317,
      "step": 2609
    },
    {
      "epoch": 0.33789688319254296,
      "grad_norm": 2.5379343788825115,
      "learning_rate": 1.0466795161701013e-06,
      "loss": 2.543,
      "step": 2610
    },
    {
      "epoch": 0.33789688319254296,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6987971067428589,
      "eval_runtime": 13.8812,
      "eval_samples_per_second": 3.17,
      "eval_steps_per_second": 0.216,
      "step": 2610
    },
    {
      "epoch": 0.3380263455998964,
      "grad_norm": 4.271675896696808,
      "learning_rate": 1.0464709760112092e-06,
      "loss": 3.1885,
      "step": 2611
    },
    {
      "epoch": 0.33815580800724987,
      "grad_norm": 2.9733542401559023,
      "learning_rate": 1.0462623497583274e-06,
      "loss": 2.5793,
      "step": 2612
    },
    {
      "epoch": 0.3382852704146034,
      "grad_norm": 2.5953257703784316,
      "learning_rate": 1.046053637454072e-06,
      "loss": 2.7405,
      "step": 2613
    },
    {
      "epoch": 0.33841473282195683,
      "grad_norm": 2.6447721112500133,
      "learning_rate": 1.045844839141077e-06,
      "loss": 2.46,
      "step": 2614
    },
    {
      "epoch": 0.3385441952293103,
      "grad_norm": 3.6974015547477017,
      "learning_rate": 1.0456359548619934e-06,
      "loss": 2.6494,
      "step": 2615
    },
    {
      "epoch": 0.3385441952293103,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7003728151321411,
      "eval_runtime": 14.3463,
      "eval_samples_per_second": 3.067,
      "eval_steps_per_second": 0.209,
      "step": 2615
    },
    {
      "epoch": 0.33867365763666374,
      "grad_norm": 5.040202116450052,
      "learning_rate": 1.04542698465949e-06,
      "loss": 2.6838,
      "step": 2616
    },
    {
      "epoch": 0.3388031200440172,
      "grad_norm": 1.5796048960942,
      "learning_rate": 1.0452179285762534e-06,
      "loss": 2.4478,
      "step": 2617
    },
    {
      "epoch": 0.3389325824513707,
      "grad_norm": 2.466287382301532,
      "learning_rate": 1.0450087866549876e-06,
      "loss": 2.8816,
      "step": 2618
    },
    {
      "epoch": 0.33906204485872415,
      "grad_norm": 2.9081834264687063,
      "learning_rate": 1.0447995589384136e-06,
      "loss": 2.6187,
      "step": 2619
    },
    {
      "epoch": 0.3391915072660776,
      "grad_norm": 5.112115427668798,
      "learning_rate": 1.044590245469271e-06,
      "loss": 3.0159,
      "step": 2620
    },
    {
      "epoch": 0.3391915072660776,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6988192796707153,
      "eval_runtime": 14.1155,
      "eval_samples_per_second": 3.117,
      "eval_steps_per_second": 0.213,
      "step": 2620
    },
    {
      "epoch": 0.33932096967343106,
      "grad_norm": 4.156557066506399,
      "learning_rate": 1.044380846290316e-06,
      "loss": 2.3723,
      "step": 2621
    },
    {
      "epoch": 0.33945043208078457,
      "grad_norm": 3.0978968693030535,
      "learning_rate": 1.0441713614443226e-06,
      "loss": 2.8125,
      "step": 2622
    },
    {
      "epoch": 0.339579894488138,
      "grad_norm": 4.5699365790234525,
      "learning_rate": 1.0439617909740826e-06,
      "loss": 2.9401,
      "step": 2623
    },
    {
      "epoch": 0.3397093568954915,
      "grad_norm": 2.4451724273045627,
      "learning_rate": 1.0437521349224048e-06,
      "loss": 2.5295,
      "step": 2624
    },
    {
      "epoch": 0.33983881930284493,
      "grad_norm": 3.5630307696173555,
      "learning_rate": 1.0435423933321157e-06,
      "loss": 2.6938,
      "step": 2625
    },
    {
      "epoch": 0.33983881930284493,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7024592161178589,
      "eval_runtime": 16.2431,
      "eval_samples_per_second": 2.709,
      "eval_steps_per_second": 0.185,
      "step": 2625
    },
    {
      "epoch": 0.3399682817101984,
      "grad_norm": 1.6036536088956528,
      "learning_rate": 1.043332566246059e-06,
      "loss": 2.6143,
      "step": 2626
    },
    {
      "epoch": 0.3400977441175519,
      "grad_norm": 1.2505945819644242,
      "learning_rate": 1.0431226537070967e-06,
      "loss": 2.428,
      "step": 2627
    },
    {
      "epoch": 0.34022720652490535,
      "grad_norm": 2.845790070062003,
      "learning_rate": 1.0429126557581073e-06,
      "loss": 3.0803,
      "step": 2628
    },
    {
      "epoch": 0.3403566689322588,
      "grad_norm": 7.691437605773018,
      "learning_rate": 1.0427025724419873e-06,
      "loss": 2.6973,
      "step": 2629
    },
    {
      "epoch": 0.34048613133961225,
      "grad_norm": 4.205271923614831,
      "learning_rate": 1.0424924038016505e-06,
      "loss": 3.1848,
      "step": 2630
    },
    {
      "epoch": 0.34048613133961225,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6981533765792847,
      "eval_runtime": 19.5639,
      "eval_samples_per_second": 2.249,
      "eval_steps_per_second": 0.153,
      "step": 2630
    },
    {
      "epoch": 0.3406155937469657,
      "grad_norm": 2.693230301642108,
      "learning_rate": 1.0422821498800282e-06,
      "loss": 2.4402,
      "step": 2631
    },
    {
      "epoch": 0.3407450561543192,
      "grad_norm": 3.529370406372305,
      "learning_rate": 1.0420718107200686e-06,
      "loss": 2.7778,
      "step": 2632
    },
    {
      "epoch": 0.34087451856167267,
      "grad_norm": 1.6320853269399986,
      "learning_rate": 1.041861386364738e-06,
      "loss": 2.4167,
      "step": 2633
    },
    {
      "epoch": 0.3410039809690261,
      "grad_norm": 2.287187177645135,
      "learning_rate": 1.04165087685702e-06,
      "loss": 2.365,
      "step": 2634
    },
    {
      "epoch": 0.3411334433763796,
      "grad_norm": 1.9301629783085759,
      "learning_rate": 1.041440282239915e-06,
      "loss": 2.6235,
      "step": 2635
    },
    {
      "epoch": 0.3411334433763796,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.698774814605713,
      "eval_runtime": 14.1817,
      "eval_samples_per_second": 3.103,
      "eval_steps_per_second": 0.212,
      "step": 2635
    },
    {
      "epoch": 0.34126290578373303,
      "grad_norm": 1.9918747611047822,
      "learning_rate": 1.0412296025564417e-06,
      "loss": 2.8333,
      "step": 2636
    },
    {
      "epoch": 0.34139236819108654,
      "grad_norm": 6.117318182738866,
      "learning_rate": 1.0410188378496356e-06,
      "loss": 3.2139,
      "step": 2637
    },
    {
      "epoch": 0.34152183059844,
      "grad_norm": 3.3724996120119086,
      "learning_rate": 1.0408079881625495e-06,
      "loss": 2.6978,
      "step": 2638
    },
    {
      "epoch": 0.34165129300579344,
      "grad_norm": 2.436548541265717,
      "learning_rate": 1.0405970535382535e-06,
      "loss": 2.7361,
      "step": 2639
    },
    {
      "epoch": 0.3417807554131469,
      "grad_norm": 1.7125169819490533,
      "learning_rate": 1.0403860340198358e-06,
      "loss": 2.552,
      "step": 2640
    },
    {
      "epoch": 0.3417807554131469,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.7010830640792847,
      "eval_runtime": 14.3574,
      "eval_samples_per_second": 3.065,
      "eval_steps_per_second": 0.209,
      "step": 2640
    },
    {
      "epoch": 0.34191021782050035,
      "grad_norm": 6.122470218936008,
      "learning_rate": 1.040174929650401e-06,
      "loss": 2.8462,
      "step": 2641
    },
    {
      "epoch": 0.34203968022785386,
      "grad_norm": 3.619371498965123,
      "learning_rate": 1.0399637404730717e-06,
      "loss": 2.6294,
      "step": 2642
    },
    {
      "epoch": 0.3421691426352073,
      "grad_norm": 1.3648947885806098,
      "learning_rate": 1.0397524665309875e-06,
      "loss": 2.304,
      "step": 2643
    },
    {
      "epoch": 0.34229860504256077,
      "grad_norm": 5.011477417461145,
      "learning_rate": 1.0395411078673053e-06,
      "loss": 2.7576,
      "step": 2644
    },
    {
      "epoch": 0.3424280674499142,
      "grad_norm": 2.817485004294429,
      "learning_rate": 1.0393296645251996e-06,
      "loss": 2.8467,
      "step": 2645
    },
    {
      "epoch": 0.3424280674499142,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6996403932571411,
      "eval_runtime": 13.7378,
      "eval_samples_per_second": 3.203,
      "eval_steps_per_second": 0.218,
      "step": 2645
    },
    {
      "epoch": 0.3425575298572677,
      "grad_norm": 2.8387188039581766,
      "learning_rate": 1.0391181365478616e-06,
      "loss": 2.7773,
      "step": 2646
    },
    {
      "epoch": 0.3426869922646212,
      "grad_norm": 2.4561634253412072,
      "learning_rate": 1.0389065239785005e-06,
      "loss": 2.5242,
      "step": 2647
    },
    {
      "epoch": 0.34281645467197464,
      "grad_norm": 4.47976755595692,
      "learning_rate": 1.0386948268603426e-06,
      "loss": 2.6035,
      "step": 2648
    },
    {
      "epoch": 0.3429459170793281,
      "grad_norm": 7.7028555556968845,
      "learning_rate": 1.0384830452366306e-06,
      "loss": 3.3071,
      "step": 2649
    },
    {
      "epoch": 0.34307537948668154,
      "grad_norm": 2.825157160526659,
      "learning_rate": 1.0382711791506262e-06,
      "loss": 2.7563,
      "step": 2650
    },
    {
      "epoch": 0.34307537948668154,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.700150966644287,
      "eval_runtime": 14.7455,
      "eval_samples_per_second": 2.984,
      "eval_steps_per_second": 0.203,
      "step": 2650
    },
    {
      "epoch": 0.343204841894035,
      "grad_norm": 2.172589344470153,
      "learning_rate": 1.0380592286456066e-06,
      "loss": 2.5469,
      "step": 2651
    },
    {
      "epoch": 0.3433343043013885,
      "grad_norm": 1.3977561893593125,
      "learning_rate": 1.0378471937648674e-06,
      "loss": 2.498,
      "step": 2652
    },
    {
      "epoch": 0.34346376670874196,
      "grad_norm": 2.730200312591785,
      "learning_rate": 1.0376350745517206e-06,
      "loss": 2.5947,
      "step": 2653
    },
    {
      "epoch": 0.3435932291160954,
      "grad_norm": 2.179287551660371,
      "learning_rate": 1.0374228710494962e-06,
      "loss": 2.7292,
      "step": 2654
    },
    {
      "epoch": 0.34372269152344886,
      "grad_norm": 2.0525204897167466,
      "learning_rate": 1.037210583301541e-06,
      "loss": 3.0515,
      "step": 2655
    },
    {
      "epoch": 0.34372269152344886,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.695268154144287,
      "eval_runtime": 14.7257,
      "eval_samples_per_second": 2.988,
      "eval_steps_per_second": 0.204,
      "step": 2655
    },
    {
      "epoch": 0.3438521539308023,
      "grad_norm": 2.530718630689439,
      "learning_rate": 1.036998211351219e-06,
      "loss": 2.6487,
      "step": 2656
    },
    {
      "epoch": 0.3439816163381558,
      "grad_norm": 2.787169732192521,
      "learning_rate": 1.0367857552419114e-06,
      "loss": 2.5308,
      "step": 2657
    },
    {
      "epoch": 0.3441110787455093,
      "grad_norm": 3.5793543978487468,
      "learning_rate": 1.0365732150170168e-06,
      "loss": 2.5918,
      "step": 2658
    },
    {
      "epoch": 0.34424054115286273,
      "grad_norm": 2.657309866469647,
      "learning_rate": 1.036360590719951e-06,
      "loss": 2.4951,
      "step": 2659
    },
    {
      "epoch": 0.3443700035602162,
      "grad_norm": 3.4454744958752133,
      "learning_rate": 1.0361478823941464e-06,
      "loss": 2.4207,
      "step": 2660
    },
    {
      "epoch": 0.3443700035602162,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6888316869735718,
      "eval_runtime": 14.0187,
      "eval_samples_per_second": 3.139,
      "eval_steps_per_second": 0.214,
      "step": 2660
    },
    {
      "epoch": 0.34449946596756964,
      "grad_norm": 2.075901874760244,
      "learning_rate": 1.0359350900830534e-06,
      "loss": 2.6682,
      "step": 2661
    },
    {
      "epoch": 0.34462892837492315,
      "grad_norm": 2.3784115166439577,
      "learning_rate": 1.035722213830139e-06,
      "loss": 2.6072,
      "step": 2662
    },
    {
      "epoch": 0.3447583907822766,
      "grad_norm": 2.3188567733504306,
      "learning_rate": 1.0355092536788875e-06,
      "loss": 2.3127,
      "step": 2663
    },
    {
      "epoch": 0.34488785318963006,
      "grad_norm": 2.7734313520199727,
      "learning_rate": 1.0352962096728001e-06,
      "loss": 2.4636,
      "step": 2664
    },
    {
      "epoch": 0.3450173155969835,
      "grad_norm": 1.8111529484358992,
      "learning_rate": 1.0350830818553957e-06,
      "loss": 2.5149,
      "step": 2665
    },
    {
      "epoch": 0.3450173155969835,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6924937963485718,
      "eval_runtime": 13.6105,
      "eval_samples_per_second": 3.233,
      "eval_steps_per_second": 0.22,
      "step": 2665
    },
    {
      "epoch": 0.345146778004337,
      "grad_norm": 1.889183736801805,
      "learning_rate": 1.0348698702702099e-06,
      "loss": 2.4775,
      "step": 2666
    },
    {
      "epoch": 0.34527624041169047,
      "grad_norm": 3.2249917674504336,
      "learning_rate": 1.0346565749607952e-06,
      "loss": 2.6787,
      "step": 2667
    },
    {
      "epoch": 0.3454057028190439,
      "grad_norm": 1.821137941358044,
      "learning_rate": 1.0344431959707219e-06,
      "loss": 2.6418,
      "step": 2668
    },
    {
      "epoch": 0.3455351652263974,
      "grad_norm": 2.849698049142795,
      "learning_rate": 1.0342297333435766e-06,
      "loss": 2.6606,
      "step": 2669
    },
    {
      "epoch": 0.34566462763375083,
      "grad_norm": 2.214619600958532,
      "learning_rate": 1.0340161871229637e-06,
      "loss": 2.5291,
      "step": 2670
    },
    {
      "epoch": 0.34566462763375083,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6917613744735718,
      "eval_runtime": 14.2198,
      "eval_samples_per_second": 3.094,
      "eval_steps_per_second": 0.211,
      "step": 2670
    },
    {
      "epoch": 0.34579409004110434,
      "grad_norm": 2.419401401318537,
      "learning_rate": 1.033802557352504e-06,
      "loss": 2.5942,
      "step": 2671
    },
    {
      "epoch": 0.3459235524484578,
      "grad_norm": 2.747847993981543,
      "learning_rate": 1.033588844075836e-06,
      "loss": 2.5269,
      "step": 2672
    },
    {
      "epoch": 0.34605301485581125,
      "grad_norm": 2.8344023600086463,
      "learning_rate": 1.033375047336615e-06,
      "loss": 2.5137,
      "step": 2673
    },
    {
      "epoch": 0.3461824772631647,
      "grad_norm": 4.751843239779162,
      "learning_rate": 1.0331611671785132e-06,
      "loss": 2.959,
      "step": 2674
    },
    {
      "epoch": 0.34631193967051815,
      "grad_norm": 1.741380372149227,
      "learning_rate": 1.0329472036452198e-06,
      "loss": 2.4663,
      "step": 2675
    },
    {
      "epoch": 0.34631193967051815,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6895862817764282,
      "eval_runtime": 13.6213,
      "eval_samples_per_second": 3.23,
      "eval_steps_per_second": 0.22,
      "step": 2675
    },
    {
      "epoch": 0.34644140207787166,
      "grad_norm": 3.0228141888865867,
      "learning_rate": 1.0327331567804414e-06,
      "loss": 2.4646,
      "step": 2676
    },
    {
      "epoch": 0.3465708644852251,
      "grad_norm": 1.9271678715472715,
      "learning_rate": 1.0325190266279013e-06,
      "loss": 2.6221,
      "step": 2677
    },
    {
      "epoch": 0.34670032689257857,
      "grad_norm": 3.7908192068887696,
      "learning_rate": 1.0323048132313402e-06,
      "loss": 2.6978,
      "step": 2678
    },
    {
      "epoch": 0.346829789299932,
      "grad_norm": 1.7386478788959716,
      "learning_rate": 1.0320905166345153e-06,
      "loss": 2.4043,
      "step": 2679
    },
    {
      "epoch": 0.3469592517072855,
      "grad_norm": 4.950923434577062,
      "learning_rate": 1.0318761368812008e-06,
      "loss": 3.041,
      "step": 2680
    },
    {
      "epoch": 0.3469592517072855,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6901633739471436,
      "eval_runtime": 14.1288,
      "eval_samples_per_second": 3.114,
      "eval_steps_per_second": 0.212,
      "step": 2680
    },
    {
      "epoch": 0.347088714114639,
      "grad_norm": 1.9711938077130988,
      "learning_rate": 1.0316616740151885e-06,
      "loss": 2.6323,
      "step": 2681
    },
    {
      "epoch": 0.34721817652199244,
      "grad_norm": 3.1359004609289243,
      "learning_rate": 1.0314471280802865e-06,
      "loss": 2.5208,
      "step": 2682
    },
    {
      "epoch": 0.3473476389293459,
      "grad_norm": 2.0582444037892142,
      "learning_rate": 1.0312324991203202e-06,
      "loss": 2.7563,
      "step": 2683
    },
    {
      "epoch": 0.34747710133669935,
      "grad_norm": 1.9005740163582243,
      "learning_rate": 1.0310177871791322e-06,
      "loss": 2.7087,
      "step": 2684
    },
    {
      "epoch": 0.3476065637440528,
      "grad_norm": 2.591412826778578,
      "learning_rate": 1.0308029923005815e-06,
      "loss": 2.522,
      "step": 2685
    },
    {
      "epoch": 0.3476065637440528,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6846369504928589,
      "eval_runtime": 15.2353,
      "eval_samples_per_second": 2.888,
      "eval_steps_per_second": 0.197,
      "step": 2685
    },
    {
      "epoch": 0.3477360261514063,
      "grad_norm": 5.741546712159457,
      "learning_rate": 1.0305881145285443e-06,
      "loss": 2.7358,
      "step": 2686
    },
    {
      "epoch": 0.34786548855875976,
      "grad_norm": 3.8499580237545556,
      "learning_rate": 1.030373153906914e-06,
      "loss": 2.6748,
      "step": 2687
    },
    {
      "epoch": 0.3479949509661132,
      "grad_norm": 5.421129292044323,
      "learning_rate": 1.0301581104796002e-06,
      "loss": 3.3887,
      "step": 2688
    },
    {
      "epoch": 0.34812441337346667,
      "grad_norm": 1.6757099464118863,
      "learning_rate": 1.0299429842905303e-06,
      "loss": 2.5768,
      "step": 2689
    },
    {
      "epoch": 0.3482538757808201,
      "grad_norm": 3.7659617938191854,
      "learning_rate": 1.0297277753836482e-06,
      "loss": 2.7148,
      "step": 2690
    },
    {
      "epoch": 0.3482538757808201,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6861683130264282,
      "eval_runtime": 13.9827,
      "eval_samples_per_second": 3.147,
      "eval_steps_per_second": 0.215,
      "step": 2690
    },
    {
      "epoch": 0.34838333818817363,
      "grad_norm": 1.6354992051198998,
      "learning_rate": 1.0295124838029143e-06,
      "loss": 2.6074,
      "step": 2691
    },
    {
      "epoch": 0.3485128005955271,
      "grad_norm": 1.8191334109481254,
      "learning_rate": 1.0292971095923065e-06,
      "loss": 2.7302,
      "step": 2692
    },
    {
      "epoch": 0.34864226300288054,
      "grad_norm": 2.7394726368716227,
      "learning_rate": 1.0290816527958194e-06,
      "loss": 3.0696,
      "step": 2693
    },
    {
      "epoch": 0.348771725410234,
      "grad_norm": 1.7938159342234976,
      "learning_rate": 1.0288661134574643e-06,
      "loss": 2.676,
      "step": 2694
    },
    {
      "epoch": 0.34890118781758744,
      "grad_norm": 1.7974756430980852,
      "learning_rate": 1.0286504916212697e-06,
      "loss": 2.6555,
      "step": 2695
    },
    {
      "epoch": 0.34890118781758744,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6839045286178589,
      "eval_runtime": 18.8183,
      "eval_samples_per_second": 2.338,
      "eval_steps_per_second": 0.159,
      "step": 2695
    },
    {
      "epoch": 0.34903065022494095,
      "grad_norm": 2.4877441014915576,
      "learning_rate": 1.0284347873312804e-06,
      "loss": 2.5808,
      "step": 2696
    },
    {
      "epoch": 0.3491601126322944,
      "grad_norm": 1.728651577605688,
      "learning_rate": 1.0282190006315584e-06,
      "loss": 2.6858,
      "step": 2697
    },
    {
      "epoch": 0.34928957503964786,
      "grad_norm": 3.904052093378296,
      "learning_rate": 1.0280031315661828e-06,
      "loss": 2.2908,
      "step": 2698
    },
    {
      "epoch": 0.3494190374470013,
      "grad_norm": 1.3845584160680322,
      "learning_rate": 1.0277871801792493e-06,
      "loss": 2.4937,
      "step": 2699
    },
    {
      "epoch": 0.34954849985435477,
      "grad_norm": 1.6504789588187299,
      "learning_rate": 1.0275711465148699e-06,
      "loss": 2.543,
      "step": 2700
    },
    {
      "epoch": 0.34954849985435477,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6812189817428589,
      "eval_runtime": 14.4984,
      "eval_samples_per_second": 3.035,
      "eval_steps_per_second": 0.207,
      "step": 2700
    },
    {
      "epoch": 0.3496779622617083,
      "grad_norm": 3.0083229568374534,
      "learning_rate": 1.027355030617174e-06,
      "loss": 2.7234,
      "step": 2701
    },
    {
      "epoch": 0.34980742466906173,
      "grad_norm": 2.090335179058034,
      "learning_rate": 1.027138832530308e-06,
      "loss": 2.811,
      "step": 2702
    },
    {
      "epoch": 0.3499368870764152,
      "grad_norm": 5.1120006546256915,
      "learning_rate": 1.0269225522984342e-06,
      "loss": 2.9458,
      "step": 2703
    },
    {
      "epoch": 0.35006634948376864,
      "grad_norm": 1.6654108556630227,
      "learning_rate": 1.0267061899657327e-06,
      "loss": 2.5515,
      "step": 2704
    },
    {
      "epoch": 0.3501958118911221,
      "grad_norm": 2.1581452489939905,
      "learning_rate": 1.0264897455763996e-06,
      "loss": 2.6143,
      "step": 2705
    },
    {
      "epoch": 0.3501958118911221,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.684037685394287,
      "eval_runtime": 13.8569,
      "eval_samples_per_second": 3.175,
      "eval_steps_per_second": 0.216,
      "step": 2705
    },
    {
      "epoch": 0.3503252742984756,
      "grad_norm": 5.093821586346566,
      "learning_rate": 1.026273219174648e-06,
      "loss": 2.676,
      "step": 2706
    },
    {
      "epoch": 0.35045473670582905,
      "grad_norm": 4.700494941897385,
      "learning_rate": 1.0260566108047081e-06,
      "loss": 2.9238,
      "step": 2707
    },
    {
      "epoch": 0.3505841991131825,
      "grad_norm": 1.7908531280720086,
      "learning_rate": 1.0258399205108263e-06,
      "loss": 2.5061,
      "step": 2708
    },
    {
      "epoch": 0.35071366152053596,
      "grad_norm": 2.7497820553652015,
      "learning_rate": 1.0256231483372662e-06,
      "loss": 2.8896,
      "step": 2709
    },
    {
      "epoch": 0.35084312392788947,
      "grad_norm": 2.153722077073625,
      "learning_rate": 1.0254062943283078e-06,
      "loss": 2.6071,
      "step": 2710
    },
    {
      "epoch": 0.35084312392788947,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6797763109207153,
      "eval_runtime": 15.018,
      "eval_samples_per_second": 2.93,
      "eval_steps_per_second": 0.2,
      "step": 2710
    },
    {
      "epoch": 0.3509725863352429,
      "grad_norm": 3.5698704746673604,
      "learning_rate": 1.0251893585282477e-06,
      "loss": 2.7341,
      "step": 2711
    },
    {
      "epoch": 0.3511020487425964,
      "grad_norm": 2.2968284368469134,
      "learning_rate": 1.0249723409814e-06,
      "loss": 2.6521,
      "step": 2712
    },
    {
      "epoch": 0.3512315111499498,
      "grad_norm": 1.7812857344635786,
      "learning_rate": 1.0247552417320942e-06,
      "loss": 2.564,
      "step": 2713
    },
    {
      "epoch": 0.3513609735573033,
      "grad_norm": 2.0443353334251047,
      "learning_rate": 1.024538060824678e-06,
      "loss": 2.9888,
      "step": 2714
    },
    {
      "epoch": 0.3514904359646568,
      "grad_norm": 1.708233465430465,
      "learning_rate": 1.0243207983035143e-06,
      "loss": 2.6377,
      "step": 2715
    },
    {
      "epoch": 0.3514904359646568,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6804865598678589,
      "eval_runtime": 13.2766,
      "eval_samples_per_second": 3.314,
      "eval_steps_per_second": 0.226,
      "step": 2715
    },
    {
      "epoch": 0.35161989837201024,
      "grad_norm": 3.1158704838603732,
      "learning_rate": 1.0241034542129838e-06,
      "loss": 2.4673,
      "step": 2716
    },
    {
      "epoch": 0.3517493607793637,
      "grad_norm": 2.1833769947337336,
      "learning_rate": 1.0238860285974833e-06,
      "loss": 2.5278,
      "step": 2717
    },
    {
      "epoch": 0.35187882318671715,
      "grad_norm": 3.022334241992606,
      "learning_rate": 1.0236685215014264e-06,
      "loss": 2.6958,
      "step": 2718
    },
    {
      "epoch": 0.3520082855940706,
      "grad_norm": 4.080487271973679,
      "learning_rate": 1.0234509329692432e-06,
      "loss": 2.6934,
      "step": 2719
    },
    {
      "epoch": 0.3521377480014241,
      "grad_norm": 2.214162767396084,
      "learning_rate": 1.0232332630453811e-06,
      "loss": 2.4307,
      "step": 2720
    },
    {
      "epoch": 0.3521377480014241,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6819514036178589,
      "eval_runtime": 14.3983,
      "eval_samples_per_second": 3.056,
      "eval_steps_per_second": 0.208,
      "step": 2720
    },
    {
      "epoch": 0.35226721040877756,
      "grad_norm": 2.7172335398715193,
      "learning_rate": 1.023015511774303e-06,
      "loss": 2.7332,
      "step": 2721
    },
    {
      "epoch": 0.352396672816131,
      "grad_norm": 2.4158222693994906,
      "learning_rate": 1.022797679200489e-06,
      "loss": 2.5942,
      "step": 2722
    },
    {
      "epoch": 0.35252613522348447,
      "grad_norm": 2.0427581008775877,
      "learning_rate": 1.0225797653684364e-06,
      "loss": 2.5781,
      "step": 2723
    },
    {
      "epoch": 0.3526555976308379,
      "grad_norm": 6.5018136173250225,
      "learning_rate": 1.022361770322658e-06,
      "loss": 3.0801,
      "step": 2724
    },
    {
      "epoch": 0.35278506003819143,
      "grad_norm": 2.3961859153367824,
      "learning_rate": 1.0221436941076835e-06,
      "loss": 2.877,
      "step": 2725
    },
    {
      "epoch": 0.35278506003819143,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.676225185394287,
      "eval_runtime": 14.0568,
      "eval_samples_per_second": 3.13,
      "eval_steps_per_second": 0.213,
      "step": 2725
    },
    {
      "epoch": 0.3529145224455449,
      "grad_norm": 5.468841195737375,
      "learning_rate": 1.0219255367680603e-06,
      "loss": 3.1128,
      "step": 2726
    },
    {
      "epoch": 0.35304398485289834,
      "grad_norm": 2.062869324882653,
      "learning_rate": 1.0217072983483507e-06,
      "loss": 2.6614,
      "step": 2727
    },
    {
      "epoch": 0.3531734472602518,
      "grad_norm": 4.964966444623541,
      "learning_rate": 1.0214889788931343e-06,
      "loss": 2.8145,
      "step": 2728
    },
    {
      "epoch": 0.35330290966760525,
      "grad_norm": 2.2202942839059445,
      "learning_rate": 1.0212705784470077e-06,
      "loss": 2.7339,
      "step": 2729
    },
    {
      "epoch": 0.35343237207495876,
      "grad_norm": 1.8581307436942252,
      "learning_rate": 1.0210520970545835e-06,
      "loss": 2.4746,
      "step": 2730
    },
    {
      "epoch": 0.35343237207495876,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6770241260528564,
      "eval_runtime": 12.9789,
      "eval_samples_per_second": 3.39,
      "eval_steps_per_second": 0.231,
      "step": 2730
    },
    {
      "epoch": 0.3535618344823122,
      "grad_norm": 1.7388438146243443,
      "learning_rate": 1.0208335347604907e-06,
      "loss": 2.6331,
      "step": 2731
    },
    {
      "epoch": 0.35369129688966566,
      "grad_norm": 1.8813465542086107,
      "learning_rate": 1.0206148916093753e-06,
      "loss": 2.6548,
      "step": 2732
    },
    {
      "epoch": 0.3538207592970191,
      "grad_norm": 5.134279151285838,
      "learning_rate": 1.0203961676458994e-06,
      "loss": 2.792,
      "step": 2733
    },
    {
      "epoch": 0.35395022170437257,
      "grad_norm": 4.5456065863723225,
      "learning_rate": 1.020177362914742e-06,
      "loss": 3.0032,
      "step": 2734
    },
    {
      "epoch": 0.3540796841117261,
      "grad_norm": 2.0698503757591,
      "learning_rate": 1.0199584774605985e-06,
      "loss": 2.6179,
      "step": 2735
    },
    {
      "epoch": 0.3540796841117261,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6804865598678589,
      "eval_runtime": 14.0312,
      "eval_samples_per_second": 3.136,
      "eval_steps_per_second": 0.214,
      "step": 2735
    },
    {
      "epoch": 0.35420914651907953,
      "grad_norm": 1.3548873945038056,
      "learning_rate": 1.0197395113281804e-06,
      "loss": 2.4907,
      "step": 2736
    },
    {
      "epoch": 0.354338608926433,
      "grad_norm": 1.593795083497264,
      "learning_rate": 1.0195204645622163e-06,
      "loss": 2.593,
      "step": 2737
    },
    {
      "epoch": 0.35446807133378644,
      "grad_norm": 2.8826349486182843,
      "learning_rate": 1.0193013372074507e-06,
      "loss": 3.1165,
      "step": 2738
    },
    {
      "epoch": 0.3545975337411399,
      "grad_norm": 2.086630510456705,
      "learning_rate": 1.0190821293086448e-06,
      "loss": 2.8477,
      "step": 2739
    },
    {
      "epoch": 0.3547269961484934,
      "grad_norm": 1.5066112311574449,
      "learning_rate": 1.018862840910576e-06,
      "loss": 2.4336,
      "step": 2740
    },
    {
      "epoch": 0.3547269961484934,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6783114671707153,
      "eval_runtime": 14.8834,
      "eval_samples_per_second": 2.956,
      "eval_steps_per_second": 0.202,
      "step": 2740
    },
    {
      "epoch": 0.35485645855584685,
      "grad_norm": 3.7675562931032984,
      "learning_rate": 1.0186434720580391e-06,
      "loss": 2.4443,
      "step": 2741
    },
    {
      "epoch": 0.3549859209632003,
      "grad_norm": 1.6962826407749612,
      "learning_rate": 1.0184240227958442e-06,
      "loss": 2.5996,
      "step": 2742
    },
    {
      "epoch": 0.35511538337055376,
      "grad_norm": 2.425459996139895,
      "learning_rate": 1.018204493168818e-06,
      "loss": 2.6594,
      "step": 2743
    },
    {
      "epoch": 0.3552448457779072,
      "grad_norm": 2.8893515810748265,
      "learning_rate": 1.0179848832218044e-06,
      "loss": 2.5491,
      "step": 2744
    },
    {
      "epoch": 0.3553743081852607,
      "grad_norm": 5.133294757648164,
      "learning_rate": 1.0177651929996632e-06,
      "loss": 3.0752,
      "step": 2745
    },
    {
      "epoch": 0.3553743081852607,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.676225185394287,
      "eval_runtime": 14.5075,
      "eval_samples_per_second": 3.033,
      "eval_steps_per_second": 0.207,
      "step": 2745
    },
    {
      "epoch": 0.3555037705926142,
      "grad_norm": 2.408280464977352,
      "learning_rate": 1.01754542254727e-06,
      "loss": 2.9233,
      "step": 2746
    },
    {
      "epoch": 0.35563323299996763,
      "grad_norm": 1.5601333048717037,
      "learning_rate": 1.0173255719095178e-06,
      "loss": 2.6821,
      "step": 2747
    },
    {
      "epoch": 0.3557626954073211,
      "grad_norm": 2.343098203115309,
      "learning_rate": 1.0171056411313155e-06,
      "loss": 3.1865,
      "step": 2748
    },
    {
      "epoch": 0.35589215781467454,
      "grad_norm": 2.387725791677623,
      "learning_rate": 1.0168856302575884e-06,
      "loss": 2.3638,
      "step": 2749
    },
    {
      "epoch": 0.35602162022202805,
      "grad_norm": 4.3263635031780145,
      "learning_rate": 1.016665539333278e-06,
      "loss": 2.9907,
      "step": 2750
    },
    {
      "epoch": 0.35602162022202805,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6748934984207153,
      "eval_runtime": 14.0396,
      "eval_samples_per_second": 3.134,
      "eval_steps_per_second": 0.214,
      "step": 2750
    },
    {
      "epoch": 0.3561510826293815,
      "grad_norm": 2.7220057185214745,
      "learning_rate": 1.0164453684033429e-06,
      "loss": 2.5071,
      "step": 2751
    },
    {
      "epoch": 0.35628054503673495,
      "grad_norm": 3.1620717437213175,
      "learning_rate": 1.0162251175127565e-06,
      "loss": 2.8092,
      "step": 2752
    },
    {
      "epoch": 0.3564100074440884,
      "grad_norm": 1.6714885458415707,
      "learning_rate": 1.0160047867065107e-06,
      "loss": 2.7773,
      "step": 2753
    },
    {
      "epoch": 0.3565394698514419,
      "grad_norm": 2.295309864492431,
      "learning_rate": 1.0157843760296114e-06,
      "loss": 2.6958,
      "step": 2754
    },
    {
      "epoch": 0.35666893225879537,
      "grad_norm": 2.2840634553892776,
      "learning_rate": 1.0155638855270824e-06,
      "loss": 2.573,
      "step": 2755
    },
    {
      "epoch": 0.35666893225879537,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6740944385528564,
      "eval_runtime": 13.7803,
      "eval_samples_per_second": 3.193,
      "eval_steps_per_second": 0.218,
      "step": 2755
    },
    {
      "epoch": 0.3567983946661488,
      "grad_norm": 1.8287983667862378,
      "learning_rate": 1.0153433152439639e-06,
      "loss": 2.5166,
      "step": 2756
    },
    {
      "epoch": 0.3569278570735023,
      "grad_norm": 2.2467825725453565,
      "learning_rate": 1.015122665225311e-06,
      "loss": 2.6299,
      "step": 2757
    },
    {
      "epoch": 0.35705731948085573,
      "grad_norm": 2.165339011762059,
      "learning_rate": 1.0149019355161964e-06,
      "loss": 2.6821,
      "step": 2758
    },
    {
      "epoch": 0.35718678188820924,
      "grad_norm": 2.440550737377332,
      "learning_rate": 1.0146811261617085e-06,
      "loss": 2.7727,
      "step": 2759
    },
    {
      "epoch": 0.3573162442955627,
      "grad_norm": 1.6163339496827576,
      "learning_rate": 1.0144602372069522e-06,
      "loss": 2.6943,
      "step": 2760
    },
    {
      "epoch": 0.3573162442955627,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6712092161178589,
      "eval_runtime": 13.6167,
      "eval_samples_per_second": 3.231,
      "eval_steps_per_second": 0.22,
      "step": 2760
    },
    {
      "epoch": 0.35744570670291614,
      "grad_norm": 1.8927438766702482,
      "learning_rate": 1.0142392686970485e-06,
      "loss": 2.6353,
      "step": 2761
    },
    {
      "epoch": 0.3575751691102696,
      "grad_norm": 1.5553088805241417,
      "learning_rate": 1.0140182206771342e-06,
      "loss": 2.5095,
      "step": 2762
    },
    {
      "epoch": 0.35770463151762305,
      "grad_norm": 1.8876681170860117,
      "learning_rate": 1.0137970931923638e-06,
      "loss": 2.6467,
      "step": 2763
    },
    {
      "epoch": 0.35783409392497656,
      "grad_norm": 4.219855432121571,
      "learning_rate": 1.0135758862879064e-06,
      "loss": 2.8311,
      "step": 2764
    },
    {
      "epoch": 0.35796355633233,
      "grad_norm": 1.655644040740648,
      "learning_rate": 1.013354600008948e-06,
      "loss": 2.4688,
      "step": 2765
    },
    {
      "epoch": 0.35796355633233,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6712092161178589,
      "eval_runtime": 14.966,
      "eval_samples_per_second": 2.94,
      "eval_steps_per_second": 0.2,
      "step": 2765
    },
    {
      "epoch": 0.35809301873968347,
      "grad_norm": 2.1231095595075424,
      "learning_rate": 1.0131332344006912e-06,
      "loss": 2.7087,
      "step": 2766
    },
    {
      "epoch": 0.3582224811470369,
      "grad_norm": 1.5107392012785503,
      "learning_rate": 1.012911789508354e-06,
      "loss": 2.4775,
      "step": 2767
    },
    {
      "epoch": 0.3583519435543904,
      "grad_norm": 1.839238804081981,
      "learning_rate": 1.0126902653771713e-06,
      "loss": 2.5251,
      "step": 2768
    },
    {
      "epoch": 0.3584814059617439,
      "grad_norm": 5.5209496826494275,
      "learning_rate": 1.0124686620523935e-06,
      "loss": 2.9309,
      "step": 2769
    },
    {
      "epoch": 0.35861086836909734,
      "grad_norm": 3.343828290739636,
      "learning_rate": 1.0122469795792882e-06,
      "loss": 2.5791,
      "step": 2770
    },
    {
      "epoch": 0.35861086836909734,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6704767942428589,
      "eval_runtime": 14.813,
      "eval_samples_per_second": 2.97,
      "eval_steps_per_second": 0.203,
      "step": 2770
    },
    {
      "epoch": 0.3587403307764508,
      "grad_norm": 1.7046274119818448,
      "learning_rate": 1.012025218003138e-06,
      "loss": 2.5901,
      "step": 2771
    },
    {
      "epoch": 0.35886979318380424,
      "grad_norm": 2.637861273350132,
      "learning_rate": 1.0118033773692423e-06,
      "loss": 2.9338,
      "step": 2772
    },
    {
      "epoch": 0.3589992555911577,
      "grad_norm": 2.1842011281667064,
      "learning_rate": 1.011581457722917e-06,
      "loss": 2.5325,
      "step": 2773
    },
    {
      "epoch": 0.3591287179985112,
      "grad_norm": 2.556678378504497,
      "learning_rate": 1.011359459109493e-06,
      "loss": 2.4214,
      "step": 2774
    },
    {
      "epoch": 0.35925818040586466,
      "grad_norm": 1.4792034026123355,
      "learning_rate": 1.0111373815743181e-06,
      "loss": 2.4813,
      "step": 2775
    },
    {
      "epoch": 0.35925818040586466,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6733176708221436,
      "eval_runtime": 14.6732,
      "eval_samples_per_second": 2.999,
      "eval_steps_per_second": 0.204,
      "step": 2775
    },
    {
      "epoch": 0.3593876428132181,
      "grad_norm": 1.9096861555874798,
      "learning_rate": 1.010915225162757e-06,
      "loss": 2.5115,
      "step": 2776
    },
    {
      "epoch": 0.35951710522057156,
      "grad_norm": 2.238511835345317,
      "learning_rate": 1.0106929899201884e-06,
      "loss": 2.9702,
      "step": 2777
    },
    {
      "epoch": 0.359646567627925,
      "grad_norm": 2.79009806027809,
      "learning_rate": 1.0104706758920093e-06,
      "loss": 2.7576,
      "step": 2778
    },
    {
      "epoch": 0.3597760300352785,
      "grad_norm": 2.3751177358846407,
      "learning_rate": 1.0102482831236314e-06,
      "loss": 2.5842,
      "step": 2779
    },
    {
      "epoch": 0.359905492442632,
      "grad_norm": 3.961157146226829,
      "learning_rate": 1.010025811660483e-06,
      "loss": 2.5696,
      "step": 2780
    },
    {
      "epoch": 0.359905492442632,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.672563076019287,
      "eval_runtime": 14.3783,
      "eval_samples_per_second": 3.06,
      "eval_steps_per_second": 0.209,
      "step": 2780
    },
    {
      "epoch": 0.36003495484998543,
      "grad_norm": 4.974593282026063,
      "learning_rate": 1.0098032615480086e-06,
      "loss": 3.1021,
      "step": 2781
    },
    {
      "epoch": 0.3601644172573389,
      "grad_norm": 2.163896203526194,
      "learning_rate": 1.0095806328316683e-06,
      "loss": 2.6873,
      "step": 2782
    },
    {
      "epoch": 0.36029387966469234,
      "grad_norm": 2.189618331821221,
      "learning_rate": 1.0093579255569387e-06,
      "loss": 2.5459,
      "step": 2783
    },
    {
      "epoch": 0.36042334207204585,
      "grad_norm": 2.1796018265233603,
      "learning_rate": 1.0091351397693123e-06,
      "loss": 2.4866,
      "step": 2784
    },
    {
      "epoch": 0.3605528044793993,
      "grad_norm": 1.6904144852131011,
      "learning_rate": 1.0089122755142974e-06,
      "loss": 2.52,
      "step": 2785
    },
    {
      "epoch": 0.3605528044793993,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6698997020721436,
      "eval_runtime": 13.7214,
      "eval_samples_per_second": 3.207,
      "eval_steps_per_second": 0.219,
      "step": 2785
    },
    {
      "epoch": 0.36068226688675276,
      "grad_norm": 1.979589460710121,
      "learning_rate": 1.0086893328374187e-06,
      "loss": 2.1804,
      "step": 2786
    },
    {
      "epoch": 0.3608117292941062,
      "grad_norm": 2.0287666268678475,
      "learning_rate": 1.0084663117842168e-06,
      "loss": 2.4736,
      "step": 2787
    },
    {
      "epoch": 0.36094119170145966,
      "grad_norm": 5.195221298775776,
      "learning_rate": 1.0082432124002483e-06,
      "loss": 2.6782,
      "step": 2788
    },
    {
      "epoch": 0.36107065410881317,
      "grad_norm": 1.7409462398376052,
      "learning_rate": 1.0080200347310853e-06,
      "loss": 2.6855,
      "step": 2789
    },
    {
      "epoch": 0.3612001165161666,
      "grad_norm": 5.078305839163032,
      "learning_rate": 1.0077967788223172e-06,
      "loss": 2.6436,
      "step": 2790
    },
    {
      "epoch": 0.3612001165161666,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6634854078292847,
      "eval_runtime": 14.538,
      "eval_samples_per_second": 3.027,
      "eval_steps_per_second": 0.206,
      "step": 2790
    },
    {
      "epoch": 0.3613295789235201,
      "grad_norm": 2.02645500316907,
      "learning_rate": 1.0075734447195478e-06,
      "loss": 2.4141,
      "step": 2791
    },
    {
      "epoch": 0.36145904133087353,
      "grad_norm": 2.1007482768477725,
      "learning_rate": 1.0073500324683983e-06,
      "loss": 2.5383,
      "step": 2792
    },
    {
      "epoch": 0.361588503738227,
      "grad_norm": 3.179793694113301,
      "learning_rate": 1.0071265421145042e-06,
      "loss": 2.6067,
      "step": 2793
    },
    {
      "epoch": 0.3617179661455805,
      "grad_norm": 1.7285730265829324,
      "learning_rate": 1.0069029737035188e-06,
      "loss": 2.6528,
      "step": 2794
    },
    {
      "epoch": 0.36184742855293395,
      "grad_norm": 2.0246104469104793,
      "learning_rate": 1.0066793272811104e-06,
      "loss": 2.4458,
      "step": 2795
    },
    {
      "epoch": 0.36184742855293395,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6627529859542847,
      "eval_runtime": 14.2601,
      "eval_samples_per_second": 3.086,
      "eval_steps_per_second": 0.21,
      "step": 2795
    },
    {
      "epoch": 0.3619768909602874,
      "grad_norm": 6.893027586606474,
      "learning_rate": 1.0064556028929628e-06,
      "loss": 2.6882,
      "step": 2796
    },
    {
      "epoch": 0.36210635336764085,
      "grad_norm": 2.6010042169066625,
      "learning_rate": 1.0062318005847769e-06,
      "loss": 2.6221,
      "step": 2797
    },
    {
      "epoch": 0.36223581577499436,
      "grad_norm": 1.931091047236256,
      "learning_rate": 1.0060079204022681e-06,
      "loss": 2.5515,
      "step": 2798
    },
    {
      "epoch": 0.3623652781823478,
      "grad_norm": 4.634160663027918,
      "learning_rate": 1.0057839623911692e-06,
      "loss": 2.7158,
      "step": 2799
    },
    {
      "epoch": 0.36249474058970127,
      "grad_norm": 1.9786446739319912,
      "learning_rate": 1.0055599265972277e-06,
      "loss": 2.3511,
      "step": 2800
    },
    {
      "epoch": 0.36249474058970127,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.659135341644287,
      "eval_runtime": 13.8539,
      "eval_samples_per_second": 3.176,
      "eval_steps_per_second": 0.217,
      "step": 2800
    },
    {
      "epoch": 0.3626242029970547,
      "grad_norm": 2.475862683678457,
      "learning_rate": 1.0053358130662076e-06,
      "loss": 2.5295,
      "step": 2801
    },
    {
      "epoch": 0.3627536654044082,
      "grad_norm": 2.3772247040170598,
      "learning_rate": 1.0051116218438886e-06,
      "loss": 2.51,
      "step": 2802
    },
    {
      "epoch": 0.3628831278117617,
      "grad_norm": 2.3539108657794583,
      "learning_rate": 1.0048873529760663e-06,
      "loss": 2.4082,
      "step": 2803
    },
    {
      "epoch": 0.36301259021911514,
      "grad_norm": 3.477079338290132,
      "learning_rate": 1.0046630065085524e-06,
      "loss": 2.8687,
      "step": 2804
    },
    {
      "epoch": 0.3631420526264686,
      "grad_norm": 2.0429439981906885,
      "learning_rate": 1.0044385824871738e-06,
      "loss": 2.3386,
      "step": 2805
    },
    {
      "epoch": 0.3631420526264686,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6570934057235718,
      "eval_runtime": 14.5406,
      "eval_samples_per_second": 3.026,
      "eval_steps_per_second": 0.206,
      "step": 2805
    },
    {
      "epoch": 0.36327151503382205,
      "grad_norm": 1.9850839606635313,
      "learning_rate": 1.004214080957774e-06,
      "loss": 2.79,
      "step": 2806
    },
    {
      "epoch": 0.3634009774411755,
      "grad_norm": 2.690719251684722,
      "learning_rate": 1.0039895019662115e-06,
      "loss": 2.6841,
      "step": 2807
    },
    {
      "epoch": 0.363530439848529,
      "grad_norm": 1.4601786832853443,
      "learning_rate": 1.0037648455583619e-06,
      "loss": 2.1992,
      "step": 2808
    },
    {
      "epoch": 0.36365990225588246,
      "grad_norm": 2.2198586946763967,
      "learning_rate": 1.0035401117801152e-06,
      "loss": 2.5422,
      "step": 2809
    },
    {
      "epoch": 0.3637893646632359,
      "grad_norm": 2.915728253811493,
      "learning_rate": 1.003315300677378e-06,
      "loss": 2.8694,
      "step": 2810
    },
    {
      "epoch": 0.3637893646632359,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6580477952957153,
      "eval_runtime": 13.5211,
      "eval_samples_per_second": 3.254,
      "eval_steps_per_second": 0.222,
      "step": 2810
    },
    {
      "epoch": 0.36391882707058937,
      "grad_norm": 2.5105635522505714,
      "learning_rate": 1.0030904122960725e-06,
      "loss": 2.4443,
      "step": 2811
    },
    {
      "epoch": 0.3640482894779428,
      "grad_norm": 3.8650002965511487,
      "learning_rate": 1.002865446682137e-06,
      "loss": 2.7808,
      "step": 2812
    },
    {
      "epoch": 0.36417775188529633,
      "grad_norm": 1.7522949256024827,
      "learning_rate": 1.0026404038815246e-06,
      "loss": 2.3232,
      "step": 2813
    },
    {
      "epoch": 0.3643072142926498,
      "grad_norm": 2.0230473849386463,
      "learning_rate": 1.0024152839402055e-06,
      "loss": 2.6321,
      "step": 2814
    },
    {
      "epoch": 0.36443667670000324,
      "grad_norm": 5.085466773697763,
      "learning_rate": 1.0021900869041648e-06,
      "loss": 2.6333,
      "step": 2815
    },
    {
      "epoch": 0.36443667670000324,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.655073642730713,
      "eval_runtime": 14.3665,
      "eval_samples_per_second": 3.063,
      "eval_steps_per_second": 0.209,
      "step": 2815
    },
    {
      "epoch": 0.3645661391073567,
      "grad_norm": 6.429554565963001,
      "learning_rate": 1.0019648128194035e-06,
      "loss": 2.8035,
      "step": 2816
    },
    {
      "epoch": 0.36469560151471014,
      "grad_norm": 8.111559914081266,
      "learning_rate": 1.0017394617319387e-06,
      "loss": 3.1309,
      "step": 2817
    },
    {
      "epoch": 0.36482506392206365,
      "grad_norm": 2.440917333251628,
      "learning_rate": 1.0015140336878024e-06,
      "loss": 2.741,
      "step": 2818
    },
    {
      "epoch": 0.3649545263294171,
      "grad_norm": 1.771033300371248,
      "learning_rate": 1.0012885287330432e-06,
      "loss": 2.4907,
      "step": 2819
    },
    {
      "epoch": 0.36508398873677056,
      "grad_norm": 13.314670461614778,
      "learning_rate": 1.001062946913725e-06,
      "loss": 2.9805,
      "step": 2820
    },
    {
      "epoch": 0.36508398873677056,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6578480005264282,
      "eval_runtime": 15.3193,
      "eval_samples_per_second": 2.872,
      "eval_steps_per_second": 0.196,
      "step": 2820
    },
    {
      "epoch": 0.365213451144124,
      "grad_norm": 1.3502093793286285,
      "learning_rate": 1.0008372882759275e-06,
      "loss": 2.3796,
      "step": 2821
    },
    {
      "epoch": 0.36534291355147747,
      "grad_norm": 2.4674474440678287,
      "learning_rate": 1.000611552865746e-06,
      "loss": 2.4067,
      "step": 2822
    },
    {
      "epoch": 0.365472375958831,
      "grad_norm": 4.508274988586452,
      "learning_rate": 1.0003857407292913e-06,
      "loss": 2.6113,
      "step": 2823
    },
    {
      "epoch": 0.36560183836618443,
      "grad_norm": 2.4223962758290494,
      "learning_rate": 1.0001598519126906e-06,
      "loss": 2.9194,
      "step": 2824
    },
    {
      "epoch": 0.3657313007735379,
      "grad_norm": 2.1409582420264317,
      "learning_rate": 9.999338864620857e-07,
      "loss": 2.3391,
      "step": 2825
    },
    {
      "epoch": 0.3657313007735379,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.655806064605713,
      "eval_runtime": 15.4518,
      "eval_samples_per_second": 2.848,
      "eval_steps_per_second": 0.194,
      "step": 2825
    },
    {
      "epoch": 0.36586076318089134,
      "grad_norm": 3.076984316415759,
      "learning_rate": 9.997078444236353e-07,
      "loss": 2.8618,
      "step": 2826
    },
    {
      "epoch": 0.3659902255882448,
      "grad_norm": 3.843688611039656,
      "learning_rate": 9.994817258435125e-07,
      "loss": 2.5103,
      "step": 2827
    },
    {
      "epoch": 0.3661196879955983,
      "grad_norm": 2.802440317812164,
      "learning_rate": 9.992555307679069e-07,
      "loss": 2.6716,
      "step": 2828
    },
    {
      "epoch": 0.36624915040295175,
      "grad_norm": 1.9016262641702826,
      "learning_rate": 9.990292592430233e-07,
      "loss": 2.6506,
      "step": 2829
    },
    {
      "epoch": 0.3663786128103052,
      "grad_norm": 2.7790468127629233,
      "learning_rate": 9.98802911315082e-07,
      "loss": 2.637,
      "step": 2830
    },
    {
      "epoch": 0.3663786128103052,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6550959348678589,
      "eval_runtime": 14.4675,
      "eval_samples_per_second": 3.041,
      "eval_steps_per_second": 0.207,
      "step": 2830
    },
    {
      "epoch": 0.36650807521765866,
      "grad_norm": 3.410805233067628,
      "learning_rate": 9.9857648703032e-07,
      "loss": 2.8799,
      "step": 2831
    },
    {
      "epoch": 0.3666375376250121,
      "grad_norm": 1.5865575286053357,
      "learning_rate": 9.98349986434988e-07,
      "loss": 2.4073,
      "step": 2832
    },
    {
      "epoch": 0.3667670000323656,
      "grad_norm": 2.480656546987658,
      "learning_rate": 9.981234095753538e-07,
      "loss": 2.752,
      "step": 2833
    },
    {
      "epoch": 0.3668964624397191,
      "grad_norm": 4.844094881588496,
      "learning_rate": 9.978967564977006e-07,
      "loss": 2.7197,
      "step": 2834
    },
    {
      "epoch": 0.3670259248470725,
      "grad_norm": 3.2388262079211523,
      "learning_rate": 9.976700272483263e-07,
      "loss": 2.7983,
      "step": 2835
    },
    {
      "epoch": 0.3670259248470725,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.657958984375,
      "eval_runtime": 14.2012,
      "eval_samples_per_second": 3.098,
      "eval_steps_per_second": 0.211,
      "step": 2835
    },
    {
      "epoch": 0.367155387254426,
      "grad_norm": 4.7590471637473115,
      "learning_rate": 9.974432218735455e-07,
      "loss": 3.3196,
      "step": 2836
    },
    {
      "epoch": 0.36728484966177943,
      "grad_norm": 2.920367032021074,
      "learning_rate": 9.972163404196876e-07,
      "loss": 2.6138,
      "step": 2837
    },
    {
      "epoch": 0.36741431206913294,
      "grad_norm": 1.6085521566111065,
      "learning_rate": 9.969893829330976e-07,
      "loss": 2.6841,
      "step": 2838
    },
    {
      "epoch": 0.3675437744764864,
      "grad_norm": 2.3473990777019917,
      "learning_rate": 9.96762349460136e-07,
      "loss": 2.7139,
      "step": 2839
    },
    {
      "epoch": 0.36767323688383985,
      "grad_norm": 1.5325845974280243,
      "learning_rate": 9.965352400471796e-07,
      "loss": 2.6448,
      "step": 2840
    },
    {
      "epoch": 0.36767323688383985,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6571155786514282,
      "eval_runtime": 14.1223,
      "eval_samples_per_second": 3.116,
      "eval_steps_per_second": 0.212,
      "step": 2840
    },
    {
      "epoch": 0.3678026992911933,
      "grad_norm": 7.2197322685688805,
      "learning_rate": 9.963080547406194e-07,
      "loss": 3.1055,
      "step": 2841
    },
    {
      "epoch": 0.3679321616985468,
      "grad_norm": 2.613122323300844,
      "learning_rate": 9.960807935868634e-07,
      "loss": 2.5474,
      "step": 2842
    },
    {
      "epoch": 0.36806162410590026,
      "grad_norm": 2.40469933508599,
      "learning_rate": 9.958534566323336e-07,
      "loss": 2.3914,
      "step": 2843
    },
    {
      "epoch": 0.3681910865132537,
      "grad_norm": 1.497259690156758,
      "learning_rate": 9.956260439234683e-07,
      "loss": 2.4963,
      "step": 2844
    },
    {
      "epoch": 0.36832054892060717,
      "grad_norm": 1.6311114945857514,
      "learning_rate": 9.953985555067216e-07,
      "loss": 2.5667,
      "step": 2845
    },
    {
      "epoch": 0.36832054892060717,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.654252529144287,
      "eval_runtime": 13.5696,
      "eval_samples_per_second": 3.243,
      "eval_steps_per_second": 0.221,
      "step": 2845
    },
    {
      "epoch": 0.3684500113279606,
      "grad_norm": 2.5058718760599636,
      "learning_rate": 9.95170991428562e-07,
      "loss": 2.7852,
      "step": 2846
    },
    {
      "epoch": 0.36857947373531413,
      "grad_norm": 3.9526452165720656,
      "learning_rate": 9.949433517354745e-07,
      "loss": 2.7664,
      "step": 2847
    },
    {
      "epoch": 0.3687089361426676,
      "grad_norm": 2.401036808136133,
      "learning_rate": 9.947156364739591e-07,
      "loss": 2.7397,
      "step": 2848
    },
    {
      "epoch": 0.36883839855002104,
      "grad_norm": 2.505679155203522,
      "learning_rate": 9.94487845690531e-07,
      "loss": 2.6333,
      "step": 2849
    },
    {
      "epoch": 0.3689678609573745,
      "grad_norm": 1.7640469435332933,
      "learning_rate": 9.942599794317212e-07,
      "loss": 2.8472,
      "step": 2850
    },
    {
      "epoch": 0.3689678609573745,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6530095338821411,
      "eval_runtime": 14.6131,
      "eval_samples_per_second": 3.011,
      "eval_steps_per_second": 0.205,
      "step": 2850
    },
    {
      "epoch": 0.36909732336472795,
      "grad_norm": 2.0032739892814853,
      "learning_rate": 9.940320377440763e-07,
      "loss": 2.5828,
      "step": 2851
    },
    {
      "epoch": 0.36922678577208146,
      "grad_norm": 3.651421515199593,
      "learning_rate": 9.938040206741574e-07,
      "loss": 2.6348,
      "step": 2852
    },
    {
      "epoch": 0.3693562481794349,
      "grad_norm": 1.9875035942644188,
      "learning_rate": 9.93575928268542e-07,
      "loss": 2.5078,
      "step": 2853
    },
    {
      "epoch": 0.36948571058678836,
      "grad_norm": 2.4908876466897727,
      "learning_rate": 9.933477605738228e-07,
      "loss": 2.8052,
      "step": 2854
    },
    {
      "epoch": 0.3696151729941418,
      "grad_norm": 2.248921545760026,
      "learning_rate": 9.931195176366073e-07,
      "loss": 2.6587,
      "step": 2855
    },
    {
      "epoch": 0.3696151729941418,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6558505296707153,
      "eval_runtime": 14.4355,
      "eval_samples_per_second": 3.048,
      "eval_steps_per_second": 0.208,
      "step": 2855
    },
    {
      "epoch": 0.36974463540149527,
      "grad_norm": 4.27973677002824,
      "learning_rate": 9.92891199503519e-07,
      "loss": 2.873,
      "step": 2856
    },
    {
      "epoch": 0.3698740978088488,
      "grad_norm": 3.8795296749133557,
      "learning_rate": 9.926628062211963e-07,
      "loss": 2.6931,
      "step": 2857
    },
    {
      "epoch": 0.37000356021620223,
      "grad_norm": 2.5705228220329825,
      "learning_rate": 9.924343378362934e-07,
      "loss": 2.6344,
      "step": 2858
    },
    {
      "epoch": 0.3701330226235557,
      "grad_norm": 1.5360341893314735,
      "learning_rate": 9.92205794395479e-07,
      "loss": 2.3833,
      "step": 2859
    },
    {
      "epoch": 0.37026248503090914,
      "grad_norm": 3.1868133263375458,
      "learning_rate": 9.919771759454386e-07,
      "loss": 2.4553,
      "step": 2860
    },
    {
      "epoch": 0.37026248503090914,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6549626588821411,
      "eval_runtime": 14.3026,
      "eval_samples_per_second": 3.076,
      "eval_steps_per_second": 0.21,
      "step": 2860
    },
    {
      "epoch": 0.3703919474382626,
      "grad_norm": 5.4817493711768925,
      "learning_rate": 9.917484825328717e-07,
      "loss": 2.853,
      "step": 2861
    },
    {
      "epoch": 0.3705214098456161,
      "grad_norm": 2.298353178048306,
      "learning_rate": 9.915197142044933e-07,
      "loss": 2.4646,
      "step": 2862
    },
    {
      "epoch": 0.37065087225296955,
      "grad_norm": 1.500264102969516,
      "learning_rate": 9.912908710070344e-07,
      "loss": 2.5222,
      "step": 2863
    },
    {
      "epoch": 0.370780334660323,
      "grad_norm": 3.958760515874364,
      "learning_rate": 9.910619529872408e-07,
      "loss": 2.8271,
      "step": 2864
    },
    {
      "epoch": 0.37090979706767646,
      "grad_norm": 1.9219125881825532,
      "learning_rate": 9.908329601918735e-07,
      "loss": 2.4551,
      "step": 2865
    },
    {
      "epoch": 0.37090979706767646,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6551846265792847,
      "eval_runtime": 14.1044,
      "eval_samples_per_second": 3.12,
      "eval_steps_per_second": 0.213,
      "step": 2865
    },
    {
      "epoch": 0.3710392594750299,
      "grad_norm": 4.470706267362769,
      "learning_rate": 9.90603892667709e-07,
      "loss": 2.8923,
      "step": 2866
    },
    {
      "epoch": 0.3711687218823834,
      "grad_norm": 5.050963397787529,
      "learning_rate": 9.903747504615391e-07,
      "loss": 2.8672,
      "step": 2867
    },
    {
      "epoch": 0.3712981842897369,
      "grad_norm": 1.865929709420239,
      "learning_rate": 9.901455336201704e-07,
      "loss": 2.6409,
      "step": 2868
    },
    {
      "epoch": 0.37142764669709033,
      "grad_norm": 2.490363858098563,
      "learning_rate": 9.899162421904254e-07,
      "loss": 2.593,
      "step": 2869
    },
    {
      "epoch": 0.3715571091044438,
      "grad_norm": 2.530135556907789,
      "learning_rate": 9.896868762191415e-07,
      "loss": 2.4377,
      "step": 2870
    },
    {
      "epoch": 0.3715571091044438,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6556285619735718,
      "eval_runtime": 14.6707,
      "eval_samples_per_second": 2.999,
      "eval_steps_per_second": 0.204,
      "step": 2870
    },
    {
      "epoch": 0.37168657151179724,
      "grad_norm": 1.8476682516909826,
      "learning_rate": 9.894574357531708e-07,
      "loss": 2.6143,
      "step": 2871
    },
    {
      "epoch": 0.37181603391915075,
      "grad_norm": 3.011132993834504,
      "learning_rate": 9.892279208393822e-07,
      "loss": 2.9265,
      "step": 2872
    },
    {
      "epoch": 0.3719454963265042,
      "grad_norm": 4.743411824262035,
      "learning_rate": 9.889983315246579e-07,
      "loss": 2.6582,
      "step": 2873
    },
    {
      "epoch": 0.37207495873385765,
      "grad_norm": 2.296851095502706,
      "learning_rate": 9.887686678558966e-07,
      "loss": 2.7363,
      "step": 2874
    },
    {
      "epoch": 0.3722044211412111,
      "grad_norm": 4.355762371389097,
      "learning_rate": 9.885389298800116e-07,
      "loss": 2.5522,
      "step": 2875
    },
    {
      "epoch": 0.3722044211412111,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6593128442764282,
      "eval_runtime": 15.0093,
      "eval_samples_per_second": 2.932,
      "eval_steps_per_second": 0.2,
      "step": 2875
    },
    {
      "epoch": 0.37233388354856456,
      "grad_norm": 1.9190621163487787,
      "learning_rate": 9.883091176439317e-07,
      "loss": 2.5459,
      "step": 2876
    },
    {
      "epoch": 0.37246334595591807,
      "grad_norm": 2.1408444637450907,
      "learning_rate": 9.880792311946002e-07,
      "loss": 2.7603,
      "step": 2877
    },
    {
      "epoch": 0.3725928083632715,
      "grad_norm": 1.3549598855291372,
      "learning_rate": 9.878492705789768e-07,
      "loss": 2.4688,
      "step": 2878
    },
    {
      "epoch": 0.372722270770625,
      "grad_norm": 3.305339304043236,
      "learning_rate": 9.876192358440352e-07,
      "loss": 2.3933,
      "step": 2879
    },
    {
      "epoch": 0.37285173317797843,
      "grad_norm": 2.3926662568220154,
      "learning_rate": 9.873891270367647e-07,
      "loss": 2.3467,
      "step": 2880
    },
    {
      "epoch": 0.37285173317797843,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.657759189605713,
      "eval_runtime": 13.7538,
      "eval_samples_per_second": 3.199,
      "eval_steps_per_second": 0.218,
      "step": 2880
    },
    {
      "epoch": 0.3729811955853319,
      "grad_norm": 1.3013718412397575,
      "learning_rate": 9.871589442041697e-07,
      "loss": 2.2065,
      "step": 2881
    },
    {
      "epoch": 0.3731106579926854,
      "grad_norm": 3.2762350523904753,
      "learning_rate": 9.869286873932696e-07,
      "loss": 2.7397,
      "step": 2882
    },
    {
      "epoch": 0.37324012040003884,
      "grad_norm": 1.64140064851061,
      "learning_rate": 9.866983566510994e-07,
      "loss": 2.5002,
      "step": 2883
    },
    {
      "epoch": 0.3733695828073923,
      "grad_norm": 4.37603285653419,
      "learning_rate": 9.864679520247084e-07,
      "loss": 2.7029,
      "step": 2884
    },
    {
      "epoch": 0.37349904521474575,
      "grad_norm": 1.673507357814267,
      "learning_rate": 9.862374735611615e-07,
      "loss": 2.5938,
      "step": 2885
    },
    {
      "epoch": 0.37349904521474575,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6661266088485718,
      "eval_runtime": 14.3128,
      "eval_samples_per_second": 3.074,
      "eval_steps_per_second": 0.21,
      "step": 2885
    },
    {
      "epoch": 0.37362850762209926,
      "grad_norm": 2.382173817967526,
      "learning_rate": 9.860069213075388e-07,
      "loss": 2.918,
      "step": 2886
    },
    {
      "epoch": 0.3737579700294527,
      "grad_norm": 6.317753514442637,
      "learning_rate": 9.85776295310935e-07,
      "loss": 3.3018,
      "step": 2887
    },
    {
      "epoch": 0.37388743243680617,
      "grad_norm": 5.054314906384003,
      "learning_rate": 9.855455956184603e-07,
      "loss": 2.8308,
      "step": 2888
    },
    {
      "epoch": 0.3740168948441596,
      "grad_norm": 1.9493958169853738,
      "learning_rate": 9.853148222772401e-07,
      "loss": 2.3599,
      "step": 2889
    },
    {
      "epoch": 0.3741463572515131,
      "grad_norm": 2.5835033412076056,
      "learning_rate": 9.85083975334414e-07,
      "loss": 2.4834,
      "step": 2890
    },
    {
      "epoch": 0.3741463572515131,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.661909580230713,
      "eval_runtime": 13.5669,
      "eval_samples_per_second": 3.243,
      "eval_steps_per_second": 0.221,
      "step": 2890
    },
    {
      "epoch": 0.3742758196588666,
      "grad_norm": 1.8001664560844015,
      "learning_rate": 9.848530548371373e-07,
      "loss": 2.5745,
      "step": 2891
    },
    {
      "epoch": 0.37440528206622004,
      "grad_norm": 2.6656714733562095,
      "learning_rate": 9.846220608325803e-07,
      "loss": 2.5488,
      "step": 2892
    },
    {
      "epoch": 0.3745347444735735,
      "grad_norm": 1.9325039528621062,
      "learning_rate": 9.843909933679281e-07,
      "loss": 2.5005,
      "step": 2893
    },
    {
      "epoch": 0.37466420688092694,
      "grad_norm": 1.975667413191557,
      "learning_rate": 9.84159852490381e-07,
      "loss": 2.8289,
      "step": 2894
    },
    {
      "epoch": 0.3747936692882804,
      "grad_norm": 3.0073670758070237,
      "learning_rate": 9.839286382471544e-07,
      "loss": 2.4468,
      "step": 2895
    },
    {
      "epoch": 0.3747936692882804,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6619762182235718,
      "eval_runtime": 14.3049,
      "eval_samples_per_second": 3.076,
      "eval_steps_per_second": 0.21,
      "step": 2895
    },
    {
      "epoch": 0.3749231316956339,
      "grad_norm": 1.8155380441287163,
      "learning_rate": 9.83697350685478e-07,
      "loss": 2.4429,
      "step": 2896
    },
    {
      "epoch": 0.37505259410298736,
      "grad_norm": 2.1761299156836804,
      "learning_rate": 9.834659898525973e-07,
      "loss": 2.4229,
      "step": 2897
    },
    {
      "epoch": 0.3751820565103408,
      "grad_norm": 3.0528446453627116,
      "learning_rate": 9.832345557957727e-07,
      "loss": 2.7722,
      "step": 2898
    },
    {
      "epoch": 0.37531151891769426,
      "grad_norm": 2.140615544097052,
      "learning_rate": 9.830030485622787e-07,
      "loss": 2.5239,
      "step": 2899
    },
    {
      "epoch": 0.3754409813250477,
      "grad_norm": 1.9308300578031676,
      "learning_rate": 9.827714681994055e-07,
      "loss": 2.6274,
      "step": 2900
    },
    {
      "epoch": 0.3754409813250477,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6619318723678589,
      "eval_runtime": 14.4291,
      "eval_samples_per_second": 3.049,
      "eval_steps_per_second": 0.208,
      "step": 2900
    },
    {
      "epoch": 0.3755704437324012,
      "grad_norm": 1.6084404192554373,
      "learning_rate": 9.825398147544585e-07,
      "loss": 2.4548,
      "step": 2901
    },
    {
      "epoch": 0.3756999061397547,
      "grad_norm": 2.652803983203279,
      "learning_rate": 9.823080882747574e-07,
      "loss": 2.6475,
      "step": 2902
    },
    {
      "epoch": 0.37582936854710813,
      "grad_norm": 2.8308892280846902,
      "learning_rate": 9.820762888076369e-07,
      "loss": 2.4497,
      "step": 2903
    },
    {
      "epoch": 0.3759588309544616,
      "grad_norm": 2.9916139786632296,
      "learning_rate": 9.818444164004465e-07,
      "loss": 2.6519,
      "step": 2904
    },
    {
      "epoch": 0.37608829336181504,
      "grad_norm": 1.4481726042951721,
      "learning_rate": 9.816124711005514e-07,
      "loss": 2.6091,
      "step": 2905
    },
    {
      "epoch": 0.37608829336181504,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6597567796707153,
      "eval_runtime": 15.3161,
      "eval_samples_per_second": 2.873,
      "eval_steps_per_second": 0.196,
      "step": 2905
    },
    {
      "epoch": 0.37621775576916855,
      "grad_norm": 1.4605777984911823,
      "learning_rate": 9.81380452955331e-07,
      "loss": 2.4836,
      "step": 2906
    },
    {
      "epoch": 0.376347218176522,
      "grad_norm": 2.4925902303032577,
      "learning_rate": 9.811483620121796e-07,
      "loss": 2.5605,
      "step": 2907
    },
    {
      "epoch": 0.37647668058387546,
      "grad_norm": 1.5914155002417345,
      "learning_rate": 9.80916198318506e-07,
      "loss": 2.4868,
      "step": 2908
    },
    {
      "epoch": 0.3766061429912289,
      "grad_norm": 1.9200362031011797,
      "learning_rate": 9.806839619217353e-07,
      "loss": 2.5215,
      "step": 2909
    },
    {
      "epoch": 0.37673560539858236,
      "grad_norm": 3.473186346207707,
      "learning_rate": 9.804516528693056e-07,
      "loss": 2.6567,
      "step": 2910
    },
    {
      "epoch": 0.37673560539858236,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6583806276321411,
      "eval_runtime": 14.5909,
      "eval_samples_per_second": 3.016,
      "eval_steps_per_second": 0.206,
      "step": 2910
    },
    {
      "epoch": 0.37686506780593587,
      "grad_norm": 3.2173623241077616,
      "learning_rate": 9.802192712086713e-07,
      "loss": 2.7097,
      "step": 2911
    },
    {
      "epoch": 0.3769945302132893,
      "grad_norm": 2.8911008257065727,
      "learning_rate": 9.799868169873011e-07,
      "loss": 2.5015,
      "step": 2912
    },
    {
      "epoch": 0.3771239926206428,
      "grad_norm": 3.0663374283331013,
      "learning_rate": 9.797542902526779e-07,
      "loss": 2.5581,
      "step": 2913
    },
    {
      "epoch": 0.37725345502799623,
      "grad_norm": 3.8681020052892023,
      "learning_rate": 9.795216910523006e-07,
      "loss": 2.4287,
      "step": 2914
    },
    {
      "epoch": 0.3773829174353497,
      "grad_norm": 1.6984825201174019,
      "learning_rate": 9.79289019433682e-07,
      "loss": 2.4229,
      "step": 2915
    },
    {
      "epoch": 0.3773829174353497,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6563166379928589,
      "eval_runtime": 19.0671,
      "eval_samples_per_second": 2.308,
      "eval_steps_per_second": 0.157,
      "step": 2915
    },
    {
      "epoch": 0.3775123798427032,
      "grad_norm": 3.078213800287615,
      "learning_rate": 9.7905627544435e-07,
      "loss": 2.8472,
      "step": 2916
    },
    {
      "epoch": 0.37764184225005665,
      "grad_norm": 3.765963855989695,
      "learning_rate": 9.788234591318476e-07,
      "loss": 2.7612,
      "step": 2917
    },
    {
      "epoch": 0.3777713046574101,
      "grad_norm": 2.748862332289295,
      "learning_rate": 9.785905705437318e-07,
      "loss": 2.7075,
      "step": 2918
    },
    {
      "epoch": 0.37790076706476355,
      "grad_norm": 1.9401350323303972,
      "learning_rate": 9.783576097275751e-07,
      "loss": 2.7952,
      "step": 2919
    },
    {
      "epoch": 0.378030229472117,
      "grad_norm": 1.8415919653801276,
      "learning_rate": 9.781245767309644e-07,
      "loss": 2.5034,
      "step": 2920
    },
    {
      "epoch": 0.378030229472117,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6569602489471436,
      "eval_runtime": 13.555,
      "eval_samples_per_second": 3.246,
      "eval_steps_per_second": 0.221,
      "step": 2920
    },
    {
      "epoch": 0.3781596918794705,
      "grad_norm": 2.074431723700032,
      "learning_rate": 9.778914716015012e-07,
      "loss": 2.532,
      "step": 2921
    },
    {
      "epoch": 0.37828915428682397,
      "grad_norm": 2.089934975816996,
      "learning_rate": 9.776582943868023e-07,
      "loss": 2.7031,
      "step": 2922
    },
    {
      "epoch": 0.3784186166941774,
      "grad_norm": 2.7033932285447397,
      "learning_rate": 9.774250451344987e-07,
      "loss": 2.7354,
      "step": 2923
    },
    {
      "epoch": 0.3785480791015309,
      "grad_norm": 2.3699101815668264,
      "learning_rate": 9.771917238922362e-07,
      "loss": 2.7817,
      "step": 2924
    },
    {
      "epoch": 0.37867754150888433,
      "grad_norm": 5.54804654108326,
      "learning_rate": 9.769583307076757e-07,
      "loss": 2.791,
      "step": 2925
    },
    {
      "epoch": 0.37867754150888433,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6641291379928589,
      "eval_runtime": 14.2993,
      "eval_samples_per_second": 3.077,
      "eval_steps_per_second": 0.21,
      "step": 2925
    },
    {
      "epoch": 0.37880700391623784,
      "grad_norm": 1.7222078821646445,
      "learning_rate": 9.767248656284923e-07,
      "loss": 2.5132,
      "step": 2926
    },
    {
      "epoch": 0.3789364663235913,
      "grad_norm": 2.534085602599922,
      "learning_rate": 9.764913287023756e-07,
      "loss": 2.6869,
      "step": 2927
    },
    {
      "epoch": 0.37906592873094475,
      "grad_norm": 3.294391051355385,
      "learning_rate": 9.762577199770308e-07,
      "loss": 2.6772,
      "step": 2928
    },
    {
      "epoch": 0.3791953911382982,
      "grad_norm": 2.7697912514523635,
      "learning_rate": 9.76024039500177e-07,
      "loss": 2.8999,
      "step": 2929
    },
    {
      "epoch": 0.3793248535456517,
      "grad_norm": 1.7330008099974472,
      "learning_rate": 9.757902873195483e-07,
      "loss": 2.3726,
      "step": 2930
    },
    {
      "epoch": 0.3793248535456517,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6541193723678589,
      "eval_runtime": 14.9899,
      "eval_samples_per_second": 2.935,
      "eval_steps_per_second": 0.2,
      "step": 2930
    },
    {
      "epoch": 0.37945431595300516,
      "grad_norm": 2.3639115719611468,
      "learning_rate": 9.755564634828932e-07,
      "loss": 2.7158,
      "step": 2931
    },
    {
      "epoch": 0.3795837783603586,
      "grad_norm": 2.9866876769816684,
      "learning_rate": 9.75322568037975e-07,
      "loss": 2.8494,
      "step": 2932
    },
    {
      "epoch": 0.37971324076771207,
      "grad_norm": 1.6283745546057775,
      "learning_rate": 9.750886010325714e-07,
      "loss": 2.4629,
      "step": 2933
    },
    {
      "epoch": 0.3798427031750655,
      "grad_norm": 6.156605835983724,
      "learning_rate": 9.748545625144754e-07,
      "loss": 2.8977,
      "step": 2934
    },
    {
      "epoch": 0.37997216558241903,
      "grad_norm": 2.3995464251617507,
      "learning_rate": 9.746204525314933e-07,
      "loss": 2.8044,
      "step": 2935
    },
    {
      "epoch": 0.37997216558241903,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6548073291778564,
      "eval_runtime": 14.5986,
      "eval_samples_per_second": 3.014,
      "eval_steps_per_second": 0.205,
      "step": 2935
    },
    {
      "epoch": 0.3801016279897725,
      "grad_norm": 3.8171861189613345,
      "learning_rate": 9.743862711314474e-07,
      "loss": 2.9854,
      "step": 2936
    },
    {
      "epoch": 0.38023109039712594,
      "grad_norm": 1.8957259724269466,
      "learning_rate": 9.741520183621743e-07,
      "loss": 2.5054,
      "step": 2937
    },
    {
      "epoch": 0.3803605528044794,
      "grad_norm": 1.5755684986821261,
      "learning_rate": 9.739176942715242e-07,
      "loss": 2.6443,
      "step": 2938
    },
    {
      "epoch": 0.38049001521183284,
      "grad_norm": 2.2671974570160325,
      "learning_rate": 9.736832989073627e-07,
      "loss": 2.3574,
      "step": 2939
    },
    {
      "epoch": 0.38061947761918635,
      "grad_norm": 2.256939764780161,
      "learning_rate": 9.7344883231757e-07,
      "loss": 2.6975,
      "step": 2940
    },
    {
      "epoch": 0.38061947761918635,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6541415452957153,
      "eval_runtime": 13.6346,
      "eval_samples_per_second": 3.227,
      "eval_steps_per_second": 0.22,
      "step": 2940
    },
    {
      "epoch": 0.3807489400265398,
      "grad_norm": 1.537000010993564,
      "learning_rate": 9.732142945500406e-07,
      "loss": 2.5781,
      "step": 2941
    },
    {
      "epoch": 0.38087840243389326,
      "grad_norm": 2.8647123705139865,
      "learning_rate": 9.729796856526835e-07,
      "loss": 2.8188,
      "step": 2942
    },
    {
      "epoch": 0.3810078648412467,
      "grad_norm": 7.533585750634245,
      "learning_rate": 9.727450056734224e-07,
      "loss": 2.7349,
      "step": 2943
    },
    {
      "epoch": 0.38113732724860017,
      "grad_norm": 1.7953750276427403,
      "learning_rate": 9.725102546601954e-07,
      "loss": 2.3374,
      "step": 2944
    },
    {
      "epoch": 0.3812667896559537,
      "grad_norm": 4.720789789105859,
      "learning_rate": 9.72275432660955e-07,
      "loss": 2.9561,
      "step": 2945
    },
    {
      "epoch": 0.3812667896559537,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.65625,
      "eval_runtime": 13.9738,
      "eval_samples_per_second": 3.149,
      "eval_steps_per_second": 0.215,
      "step": 2945
    },
    {
      "epoch": 0.38139625206330713,
      "grad_norm": 3.0738385522472806,
      "learning_rate": 9.720405397236691e-07,
      "loss": 2.8416,
      "step": 2946
    },
    {
      "epoch": 0.3815257144706606,
      "grad_norm": 2.836532489067419,
      "learning_rate": 9.718055758963181e-07,
      "loss": 2.5464,
      "step": 2947
    },
    {
      "epoch": 0.38165517687801404,
      "grad_norm": 4.281916370043952,
      "learning_rate": 9.71570541226899e-07,
      "loss": 2.4062,
      "step": 2948
    },
    {
      "epoch": 0.3817846392853675,
      "grad_norm": 2.122999598567164,
      "learning_rate": 9.713354357634226e-07,
      "loss": 2.5308,
      "step": 2949
    },
    {
      "epoch": 0.381914101692721,
      "grad_norm": 1.8628158389704288,
      "learning_rate": 9.71100259553913e-07,
      "loss": 2.5471,
      "step": 2950
    },
    {
      "epoch": 0.381914101692721,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6561611890792847,
      "eval_runtime": 14.8935,
      "eval_samples_per_second": 2.954,
      "eval_steps_per_second": 0.201,
      "step": 2950
    },
    {
      "epoch": 0.38204356410007445,
      "grad_norm": 5.113849581773972,
      "learning_rate": 9.708650126464107e-07,
      "loss": 2.939,
      "step": 2951
    },
    {
      "epoch": 0.3821730265074279,
      "grad_norm": 1.7565864308817931,
      "learning_rate": 9.706296950889689e-07,
      "loss": 2.6011,
      "step": 2952
    },
    {
      "epoch": 0.38230248891478136,
      "grad_norm": 1.5996920854468977,
      "learning_rate": 9.703943069296563e-07,
      "loss": 2.5493,
      "step": 2953
    },
    {
      "epoch": 0.3824319513221348,
      "grad_norm": 2.6851476414595106,
      "learning_rate": 9.701588482165559e-07,
      "loss": 2.7559,
      "step": 2954
    },
    {
      "epoch": 0.3825614137294883,
      "grad_norm": 1.3428752731818252,
      "learning_rate": 9.699233189977645e-07,
      "loss": 2.4636,
      "step": 2955
    },
    {
      "epoch": 0.3825614137294883,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6582697629928589,
      "eval_runtime": 15.8023,
      "eval_samples_per_second": 2.784,
      "eval_steps_per_second": 0.19,
      "step": 2955
    },
    {
      "epoch": 0.3826908761368418,
      "grad_norm": 2.2261597800303186,
      "learning_rate": 9.696877193213938e-07,
      "loss": 2.5762,
      "step": 2956
    },
    {
      "epoch": 0.3828203385441952,
      "grad_norm": 2.4269587201467084,
      "learning_rate": 9.694520492355702e-07,
      "loss": 2.7566,
      "step": 2957
    },
    {
      "epoch": 0.3829498009515487,
      "grad_norm": 6.1915831615432735,
      "learning_rate": 9.692163087884336e-07,
      "loss": 2.6562,
      "step": 2958
    },
    {
      "epoch": 0.38307926335890213,
      "grad_norm": 1.7659446402090824,
      "learning_rate": 9.689804980281392e-07,
      "loss": 2.5415,
      "step": 2959
    },
    {
      "epoch": 0.38320872576625564,
      "grad_norm": 1.6182461570124533,
      "learning_rate": 9.687446170028559e-07,
      "loss": 2.5049,
      "step": 2960
    },
    {
      "epoch": 0.38320872576625564,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6582919359207153,
      "eval_runtime": 14.7714,
      "eval_samples_per_second": 2.979,
      "eval_steps_per_second": 0.203,
      "step": 2960
    },
    {
      "epoch": 0.3833381881736091,
      "grad_norm": 2.484408258114913,
      "learning_rate": 9.685086657607668e-07,
      "loss": 2.4341,
      "step": 2961
    },
    {
      "epoch": 0.38346765058096255,
      "grad_norm": 1.4219949272051484,
      "learning_rate": 9.682726443500707e-07,
      "loss": 2.3198,
      "step": 2962
    },
    {
      "epoch": 0.383597112988316,
      "grad_norm": 2.1599996596051394,
      "learning_rate": 9.680365528189787e-07,
      "loss": 2.5337,
      "step": 2963
    },
    {
      "epoch": 0.38372657539566946,
      "grad_norm": 2.849560221647801,
      "learning_rate": 9.678003912157178e-07,
      "loss": 2.4512,
      "step": 2964
    },
    {
      "epoch": 0.38385603780302296,
      "grad_norm": 2.0525007196477527,
      "learning_rate": 9.67564159588529e-07,
      "loss": 2.5657,
      "step": 2965
    },
    {
      "epoch": 0.38385603780302296,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6569157838821411,
      "eval_runtime": 14.6266,
      "eval_samples_per_second": 3.008,
      "eval_steps_per_second": 0.205,
      "step": 2965
    },
    {
      "epoch": 0.3839855002103764,
      "grad_norm": 11.853506320848801,
      "learning_rate": 9.67327857985667e-07,
      "loss": 3.2476,
      "step": 2966
    },
    {
      "epoch": 0.38411496261772987,
      "grad_norm": 1.395096003546566,
      "learning_rate": 9.670914864554014e-07,
      "loss": 2.6218,
      "step": 2967
    },
    {
      "epoch": 0.3842444250250833,
      "grad_norm": 2.418421507677662,
      "learning_rate": 9.668550450460156e-07,
      "loss": 2.813,
      "step": 2968
    },
    {
      "epoch": 0.38437388743243683,
      "grad_norm": 1.8690147924311018,
      "learning_rate": 9.666185338058083e-07,
      "loss": 2.6091,
      "step": 2969
    },
    {
      "epoch": 0.3845033498397903,
      "grad_norm": 1.9932098214134655,
      "learning_rate": 9.663819527830906e-07,
      "loss": 2.5935,
      "step": 2970
    },
    {
      "epoch": 0.3845033498397903,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6590021848678589,
      "eval_runtime": 14.1727,
      "eval_samples_per_second": 3.105,
      "eval_steps_per_second": 0.212,
      "step": 2970
    },
    {
      "epoch": 0.38463281224714374,
      "grad_norm": 3.7376183681945423,
      "learning_rate": 9.661453020261898e-07,
      "loss": 2.376,
      "step": 2971
    },
    {
      "epoch": 0.3847622746544972,
      "grad_norm": 2.1725870402709737,
      "learning_rate": 9.659085815834463e-07,
      "loss": 2.5225,
      "step": 2972
    },
    {
      "epoch": 0.38489173706185065,
      "grad_norm": 2.779145919928225,
      "learning_rate": 9.656717915032153e-07,
      "loss": 2.4167,
      "step": 2973
    },
    {
      "epoch": 0.38502119946920416,
      "grad_norm": 2.5133586193138457,
      "learning_rate": 9.654349318338655e-07,
      "loss": 2.59,
      "step": 2974
    },
    {
      "epoch": 0.3851506618765576,
      "grad_norm": 1.9299002870059516,
      "learning_rate": 9.651980026237805e-07,
      "loss": 2.6021,
      "step": 2975
    },
    {
      "epoch": 0.3851506618765576,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6560946702957153,
      "eval_runtime": 14.4097,
      "eval_samples_per_second": 3.053,
      "eval_steps_per_second": 0.208,
      "step": 2975
    },
    {
      "epoch": 0.38528012428391106,
      "grad_norm": 1.2785603696142616,
      "learning_rate": 9.649610039213584e-07,
      "loss": 2.5317,
      "step": 2976
    },
    {
      "epoch": 0.3854095866912645,
      "grad_norm": 3.9870480037821854,
      "learning_rate": 9.647239357750101e-07,
      "loss": 3.2422,
      "step": 2977
    },
    {
      "epoch": 0.38553904909861797,
      "grad_norm": 1.8537570662618423,
      "learning_rate": 9.644867982331623e-07,
      "loss": 2.684,
      "step": 2978
    },
    {
      "epoch": 0.3856685115059715,
      "grad_norm": 7.254850958977091,
      "learning_rate": 9.642495913442548e-07,
      "loss": 2.9907,
      "step": 2979
    },
    {
      "epoch": 0.38579797391332493,
      "grad_norm": 3.0255818677492896,
      "learning_rate": 9.64012315156742e-07,
      "loss": 2.6094,
      "step": 2980
    },
    {
      "epoch": 0.38579797391332493,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6624866724014282,
      "eval_runtime": 14.2266,
      "eval_samples_per_second": 3.093,
      "eval_steps_per_second": 0.211,
      "step": 2980
    },
    {
      "epoch": 0.3859274363206784,
      "grad_norm": 3.5454611451475273,
      "learning_rate": 9.637749697190924e-07,
      "loss": 3.0154,
      "step": 2981
    },
    {
      "epoch": 0.38605689872803184,
      "grad_norm": 2.1053962938503226,
      "learning_rate": 9.635375550797885e-07,
      "loss": 2.6953,
      "step": 2982
    },
    {
      "epoch": 0.3861863611353853,
      "grad_norm": 1.871193696107483,
      "learning_rate": 9.633000712873272e-07,
      "loss": 2.7642,
      "step": 2983
    },
    {
      "epoch": 0.3863158235427388,
      "grad_norm": 2.864871562016539,
      "learning_rate": 9.630625183902193e-07,
      "loss": 2.5371,
      "step": 2984
    },
    {
      "epoch": 0.38644528595009225,
      "grad_norm": 1.6673722158700757,
      "learning_rate": 9.628248964369898e-07,
      "loss": 2.5378,
      "step": 2985
    },
    {
      "epoch": 0.38644528595009225,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6602672338485718,
      "eval_runtime": 14.331,
      "eval_samples_per_second": 3.07,
      "eval_steps_per_second": 0.209,
      "step": 2985
    },
    {
      "epoch": 0.3865747483574457,
      "grad_norm": 2.3111444259834415,
      "learning_rate": 9.625872054761777e-07,
      "loss": 2.7214,
      "step": 2986
    },
    {
      "epoch": 0.38670421076479916,
      "grad_norm": 1.9096602296379652,
      "learning_rate": 9.623494455563365e-07,
      "loss": 2.5627,
      "step": 2987
    },
    {
      "epoch": 0.3868336731721526,
      "grad_norm": 2.404692182300404,
      "learning_rate": 9.62111616726033e-07,
      "loss": 2.6292,
      "step": 2988
    },
    {
      "epoch": 0.3869631355795061,
      "grad_norm": 3.9961279306940005,
      "learning_rate": 9.61873719033849e-07,
      "loss": 2.9214,
      "step": 2989
    },
    {
      "epoch": 0.3870925979868596,
      "grad_norm": 2.047446637335981,
      "learning_rate": 9.6163575252838e-07,
      "loss": 2.6404,
      "step": 2990
    },
    {
      "epoch": 0.3870925979868596,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.658935546875,
      "eval_runtime": 15.1112,
      "eval_samples_per_second": 2.912,
      "eval_steps_per_second": 0.199,
      "step": 2990
    },
    {
      "epoch": 0.38722206039421303,
      "grad_norm": 2.2424753155688917,
      "learning_rate": 9.61397717258235e-07,
      "loss": 2.645,
      "step": 2991
    },
    {
      "epoch": 0.3873515228015665,
      "grad_norm": 1.6318271188843687,
      "learning_rate": 9.611596132720377e-07,
      "loss": 2.2922,
      "step": 2992
    },
    {
      "epoch": 0.38748098520891994,
      "grad_norm": 1.723731338415686,
      "learning_rate": 9.609214406184258e-07,
      "loss": 2.4976,
      "step": 2993
    },
    {
      "epoch": 0.38761044761627345,
      "grad_norm": 4.461887424499677,
      "learning_rate": 9.60683199346051e-07,
      "loss": 2.7773,
      "step": 2994
    },
    {
      "epoch": 0.3877399100236269,
      "grad_norm": 3.936086616572627,
      "learning_rate": 9.604448895035787e-07,
      "loss": 2.7976,
      "step": 2995
    },
    {
      "epoch": 0.3877399100236269,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.658247470855713,
      "eval_runtime": 14.6957,
      "eval_samples_per_second": 2.994,
      "eval_steps_per_second": 0.204,
      "step": 2995
    },
    {
      "epoch": 0.38786937243098035,
      "grad_norm": 2.146365617431816,
      "learning_rate": 9.602065111396886e-07,
      "loss": 2.6599,
      "step": 2996
    },
    {
      "epoch": 0.3879988348383338,
      "grad_norm": 2.0225830095154542,
      "learning_rate": 9.59968064303074e-07,
      "loss": 2.6436,
      "step": 2997
    },
    {
      "epoch": 0.38812829724568726,
      "grad_norm": 2.3690298915166017,
      "learning_rate": 9.59729549042443e-07,
      "loss": 2.7358,
      "step": 2998
    },
    {
      "epoch": 0.38825775965304077,
      "grad_norm": 2.1008036567206863,
      "learning_rate": 9.594909654065168e-07,
      "loss": 2.5073,
      "step": 2999
    },
    {
      "epoch": 0.3883872220603942,
      "grad_norm": 1.5558000073599048,
      "learning_rate": 9.592523134440312e-07,
      "loss": 2.4048,
      "step": 3000
    },
    {
      "epoch": 0.3883872220603942,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.657470703125,
      "eval_runtime": 13.3324,
      "eval_samples_per_second": 3.3,
      "eval_steps_per_second": 0.225,
      "step": 3000
    },
    {
      "epoch": 0.3885166844677477,
      "grad_norm": 2.171178650336728,
      "learning_rate": 9.590135932037357e-07,
      "loss": 2.7949,
      "step": 3001
    },
    {
      "epoch": 0.38864614687510113,
      "grad_norm": 1.4455267542303654,
      "learning_rate": 9.587748047343934e-07,
      "loss": 2.1575,
      "step": 3002
    },
    {
      "epoch": 0.3887756092824546,
      "grad_norm": 1.5364013753655184,
      "learning_rate": 9.58535948084782e-07,
      "loss": 2.4132,
      "step": 3003
    },
    {
      "epoch": 0.3889050716898081,
      "grad_norm": 1.7077560519608992,
      "learning_rate": 9.582970233036928e-07,
      "loss": 2.368,
      "step": 3004
    },
    {
      "epoch": 0.38903453409716154,
      "grad_norm": 1.5600897361404518,
      "learning_rate": 9.580580304399311e-07,
      "loss": 2.5352,
      "step": 3005
    },
    {
      "epoch": 0.38903453409716154,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.655961513519287,
      "eval_runtime": 13.3538,
      "eval_samples_per_second": 3.295,
      "eval_steps_per_second": 0.225,
      "step": 3005
    },
    {
      "epoch": 0.389163996504515,
      "grad_norm": 3.117618267730982,
      "learning_rate": 9.578189695423157e-07,
      "loss": 2.6108,
      "step": 3006
    },
    {
      "epoch": 0.38929345891186845,
      "grad_norm": 1.5244498534377757,
      "learning_rate": 9.575798406596797e-07,
      "loss": 2.5015,
      "step": 3007
    },
    {
      "epoch": 0.3894229213192219,
      "grad_norm": 2.149137809617406,
      "learning_rate": 9.573406438408706e-07,
      "loss": 2.5037,
      "step": 3008
    },
    {
      "epoch": 0.3895523837265754,
      "grad_norm": 1.7104875842506515,
      "learning_rate": 9.571013791347487e-07,
      "loss": 2.4773,
      "step": 3009
    },
    {
      "epoch": 0.38968184613392887,
      "grad_norm": 2.0984190124201585,
      "learning_rate": 9.568620465901885e-07,
      "loss": 2.7288,
      "step": 3010
    },
    {
      "epoch": 0.38968184613392887,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6532981395721436,
      "eval_runtime": 14.4558,
      "eval_samples_per_second": 3.044,
      "eval_steps_per_second": 0.208,
      "step": 3010
    },
    {
      "epoch": 0.3898113085412823,
      "grad_norm": 3.1110778131066494,
      "learning_rate": 9.566226462560787e-07,
      "loss": 2.6516,
      "step": 3011
    },
    {
      "epoch": 0.3899407709486358,
      "grad_norm": 2.8595897648197517,
      "learning_rate": 9.563831781813216e-07,
      "loss": 2.6724,
      "step": 3012
    },
    {
      "epoch": 0.3900702333559893,
      "grad_norm": 3.9573390702001343,
      "learning_rate": 9.561436424148337e-07,
      "loss": 2.3965,
      "step": 3013
    },
    {
      "epoch": 0.39019969576334274,
      "grad_norm": 2.556971687568334,
      "learning_rate": 9.559040390055449e-07,
      "loss": 2.5435,
      "step": 3014
    },
    {
      "epoch": 0.3903291581706962,
      "grad_norm": 3.613002590592715,
      "learning_rate": 9.556643680023992e-07,
      "loss": 2.4116,
      "step": 3015
    },
    {
      "epoch": 0.3903291581706962,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6519221067428589,
      "eval_runtime": 14.7133,
      "eval_samples_per_second": 2.99,
      "eval_steps_per_second": 0.204,
      "step": 3015
    },
    {
      "epoch": 0.39045862057804964,
      "grad_norm": 1.558538799388353,
      "learning_rate": 9.554246294543536e-07,
      "loss": 2.3438,
      "step": 3016
    },
    {
      "epoch": 0.3905880829854031,
      "grad_norm": 1.6024890999740347,
      "learning_rate": 9.551848234103803e-07,
      "loss": 2.321,
      "step": 3017
    },
    {
      "epoch": 0.3907175453927566,
      "grad_norm": 1.5739377191691255,
      "learning_rate": 9.549449499194645e-07,
      "loss": 2.3508,
      "step": 3018
    },
    {
      "epoch": 0.39084700780011006,
      "grad_norm": 3.4725995709644546,
      "learning_rate": 9.547050090306048e-07,
      "loss": 3.0415,
      "step": 3019
    },
    {
      "epoch": 0.3909764702074635,
      "grad_norm": 1.8670577685348317,
      "learning_rate": 9.544650007928142e-07,
      "loss": 2.4199,
      "step": 3020
    },
    {
      "epoch": 0.3909764702074635,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6511008739471436,
      "eval_runtime": 15.0301,
      "eval_samples_per_second": 2.927,
      "eval_steps_per_second": 0.2,
      "step": 3020
    },
    {
      "epoch": 0.39110593261481696,
      "grad_norm": 3.1261561040662413,
      "learning_rate": 9.542249252551191e-07,
      "loss": 2.8362,
      "step": 3021
    },
    {
      "epoch": 0.3912353950221704,
      "grad_norm": 2.451160499309691,
      "learning_rate": 9.5398478246656e-07,
      "loss": 2.6284,
      "step": 3022
    },
    {
      "epoch": 0.3913648574295239,
      "grad_norm": 3.861183752075475,
      "learning_rate": 9.537445724761909e-07,
      "loss": 2.9661,
      "step": 3023
    },
    {
      "epoch": 0.3914943198368774,
      "grad_norm": 2.485001916311851,
      "learning_rate": 9.535042953330796e-07,
      "loss": 2.8076,
      "step": 3024
    },
    {
      "epoch": 0.39162378224423083,
      "grad_norm": 1.9943353752441344,
      "learning_rate": 9.532639510863073e-07,
      "loss": 2.3853,
      "step": 3025
    },
    {
      "epoch": 0.39162378224423083,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6526545286178589,
      "eval_runtime": 21.7862,
      "eval_samples_per_second": 2.02,
      "eval_steps_per_second": 0.138,
      "step": 3025
    },
    {
      "epoch": 0.3917532446515843,
      "grad_norm": 1.8786943520749657,
      "learning_rate": 9.530235397849696e-07,
      "loss": 2.7849,
      "step": 3026
    },
    {
      "epoch": 0.39188270705893774,
      "grad_norm": 1.395797474944203,
      "learning_rate": 9.527830614781751e-07,
      "loss": 2.5081,
      "step": 3027
    },
    {
      "epoch": 0.39201216946629125,
      "grad_norm": 2.217177230705167,
      "learning_rate": 9.525425162150465e-07,
      "loss": 2.4597,
      "step": 3028
    },
    {
      "epoch": 0.3921416318736447,
      "grad_norm": 1.0821303446695993,
      "learning_rate": 9.523019040447198e-07,
      "loss": 2.3064,
      "step": 3029
    },
    {
      "epoch": 0.39227109428099816,
      "grad_norm": 1.3666860412646131,
      "learning_rate": 9.520612250163451e-07,
      "loss": 2.3789,
      "step": 3030
    },
    {
      "epoch": 0.39227109428099816,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6512340307235718,
      "eval_runtime": 20.8251,
      "eval_samples_per_second": 2.113,
      "eval_steps_per_second": 0.144,
      "step": 3030
    },
    {
      "epoch": 0.3924005566883516,
      "grad_norm": 3.202200381584669,
      "learning_rate": 9.518204791790861e-07,
      "loss": 2.6785,
      "step": 3031
    },
    {
      "epoch": 0.39253001909570506,
      "grad_norm": 6.51169773920298,
      "learning_rate": 9.515796665821197e-07,
      "loss": 2.7849,
      "step": 3032
    },
    {
      "epoch": 0.39265948150305857,
      "grad_norm": 1.4196275999762826,
      "learning_rate": 9.513387872746371e-07,
      "loss": 2.46,
      "step": 3033
    },
    {
      "epoch": 0.392788943910412,
      "grad_norm": 1.8440774219857519,
      "learning_rate": 9.510978413058426e-07,
      "loss": 2.7456,
      "step": 3034
    },
    {
      "epoch": 0.3929184063177655,
      "grad_norm": 2.2258732752178023,
      "learning_rate": 9.508568287249542e-07,
      "loss": 2.7627,
      "step": 3035
    },
    {
      "epoch": 0.3929184063177655,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6516779661178589,
      "eval_runtime": 14.6401,
      "eval_samples_per_second": 3.005,
      "eval_steps_per_second": 0.205,
      "step": 3035
    },
    {
      "epoch": 0.39304786872511893,
      "grad_norm": 4.526151102252629,
      "learning_rate": 9.506157495812037e-07,
      "loss": 3.0513,
      "step": 3036
    },
    {
      "epoch": 0.3931773311324724,
      "grad_norm": 1.5664132059307387,
      "learning_rate": 9.503746039238368e-07,
      "loss": 2.6113,
      "step": 3037
    },
    {
      "epoch": 0.3933067935398259,
      "grad_norm": 1.6605676456023712,
      "learning_rate": 9.501333918021116e-07,
      "loss": 2.3708,
      "step": 3038
    },
    {
      "epoch": 0.39343625594717935,
      "grad_norm": 2.1685045803081646,
      "learning_rate": 9.498921132653011e-07,
      "loss": 2.9172,
      "step": 3039
    },
    {
      "epoch": 0.3935657183545328,
      "grad_norm": 1.869239013237251,
      "learning_rate": 9.496507683626913e-07,
      "loss": 2.5569,
      "step": 3040
    },
    {
      "epoch": 0.3935657183545328,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.649702548980713,
      "eval_runtime": 14.7449,
      "eval_samples_per_second": 2.984,
      "eval_steps_per_second": 0.203,
      "step": 3040
    },
    {
      "epoch": 0.39369518076188625,
      "grad_norm": 1.9271257473699925,
      "learning_rate": 9.494093571435814e-07,
      "loss": 2.5493,
      "step": 3041
    },
    {
      "epoch": 0.3938246431692397,
      "grad_norm": 2.2037140668267723,
      "learning_rate": 9.491678796572851e-07,
      "loss": 2.4985,
      "step": 3042
    },
    {
      "epoch": 0.3939541055765932,
      "grad_norm": 2.7915996843622084,
      "learning_rate": 9.489263359531286e-07,
      "loss": 2.7314,
      "step": 3043
    },
    {
      "epoch": 0.39408356798394667,
      "grad_norm": 5.082239798908639,
      "learning_rate": 9.486847260804524e-07,
      "loss": 2.7034,
      "step": 3044
    },
    {
      "epoch": 0.3942130303913001,
      "grad_norm": 2.07919215026421,
      "learning_rate": 9.484430500886101e-07,
      "loss": 2.4578,
      "step": 3045
    },
    {
      "epoch": 0.3942130303913001,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6489479541778564,
      "eval_runtime": 16.0474,
      "eval_samples_per_second": 2.742,
      "eval_steps_per_second": 0.187,
      "step": 3045
    },
    {
      "epoch": 0.3943424927986536,
      "grad_norm": 2.0770259414754824,
      "learning_rate": 9.48201308026969e-07,
      "loss": 2.5146,
      "step": 3046
    },
    {
      "epoch": 0.39447195520600703,
      "grad_norm": 1.9094743532085647,
      "learning_rate": 9.479594999449096e-07,
      "loss": 2.6597,
      "step": 3047
    },
    {
      "epoch": 0.39460141761336054,
      "grad_norm": 2.166069735039154,
      "learning_rate": 9.477176258918264e-07,
      "loss": 2.5308,
      "step": 3048
    },
    {
      "epoch": 0.394730880020714,
      "grad_norm": 1.7517625118258298,
      "learning_rate": 9.474756859171268e-07,
      "loss": 2.4011,
      "step": 3049
    },
    {
      "epoch": 0.39486034242806745,
      "grad_norm": 2.0901260404474318,
      "learning_rate": 9.47233680070232e-07,
      "loss": 2.5554,
      "step": 3050
    },
    {
      "epoch": 0.39486034242806745,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.648193359375,
      "eval_runtime": 19.4171,
      "eval_samples_per_second": 2.266,
      "eval_steps_per_second": 0.155,
      "step": 3050
    },
    {
      "epoch": 0.3949898048354209,
      "grad_norm": 1.4068251970442467,
      "learning_rate": 9.469916084005771e-07,
      "loss": 2.4116,
      "step": 3051
    },
    {
      "epoch": 0.39511926724277435,
      "grad_norm": 1.4248039613628964,
      "learning_rate": 9.467494709576091e-07,
      "loss": 2.4207,
      "step": 3052
    },
    {
      "epoch": 0.39524872965012786,
      "grad_norm": 2.0154883756342667,
      "learning_rate": 9.465072677907903e-07,
      "loss": 2.4954,
      "step": 3053
    },
    {
      "epoch": 0.3953781920574813,
      "grad_norm": 2.2741005264310097,
      "learning_rate": 9.462649989495956e-07,
      "loss": 2.8135,
      "step": 3054
    },
    {
      "epoch": 0.39550765446483477,
      "grad_norm": 2.1056246453063996,
      "learning_rate": 9.46022664483513e-07,
      "loss": 2.6316,
      "step": 3055
    },
    {
      "epoch": 0.39550765446483477,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6426669359207153,
      "eval_runtime": 15.5456,
      "eval_samples_per_second": 2.83,
      "eval_steps_per_second": 0.193,
      "step": 3055
    },
    {
      "epoch": 0.3956371168721882,
      "grad_norm": 2.4372041345882907,
      "learning_rate": 9.457802644420444e-07,
      "loss": 2.3853,
      "step": 3056
    },
    {
      "epoch": 0.39576657927954173,
      "grad_norm": 2.6901249077029696,
      "learning_rate": 9.455377988747047e-07,
      "loss": 2.7087,
      "step": 3057
    },
    {
      "epoch": 0.3958960416868952,
      "grad_norm": 1.5380116052703254,
      "learning_rate": 9.452952678310226e-07,
      "loss": 2.4177,
      "step": 3058
    },
    {
      "epoch": 0.39602550409424864,
      "grad_norm": 1.2958424771716672,
      "learning_rate": 9.450526713605402e-07,
      "loss": 2.1733,
      "step": 3059
    },
    {
      "epoch": 0.3961549665016021,
      "grad_norm": 2.451527162775795,
      "learning_rate": 9.448100095128124e-07,
      "loss": 2.6582,
      "step": 3060
    },
    {
      "epoch": 0.3961549665016021,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6427334547042847,
      "eval_runtime": 14.6392,
      "eval_samples_per_second": 3.006,
      "eval_steps_per_second": 0.205,
      "step": 3060
    },
    {
      "epoch": 0.39628442890895554,
      "grad_norm": 4.451044959814333,
      "learning_rate": 9.445672823374078e-07,
      "loss": 2.6443,
      "step": 3061
    },
    {
      "epoch": 0.39641389131630905,
      "grad_norm": 2.0231162149575024,
      "learning_rate": 9.443244898839085e-07,
      "loss": 2.5566,
      "step": 3062
    },
    {
      "epoch": 0.3965433537236625,
      "grad_norm": 3.6134534339377073,
      "learning_rate": 9.440816322019097e-07,
      "loss": 3.1411,
      "step": 3063
    },
    {
      "epoch": 0.39667281613101596,
      "grad_norm": 2.5762787718822393,
      "learning_rate": 9.438387093410201e-07,
      "loss": 2.9038,
      "step": 3064
    },
    {
      "epoch": 0.3968022785383694,
      "grad_norm": 2.222180988623937,
      "learning_rate": 9.435957213508615e-07,
      "loss": 2.6431,
      "step": 3065
    },
    {
      "epoch": 0.3968022785383694,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6433993577957153,
      "eval_runtime": 14.8405,
      "eval_samples_per_second": 2.965,
      "eval_steps_per_second": 0.202,
      "step": 3065
    },
    {
      "epoch": 0.39693174094572287,
      "grad_norm": 2.062054458392921,
      "learning_rate": 9.43352668281069e-07,
      "loss": 2.5671,
      "step": 3066
    },
    {
      "epoch": 0.3970612033530764,
      "grad_norm": 2.2504375012902216,
      "learning_rate": 9.431095501812915e-07,
      "loss": 2.5828,
      "step": 3067
    },
    {
      "epoch": 0.39719066576042983,
      "grad_norm": 2.787612981836153,
      "learning_rate": 9.428663671011907e-07,
      "loss": 2.4934,
      "step": 3068
    },
    {
      "epoch": 0.3973201281677833,
      "grad_norm": 1.6814451247588047,
      "learning_rate": 9.426231190904414e-07,
      "loss": 2.478,
      "step": 3069
    },
    {
      "epoch": 0.39744959057513674,
      "grad_norm": 3.913986592034534,
      "learning_rate": 9.423798061987321e-07,
      "loss": 3.0737,
      "step": 3070
    },
    {
      "epoch": 0.39744959057513674,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6434437036514282,
      "eval_runtime": 14.0731,
      "eval_samples_per_second": 3.127,
      "eval_steps_per_second": 0.213,
      "step": 3070
    },
    {
      "epoch": 0.3975790529824902,
      "grad_norm": 4.0420543144703736,
      "learning_rate": 9.421364284757644e-07,
      "loss": 2.5369,
      "step": 3071
    },
    {
      "epoch": 0.3977085153898437,
      "grad_norm": 1.981399142103187,
      "learning_rate": 9.418929859712531e-07,
      "loss": 2.4753,
      "step": 3072
    },
    {
      "epoch": 0.39783797779719715,
      "grad_norm": 2.6473778429723227,
      "learning_rate": 9.416494787349265e-07,
      "loss": 3.0818,
      "step": 3073
    },
    {
      "epoch": 0.3979674402045506,
      "grad_norm": 2.2789585146723894,
      "learning_rate": 9.414059068165254e-07,
      "loss": 2.4092,
      "step": 3074
    },
    {
      "epoch": 0.39809690261190406,
      "grad_norm": 4.9804698868183985,
      "learning_rate": 9.411622702658045e-07,
      "loss": 3.074,
      "step": 3075
    },
    {
      "epoch": 0.39809690261190406,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6427112817764282,
      "eval_runtime": 15.6032,
      "eval_samples_per_second": 2.82,
      "eval_steps_per_second": 0.192,
      "step": 3075
    },
    {
      "epoch": 0.3982263650192575,
      "grad_norm": 2.0084135170836426,
      "learning_rate": 9.409185691325317e-07,
      "loss": 2.554,
      "step": 3076
    },
    {
      "epoch": 0.398355827426611,
      "grad_norm": 4.293940724585163,
      "learning_rate": 9.40674803466488e-07,
      "loss": 2.9316,
      "step": 3077
    },
    {
      "epoch": 0.3984852898339645,
      "grad_norm": 1.7772927181216995,
      "learning_rate": 9.40430973317467e-07,
      "loss": 2.3558,
      "step": 3078
    },
    {
      "epoch": 0.3986147522413179,
      "grad_norm": 1.8421636161400678,
      "learning_rate": 9.401870787352763e-07,
      "loss": 2.5823,
      "step": 3079
    },
    {
      "epoch": 0.3987442146486714,
      "grad_norm": 2.2826887420234674,
      "learning_rate": 9.399431197697361e-07,
      "loss": 2.7739,
      "step": 3080
    },
    {
      "epoch": 0.3987442146486714,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6447087526321411,
      "eval_runtime": 14.1136,
      "eval_samples_per_second": 3.118,
      "eval_steps_per_second": 0.213,
      "step": 3080
    },
    {
      "epoch": 0.39887367705602483,
      "grad_norm": 2.1002790638395177,
      "learning_rate": 9.396990964706802e-07,
      "loss": 2.4294,
      "step": 3081
    },
    {
      "epoch": 0.39900313946337834,
      "grad_norm": 1.6863920592929518,
      "learning_rate": 9.394550088879554e-07,
      "loss": 2.2727,
      "step": 3082
    },
    {
      "epoch": 0.3991326018707318,
      "grad_norm": 2.226894359453203,
      "learning_rate": 9.392108570714209e-07,
      "loss": 2.3413,
      "step": 3083
    },
    {
      "epoch": 0.39926206427808525,
      "grad_norm": 1.8551535970047062,
      "learning_rate": 9.389666410709503e-07,
      "loss": 2.5388,
      "step": 3084
    },
    {
      "epoch": 0.3993915266854387,
      "grad_norm": 2.763576814244469,
      "learning_rate": 9.387223609364295e-07,
      "loss": 2.4541,
      "step": 3085
    },
    {
      "epoch": 0.3993915266854387,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.646928310394287,
      "eval_runtime": 13.484,
      "eval_samples_per_second": 3.263,
      "eval_steps_per_second": 0.222,
      "step": 3085
    },
    {
      "epoch": 0.39952098909279216,
      "grad_norm": 1.7293047120687537,
      "learning_rate": 9.384780167177575e-07,
      "loss": 2.4565,
      "step": 3086
    },
    {
      "epoch": 0.39965045150014566,
      "grad_norm": 2.499157859208796,
      "learning_rate": 9.382336084648467e-07,
      "loss": 2.7568,
      "step": 3087
    },
    {
      "epoch": 0.3997799139074991,
      "grad_norm": 2.4906710804355856,
      "learning_rate": 9.379891362276224e-07,
      "loss": 2.7517,
      "step": 3088
    },
    {
      "epoch": 0.39990937631485257,
      "grad_norm": 2.6127722898284054,
      "learning_rate": 9.377446000560231e-07,
      "loss": 2.6042,
      "step": 3089
    },
    {
      "epoch": 0.400038838722206,
      "grad_norm": 2.4247062554391317,
      "learning_rate": 9.375000000000001e-07,
      "loss": 2.7288,
      "step": 3090
    },
    {
      "epoch": 0.400038838722206,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6405361890792847,
      "eval_runtime": 14.954,
      "eval_samples_per_second": 2.942,
      "eval_steps_per_second": 0.201,
      "step": 3090
    },
    {
      "epoch": 0.4001683011295595,
      "grad_norm": 3.302854678868893,
      "learning_rate": 9.372553361095181e-07,
      "loss": 2.7615,
      "step": 3091
    },
    {
      "epoch": 0.400297763536913,
      "grad_norm": 2.477649676457822,
      "learning_rate": 9.370106084345545e-07,
      "loss": 2.655,
      "step": 3092
    },
    {
      "epoch": 0.40042722594426644,
      "grad_norm": 3.4201098845011346,
      "learning_rate": 9.367658170251e-07,
      "loss": 2.9888,
      "step": 3093
    },
    {
      "epoch": 0.4005566883516199,
      "grad_norm": 2.4589305440829006,
      "learning_rate": 9.365209619311581e-07,
      "loss": 2.6514,
      "step": 3094
    },
    {
      "epoch": 0.40068615075897335,
      "grad_norm": 4.229198532421323,
      "learning_rate": 9.362760432027456e-07,
      "loss": 2.6074,
      "step": 3095
    },
    {
      "epoch": 0.40068615075897335,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6382724046707153,
      "eval_runtime": 16.6591,
      "eval_samples_per_second": 2.641,
      "eval_steps_per_second": 0.18,
      "step": 3095
    },
    {
      "epoch": 0.4008156131663268,
      "grad_norm": 2.4787746453106436,
      "learning_rate": 9.36031060889892e-07,
      "loss": 2.7883,
      "step": 3096
    },
    {
      "epoch": 0.4009450755736803,
      "grad_norm": 2.7178854594092137,
      "learning_rate": 9.357860150426397e-07,
      "loss": 2.728,
      "step": 3097
    },
    {
      "epoch": 0.40107453798103376,
      "grad_norm": 1.872971907836054,
      "learning_rate": 9.355409057110447e-07,
      "loss": 2.5439,
      "step": 3098
    },
    {
      "epoch": 0.4012040003883872,
      "grad_norm": 2.2670914242792937,
      "learning_rate": 9.352957329451752e-07,
      "loss": 2.6538,
      "step": 3099
    },
    {
      "epoch": 0.40133346279574067,
      "grad_norm": 1.5071107420723142,
      "learning_rate": 9.35050496795113e-07,
      "loss": 2.3447,
      "step": 3100
    },
    {
      "epoch": 0.40133346279574067,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.64111328125,
      "eval_runtime": 16.8423,
      "eval_samples_per_second": 2.612,
      "eval_steps_per_second": 0.178,
      "step": 3100
    },
    {
      "epoch": 0.4014629252030942,
      "grad_norm": 1.8478456067846563,
      "learning_rate": 9.348051973109522e-07,
      "loss": 2.4612,
      "step": 3101
    },
    {
      "epoch": 0.40159238761044763,
      "grad_norm": 1.5693193171751734,
      "learning_rate": 9.345598345428004e-07,
      "loss": 2.4302,
      "step": 3102
    },
    {
      "epoch": 0.4017218500178011,
      "grad_norm": 2.025385736654082,
      "learning_rate": 9.34314408540778e-07,
      "loss": 2.5471,
      "step": 3103
    },
    {
      "epoch": 0.40185131242515454,
      "grad_norm": 1.2568905976090412,
      "learning_rate": 9.340689193550182e-07,
      "loss": 2.4692,
      "step": 3104
    },
    {
      "epoch": 0.401980774832508,
      "grad_norm": 2.9181681902493724,
      "learning_rate": 9.33823367035667e-07,
      "loss": 2.6482,
      "step": 3105
    },
    {
      "epoch": 0.401980774832508,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.638383388519287,
      "eval_runtime": 14.5494,
      "eval_samples_per_second": 3.024,
      "eval_steps_per_second": 0.206,
      "step": 3105
    },
    {
      "epoch": 0.4021102372398615,
      "grad_norm": 3.589858781874059,
      "learning_rate": 9.335777516328833e-07,
      "loss": 2.6055,
      "step": 3106
    },
    {
      "epoch": 0.40223969964721495,
      "grad_norm": 2.2409088019430112,
      "learning_rate": 9.333320731968396e-07,
      "loss": 2.5139,
      "step": 3107
    },
    {
      "epoch": 0.4023691620545684,
      "grad_norm": 1.2046073215318571,
      "learning_rate": 9.330863317777201e-07,
      "loss": 2.2661,
      "step": 3108
    },
    {
      "epoch": 0.40249862446192186,
      "grad_norm": 1.8262955699883103,
      "learning_rate": 9.328405274257228e-07,
      "loss": 2.4307,
      "step": 3109
    },
    {
      "epoch": 0.4026280868692753,
      "grad_norm": 2.300838876277139,
      "learning_rate": 9.325946601910582e-07,
      "loss": 2.4114,
      "step": 3110
    },
    {
      "epoch": 0.4026280868692753,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6368962526321411,
      "eval_runtime": 13.8464,
      "eval_samples_per_second": 3.178,
      "eval_steps_per_second": 0.217,
      "step": 3110
    },
    {
      "epoch": 0.4027575492766288,
      "grad_norm": 14.40113087060862,
      "learning_rate": 9.323487301239495e-07,
      "loss": 3.0903,
      "step": 3111
    },
    {
      "epoch": 0.4028870116839823,
      "grad_norm": 1.8789966606954218,
      "learning_rate": 9.32102737274633e-07,
      "loss": 2.479,
      "step": 3112
    },
    {
      "epoch": 0.40301647409133573,
      "grad_norm": 3.782303856963805,
      "learning_rate": 9.318566816933578e-07,
      "loss": 2.623,
      "step": 3113
    },
    {
      "epoch": 0.4031459364986892,
      "grad_norm": 1.5264590123020203,
      "learning_rate": 9.316105634303856e-07,
      "loss": 2.4453,
      "step": 3114
    },
    {
      "epoch": 0.40327539890604264,
      "grad_norm": 2.663249131366837,
      "learning_rate": 9.313643825359911e-07,
      "loss": 2.6101,
      "step": 3115
    },
    {
      "epoch": 0.40327539890604264,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6377174854278564,
      "eval_runtime": 19.1181,
      "eval_samples_per_second": 2.301,
      "eval_steps_per_second": 0.157,
      "step": 3115
    },
    {
      "epoch": 0.40340486131339615,
      "grad_norm": 1.3755247525337402,
      "learning_rate": 9.311181390604618e-07,
      "loss": 2.4829,
      "step": 3116
    },
    {
      "epoch": 0.4035343237207496,
      "grad_norm": 1.9212891466534978,
      "learning_rate": 9.308718330540978e-07,
      "loss": 2.4246,
      "step": 3117
    },
    {
      "epoch": 0.40366378612810305,
      "grad_norm": 2.544753759419078,
      "learning_rate": 9.306254645672122e-07,
      "loss": 2.615,
      "step": 3118
    },
    {
      "epoch": 0.4037932485354565,
      "grad_norm": 1.9523295145421204,
      "learning_rate": 9.303790336501307e-07,
      "loss": 2.6646,
      "step": 3119
    },
    {
      "epoch": 0.40392271094280996,
      "grad_norm": 1.416475618164198,
      "learning_rate": 9.301325403531918e-07,
      "loss": 2.3164,
      "step": 3120
    },
    {
      "epoch": 0.40392271094280996,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6353648900985718,
      "eval_runtime": 14.8817,
      "eval_samples_per_second": 2.957,
      "eval_steps_per_second": 0.202,
      "step": 3120
    },
    {
      "epoch": 0.40405217335016347,
      "grad_norm": 2.3078047604522482,
      "learning_rate": 9.298859847267469e-07,
      "loss": 2.4058,
      "step": 3121
    },
    {
      "epoch": 0.4041816357575169,
      "grad_norm": 9.388826804965841,
      "learning_rate": 9.296393668211596e-07,
      "loss": 2.8711,
      "step": 3122
    },
    {
      "epoch": 0.4043110981648704,
      "grad_norm": 2.0987676802321404,
      "learning_rate": 9.293926866868072e-07,
      "loss": 2.5203,
      "step": 3123
    },
    {
      "epoch": 0.40444056057222383,
      "grad_norm": 2.255919448500308,
      "learning_rate": 9.291459443740786e-07,
      "loss": 2.4963,
      "step": 3124
    },
    {
      "epoch": 0.4045700229795773,
      "grad_norm": 4.102169383890809,
      "learning_rate": 9.288991399333761e-07,
      "loss": 2.9155,
      "step": 3125
    },
    {
      "epoch": 0.4045700229795773,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6348544359207153,
      "eval_runtime": 13.892,
      "eval_samples_per_second": 3.167,
      "eval_steps_per_second": 0.216,
      "step": 3125
    },
    {
      "epoch": 0.4046994853869308,
      "grad_norm": 2.3625896025957704,
      "learning_rate": 9.286522734151146e-07,
      "loss": 2.4388,
      "step": 3126
    },
    {
      "epoch": 0.40482894779428424,
      "grad_norm": 2.33447697730886,
      "learning_rate": 9.284053448697213e-07,
      "loss": 2.6108,
      "step": 3127
    },
    {
      "epoch": 0.4049584102016377,
      "grad_norm": 4.5414991479556095,
      "learning_rate": 9.281583543476368e-07,
      "loss": 2.8462,
      "step": 3128
    },
    {
      "epoch": 0.40508787260899115,
      "grad_norm": 1.8421529376703967,
      "learning_rate": 9.279113018993136e-07,
      "loss": 2.5676,
      "step": 3129
    },
    {
      "epoch": 0.4052173350163446,
      "grad_norm": 1.7697452402571285,
      "learning_rate": 9.276641875752171e-07,
      "loss": 2.6553,
      "step": 3130
    },
    {
      "epoch": 0.4052173350163446,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.638427734375,
      "eval_runtime": 23.0472,
      "eval_samples_per_second": 1.909,
      "eval_steps_per_second": 0.13,
      "step": 3130
    },
    {
      "epoch": 0.4053467974236981,
      "grad_norm": 2.0337708679645736,
      "learning_rate": 9.274170114258258e-07,
      "loss": 2.4209,
      "step": 3131
    },
    {
      "epoch": 0.40547625983105157,
      "grad_norm": 1.66956425256569,
      "learning_rate": 9.271697735016302e-07,
      "loss": 2.5112,
      "step": 3132
    },
    {
      "epoch": 0.405605722238405,
      "grad_norm": 1.2457792927470352,
      "learning_rate": 9.269224738531335e-07,
      "loss": 2.3901,
      "step": 3133
    },
    {
      "epoch": 0.4057351846457585,
      "grad_norm": 1.63242837213429,
      "learning_rate": 9.266751125308518e-07,
      "loss": 2.4534,
      "step": 3134
    },
    {
      "epoch": 0.4058646470531119,
      "grad_norm": 1.8314878495342952,
      "learning_rate": 9.264276895853138e-07,
      "loss": 2.3594,
      "step": 3135
    },
    {
      "epoch": 0.4058646470531119,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6320579051971436,
      "eval_runtime": 14.5662,
      "eval_samples_per_second": 3.021,
      "eval_steps_per_second": 0.206,
      "step": 3135
    },
    {
      "epoch": 0.40599410946046544,
      "grad_norm": 2.6779428074927383,
      "learning_rate": 9.261802050670606e-07,
      "loss": 2.8491,
      "step": 3136
    },
    {
      "epoch": 0.4061235718678189,
      "grad_norm": 3.7541144768700803,
      "learning_rate": 9.259326590266457e-07,
      "loss": 2.5808,
      "step": 3137
    },
    {
      "epoch": 0.40625303427517234,
      "grad_norm": 1.2850824932258837,
      "learning_rate": 9.256850515146357e-07,
      "loss": 2.5637,
      "step": 3138
    },
    {
      "epoch": 0.4063824966825258,
      "grad_norm": 2.6325487961655214,
      "learning_rate": 9.254373825816093e-07,
      "loss": 2.5239,
      "step": 3139
    },
    {
      "epoch": 0.40651195908987925,
      "grad_norm": 1.7155327463440555,
      "learning_rate": 9.25189652278158e-07,
      "loss": 2.5234,
      "step": 3140
    },
    {
      "epoch": 0.40651195908987925,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6346768140792847,
      "eval_runtime": 16.4256,
      "eval_samples_per_second": 2.679,
      "eval_steps_per_second": 0.183,
      "step": 3140
    },
    {
      "epoch": 0.40664142149723276,
      "grad_norm": 2.6385478954307082,
      "learning_rate": 9.249418606548855e-07,
      "loss": 2.6528,
      "step": 3141
    },
    {
      "epoch": 0.4067708839045862,
      "grad_norm": 2.1806615397337565,
      "learning_rate": 9.246940077624085e-07,
      "loss": 2.6929,
      "step": 3142
    },
    {
      "epoch": 0.40690034631193966,
      "grad_norm": 1.9939039450824454,
      "learning_rate": 9.244460936513557e-07,
      "loss": 2.6274,
      "step": 3143
    },
    {
      "epoch": 0.4070298087192931,
      "grad_norm": 1.5289750685486185,
      "learning_rate": 9.241981183723691e-07,
      "loss": 2.6245,
      "step": 3144
    },
    {
      "epoch": 0.4071592711266466,
      "grad_norm": 1.8233159160424126,
      "learning_rate": 9.239500819761022e-07,
      "loss": 2.6013,
      "step": 3145
    },
    {
      "epoch": 0.4071592711266466,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6326793432235718,
      "eval_runtime": 14.4513,
      "eval_samples_per_second": 3.045,
      "eval_steps_per_second": 0.208,
      "step": 3145
    },
    {
      "epoch": 0.4072887335340001,
      "grad_norm": 2.5374459117830663,
      "learning_rate": 9.237019845132216e-07,
      "loss": 2.6333,
      "step": 3146
    },
    {
      "epoch": 0.40741819594135353,
      "grad_norm": 1.6299393691109445,
      "learning_rate": 9.234538260344061e-07,
      "loss": 2.5466,
      "step": 3147
    },
    {
      "epoch": 0.407547658348707,
      "grad_norm": 2.676433195955329,
      "learning_rate": 9.232056065903476e-07,
      "loss": 2.7439,
      "step": 3148
    },
    {
      "epoch": 0.40767712075606044,
      "grad_norm": 1.9111031935904645,
      "learning_rate": 9.229573262317494e-07,
      "loss": 2.6409,
      "step": 3149
    },
    {
      "epoch": 0.40780658316341395,
      "grad_norm": 2.119199479842497,
      "learning_rate": 9.227089850093279e-07,
      "loss": 2.3306,
      "step": 3150
    },
    {
      "epoch": 0.40780658316341395,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6363414525985718,
      "eval_runtime": 14.2686,
      "eval_samples_per_second": 3.084,
      "eval_steps_per_second": 0.21,
      "step": 3150
    },
    {
      "epoch": 0.4079360455707674,
      "grad_norm": 2.306589793950446,
      "learning_rate": 9.22460582973812e-07,
      "loss": 2.2992,
      "step": 3151
    },
    {
      "epoch": 0.40806550797812086,
      "grad_norm": 2.5095156128091207,
      "learning_rate": 9.222121201759429e-07,
      "loss": 2.7454,
      "step": 3152
    },
    {
      "epoch": 0.4081949703854743,
      "grad_norm": 1.7471666423852787,
      "learning_rate": 9.219635966664739e-07,
      "loss": 2.4064,
      "step": 3153
    },
    {
      "epoch": 0.40832443279282776,
      "grad_norm": 2.346173468590584,
      "learning_rate": 9.217150124961713e-07,
      "loss": 2.7056,
      "step": 3154
    },
    {
      "epoch": 0.40845389520018127,
      "grad_norm": 1.8827726403951508,
      "learning_rate": 9.214663677158131e-07,
      "loss": 2.3745,
      "step": 3155
    },
    {
      "epoch": 0.40845389520018127,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6353427171707153,
      "eval_runtime": 21.5011,
      "eval_samples_per_second": 2.046,
      "eval_steps_per_second": 0.14,
      "step": 3155
    },
    {
      "epoch": 0.4085833576075347,
      "grad_norm": 3.1129255407423386,
      "learning_rate": 9.212176623761901e-07,
      "loss": 2.6379,
      "step": 3156
    },
    {
      "epoch": 0.4087128200148882,
      "grad_norm": 2.6454766740203417,
      "learning_rate": 9.209688965281057e-07,
      "loss": 2.6465,
      "step": 3157
    },
    {
      "epoch": 0.40884228242224163,
      "grad_norm": 3.0469970888490097,
      "learning_rate": 9.20720070222375e-07,
      "loss": 2.7793,
      "step": 3158
    },
    {
      "epoch": 0.4089717448295951,
      "grad_norm": 1.6882128429732328,
      "learning_rate": 9.20471183509826e-07,
      "loss": 2.4279,
      "step": 3159
    },
    {
      "epoch": 0.4091012072369486,
      "grad_norm": 2.0014236740102196,
      "learning_rate": 9.20222236441299e-07,
      "loss": 2.5684,
      "step": 3160
    },
    {
      "epoch": 0.4091012072369486,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6346324682235718,
      "eval_runtime": 15.146,
      "eval_samples_per_second": 2.905,
      "eval_steps_per_second": 0.198,
      "step": 3160
    },
    {
      "epoch": 0.40923066964430205,
      "grad_norm": 1.6349466783922244,
      "learning_rate": 9.199732290676462e-07,
      "loss": 2.249,
      "step": 3161
    },
    {
      "epoch": 0.4093601320516555,
      "grad_norm": 2.421945812427243,
      "learning_rate": 9.197241614397326e-07,
      "loss": 2.6121,
      "step": 3162
    },
    {
      "epoch": 0.40948959445900895,
      "grad_norm": 1.6542954455696464,
      "learning_rate": 9.194750336084351e-07,
      "loss": 2.356,
      "step": 3163
    },
    {
      "epoch": 0.4096190568663624,
      "grad_norm": 1.9871650766107825,
      "learning_rate": 9.192258456246434e-07,
      "loss": 2.6946,
      "step": 3164
    },
    {
      "epoch": 0.4097485192737159,
      "grad_norm": 2.4313645490977867,
      "learning_rate": 9.189765975392589e-07,
      "loss": 2.6299,
      "step": 3165
    },
    {
      "epoch": 0.4097485192737159,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6305264234542847,
      "eval_runtime": 15.5697,
      "eval_samples_per_second": 2.826,
      "eval_steps_per_second": 0.193,
      "step": 3165
    },
    {
      "epoch": 0.40987798168106937,
      "grad_norm": 2.2305933453543796,
      "learning_rate": 9.187272894031959e-07,
      "loss": 2.5015,
      "step": 3166
    },
    {
      "epoch": 0.4100074440884228,
      "grad_norm": 2.265110689173424,
      "learning_rate": 9.184779212673803e-07,
      "loss": 2.5164,
      "step": 3167
    },
    {
      "epoch": 0.4101369064957763,
      "grad_norm": 1.7082441886990973,
      "learning_rate": 9.182284931827509e-07,
      "loss": 2.5229,
      "step": 3168
    },
    {
      "epoch": 0.41026636890312973,
      "grad_norm": 2.488255843895349,
      "learning_rate": 9.179790052002582e-07,
      "loss": 2.9529,
      "step": 3169
    },
    {
      "epoch": 0.41039583131048324,
      "grad_norm": 4.651526710767568,
      "learning_rate": 9.177294573708652e-07,
      "loss": 3.5449,
      "step": 3170
    },
    {
      "epoch": 0.41039583131048324,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6325905323028564,
      "eval_runtime": 14.282,
      "eval_samples_per_second": 3.081,
      "eval_steps_per_second": 0.21,
      "step": 3170
    },
    {
      "epoch": 0.4105252937178367,
      "grad_norm": 2.4930524696703324,
      "learning_rate": 9.174798497455475e-07,
      "loss": 2.5156,
      "step": 3171
    },
    {
      "epoch": 0.41065475612519015,
      "grad_norm": 1.5528227466020093,
      "learning_rate": 9.172301823752919e-07,
      "loss": 2.478,
      "step": 3172
    },
    {
      "epoch": 0.4107842185325436,
      "grad_norm": 2.677157226310644,
      "learning_rate": 9.169804553110983e-07,
      "loss": 2.3484,
      "step": 3173
    },
    {
      "epoch": 0.41091368093989705,
      "grad_norm": 1.9770795272313717,
      "learning_rate": 9.167306686039786e-07,
      "loss": 2.7358,
      "step": 3174
    },
    {
      "epoch": 0.41104314334725056,
      "grad_norm": 3.0029429384104036,
      "learning_rate": 9.164808223049567e-07,
      "loss": 3.1643,
      "step": 3175
    },
    {
      "epoch": 0.41104314334725056,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6319690942764282,
      "eval_runtime": 13.4563,
      "eval_samples_per_second": 3.27,
      "eval_steps_per_second": 0.223,
      "step": 3175
    },
    {
      "epoch": 0.411172605754604,
      "grad_norm": 1.4529269509986145,
      "learning_rate": 9.162309164650689e-07,
      "loss": 2.3882,
      "step": 3176
    },
    {
      "epoch": 0.41130206816195747,
      "grad_norm": 2.138975165999896,
      "learning_rate": 9.159809511353633e-07,
      "loss": 2.5444,
      "step": 3177
    },
    {
      "epoch": 0.4114315305693109,
      "grad_norm": 1.8334174052761953,
      "learning_rate": 9.157309263669006e-07,
      "loss": 2.4827,
      "step": 3178
    },
    {
      "epoch": 0.4115609929766644,
      "grad_norm": 3.53241619656035,
      "learning_rate": 9.154808422107534e-07,
      "loss": 2.3521,
      "step": 3179
    },
    {
      "epoch": 0.4116904553840179,
      "grad_norm": 1.4455666561151648,
      "learning_rate": 9.152306987180062e-07,
      "loss": 2.6204,
      "step": 3180
    },
    {
      "epoch": 0.4116904553840179,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6283291578292847,
      "eval_runtime": 13.5192,
      "eval_samples_per_second": 3.255,
      "eval_steps_per_second": 0.222,
      "step": 3180
    },
    {
      "epoch": 0.41181991779137134,
      "grad_norm": 2.0674732159398754,
      "learning_rate": 9.149804959397561e-07,
      "loss": 2.5933,
      "step": 3181
    },
    {
      "epoch": 0.4119493801987248,
      "grad_norm": 1.6228242051521333,
      "learning_rate": 9.147302339271124e-07,
      "loss": 2.5283,
      "step": 3182
    },
    {
      "epoch": 0.41207884260607824,
      "grad_norm": 4.115468897474255,
      "learning_rate": 9.144799127311954e-07,
      "loss": 2.7917,
      "step": 3183
    },
    {
      "epoch": 0.4122083050134317,
      "grad_norm": 1.7892069658451697,
      "learning_rate": 9.142295324031389e-07,
      "loss": 2.4824,
      "step": 3184
    },
    {
      "epoch": 0.4123377674207852,
      "grad_norm": 1.8130998555584166,
      "learning_rate": 9.139790929940881e-07,
      "loss": 2.4771,
      "step": 3185
    },
    {
      "epoch": 0.4123377674207852,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6305042505264282,
      "eval_runtime": 14.3621,
      "eval_samples_per_second": 3.064,
      "eval_steps_per_second": 0.209,
      "step": 3185
    },
    {
      "epoch": 0.41246722982813866,
      "grad_norm": 1.673273409526274,
      "learning_rate": 9.137285945552002e-07,
      "loss": 2.5122,
      "step": 3186
    },
    {
      "epoch": 0.4125966922354921,
      "grad_norm": 1.565240354966385,
      "learning_rate": 9.134780371376445e-07,
      "loss": 2.5337,
      "step": 3187
    },
    {
      "epoch": 0.41272615464284557,
      "grad_norm": 1.4944363374687455,
      "learning_rate": 9.132274207926025e-07,
      "loss": 2.4297,
      "step": 3188
    },
    {
      "epoch": 0.4128556170501991,
      "grad_norm": 1.3390616257700647,
      "learning_rate": 9.129767455712679e-07,
      "loss": 2.4348,
      "step": 3189
    },
    {
      "epoch": 0.41298507945755253,
      "grad_norm": 1.8215280567863243,
      "learning_rate": 9.127260115248459e-07,
      "loss": 2.3762,
      "step": 3190
    },
    {
      "epoch": 0.41298507945755253,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6277521848678589,
      "eval_runtime": 14.4618,
      "eval_samples_per_second": 3.042,
      "eval_steps_per_second": 0.207,
      "step": 3190
    },
    {
      "epoch": 0.413114541864906,
      "grad_norm": 1.7308143449323479,
      "learning_rate": 9.124752187045541e-07,
      "loss": 2.5685,
      "step": 3191
    },
    {
      "epoch": 0.41324400427225944,
      "grad_norm": 2.353118060280929,
      "learning_rate": 9.122243671616218e-07,
      "loss": 2.8176,
      "step": 3192
    },
    {
      "epoch": 0.4133734666796129,
      "grad_norm": 1.4677428234050347,
      "learning_rate": 9.119734569472911e-07,
      "loss": 2.5789,
      "step": 3193
    },
    {
      "epoch": 0.4135029290869664,
      "grad_norm": 1.5533853174449563,
      "learning_rate": 9.117224881128151e-07,
      "loss": 2.5591,
      "step": 3194
    },
    {
      "epoch": 0.41363239149431985,
      "grad_norm": 1.8359214209643377,
      "learning_rate": 9.114714607094591e-07,
      "loss": 2.6831,
      "step": 3195
    },
    {
      "epoch": 0.41363239149431985,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6283957958221436,
      "eval_runtime": 13.8051,
      "eval_samples_per_second": 3.187,
      "eval_steps_per_second": 0.217,
      "step": 3195
    },
    {
      "epoch": 0.4137618539016733,
      "grad_norm": 2.439111361132348,
      "learning_rate": 9.112203747885007e-07,
      "loss": 2.9541,
      "step": 3196
    },
    {
      "epoch": 0.41389131630902676,
      "grad_norm": 2.2204867672810744,
      "learning_rate": 9.109692304012295e-07,
      "loss": 2.6792,
      "step": 3197
    },
    {
      "epoch": 0.4140207787163802,
      "grad_norm": 1.519237431008716,
      "learning_rate": 9.107180275989466e-07,
      "loss": 2.3406,
      "step": 3198
    },
    {
      "epoch": 0.4141502411237337,
      "grad_norm": 2.9943164027139324,
      "learning_rate": 9.104667664329654e-07,
      "loss": 2.9312,
      "step": 3199
    },
    {
      "epoch": 0.4142797035310872,
      "grad_norm": 3.385237029492058,
      "learning_rate": 9.102154469546109e-07,
      "loss": 2.8723,
      "step": 3200
    },
    {
      "epoch": 0.4142797035310872,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6263316869735718,
      "eval_runtime": 15.6112,
      "eval_samples_per_second": 2.818,
      "eval_steps_per_second": 0.192,
      "step": 3200
    },
    {
      "epoch": 0.4144091659384406,
      "grad_norm": 2.0605726991878366,
      "learning_rate": 9.099640692152203e-07,
      "loss": 2.3472,
      "step": 3201
    },
    {
      "epoch": 0.4145386283457941,
      "grad_norm": 1.9531322408622698,
      "learning_rate": 9.097126332661427e-07,
      "loss": 2.5063,
      "step": 3202
    },
    {
      "epoch": 0.41466809075314753,
      "grad_norm": 1.7827777854279698,
      "learning_rate": 9.094611391587388e-07,
      "loss": 2.6777,
      "step": 3203
    },
    {
      "epoch": 0.41479755316050104,
      "grad_norm": 1.6130381349912482,
      "learning_rate": 9.092095869443814e-07,
      "loss": 2.3628,
      "step": 3204
    },
    {
      "epoch": 0.4149270155678545,
      "grad_norm": 1.436695709810826,
      "learning_rate": 9.089579766744554e-07,
      "loss": 2.3994,
      "step": 3205
    },
    {
      "epoch": 0.4149270155678545,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6241787672042847,
      "eval_runtime": 14.7091,
      "eval_samples_per_second": 2.991,
      "eval_steps_per_second": 0.204,
      "step": 3205
    },
    {
      "epoch": 0.41505647797520795,
      "grad_norm": 2.0319122385459605,
      "learning_rate": 9.08706308400357e-07,
      "loss": 2.792,
      "step": 3206
    },
    {
      "epoch": 0.4151859403825614,
      "grad_norm": 2.1304839797660797,
      "learning_rate": 9.084545821734946e-07,
      "loss": 2.5764,
      "step": 3207
    },
    {
      "epoch": 0.41531540278991486,
      "grad_norm": 3.627352745893683,
      "learning_rate": 9.082027980452884e-07,
      "loss": 2.564,
      "step": 3208
    },
    {
      "epoch": 0.41544486519726836,
      "grad_norm": 1.8001616268210472,
      "learning_rate": 9.079509560671706e-07,
      "loss": 2.4189,
      "step": 3209
    },
    {
      "epoch": 0.4155743276046218,
      "grad_norm": 2.232693616134311,
      "learning_rate": 9.076990562905847e-07,
      "loss": 2.699,
      "step": 3210
    },
    {
      "epoch": 0.4155743276046218,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6234463453292847,
      "eval_runtime": 19.7147,
      "eval_samples_per_second": 2.232,
      "eval_steps_per_second": 0.152,
      "step": 3210
    },
    {
      "epoch": 0.41570379001197527,
      "grad_norm": 1.5920193896463253,
      "learning_rate": 9.074470987669866e-07,
      "loss": 2.4985,
      "step": 3211
    },
    {
      "epoch": 0.4158332524193287,
      "grad_norm": 3.34654975967097,
      "learning_rate": 9.071950835478437e-07,
      "loss": 2.5562,
      "step": 3212
    },
    {
      "epoch": 0.4159627148266822,
      "grad_norm": 2.101812673255751,
      "learning_rate": 9.069430106846349e-07,
      "loss": 2.4448,
      "step": 3213
    },
    {
      "epoch": 0.4160921772340357,
      "grad_norm": 1.528590510355101,
      "learning_rate": 9.066908802288518e-07,
      "loss": 2.2534,
      "step": 3214
    },
    {
      "epoch": 0.41622163964138914,
      "grad_norm": 1.7738671714681336,
      "learning_rate": 9.064386922319965e-07,
      "loss": 2.4614,
      "step": 3215
    },
    {
      "epoch": 0.41622163964138914,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6248224973678589,
      "eval_runtime": 17.8083,
      "eval_samples_per_second": 2.471,
      "eval_steps_per_second": 0.168,
      "step": 3215
    },
    {
      "epoch": 0.4163511020487426,
      "grad_norm": 1.8341224537735463,
      "learning_rate": 9.061864467455839e-07,
      "loss": 2.8481,
      "step": 3216
    },
    {
      "epoch": 0.41648056445609605,
      "grad_norm": 1.7480148092472603,
      "learning_rate": 9.059341438211403e-07,
      "loss": 2.5042,
      "step": 3217
    },
    {
      "epoch": 0.4166100268634495,
      "grad_norm": 1.3938576142248862,
      "learning_rate": 9.056817835102035e-07,
      "loss": 2.5427,
      "step": 3218
    },
    {
      "epoch": 0.416739489270803,
      "grad_norm": 2.7013522583805467,
      "learning_rate": 9.054293658643228e-07,
      "loss": 2.6804,
      "step": 3219
    },
    {
      "epoch": 0.41686895167815646,
      "grad_norm": 7.9317473198485935,
      "learning_rate": 9.051768909350606e-07,
      "loss": 2.9421,
      "step": 3220
    },
    {
      "epoch": 0.41686895167815646,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6247780323028564,
      "eval_runtime": 13.8241,
      "eval_samples_per_second": 3.183,
      "eval_steps_per_second": 0.217,
      "step": 3220
    },
    {
      "epoch": 0.4169984140855099,
      "grad_norm": 1.9709065424588976,
      "learning_rate": 9.049243587739894e-07,
      "loss": 2.8203,
      "step": 3221
    },
    {
      "epoch": 0.41712787649286337,
      "grad_norm": 1.9428330311588493,
      "learning_rate": 9.046717694326942e-07,
      "loss": 2.4033,
      "step": 3222
    },
    {
      "epoch": 0.4172573389002168,
      "grad_norm": 1.6212262976038094,
      "learning_rate": 9.044191229627712e-07,
      "loss": 2.511,
      "step": 3223
    },
    {
      "epoch": 0.41738680130757033,
      "grad_norm": 3.393026697002203,
      "learning_rate": 9.041664194158288e-07,
      "loss": 2.4211,
      "step": 3224
    },
    {
      "epoch": 0.4175162637149238,
      "grad_norm": 2.302173837703524,
      "learning_rate": 9.039136588434869e-07,
      "loss": 2.4407,
      "step": 3225
    },
    {
      "epoch": 0.4175162637149238,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6212047338485718,
      "eval_runtime": 14.0856,
      "eval_samples_per_second": 3.124,
      "eval_steps_per_second": 0.213,
      "step": 3225
    },
    {
      "epoch": 0.41764572612227724,
      "grad_norm": 1.9221139960252982,
      "learning_rate": 9.036608412973768e-07,
      "loss": 2.5552,
      "step": 3226
    },
    {
      "epoch": 0.4177751885296307,
      "grad_norm": 2.476350930165937,
      "learning_rate": 9.034079668291418e-07,
      "loss": 2.5461,
      "step": 3227
    },
    {
      "epoch": 0.41790465093698415,
      "grad_norm": 1.47229003898402,
      "learning_rate": 9.031550354904362e-07,
      "loss": 2.4373,
      "step": 3228
    },
    {
      "epoch": 0.41803411334433765,
      "grad_norm": 2.183663206223865,
      "learning_rate": 9.029020473329268e-07,
      "loss": 2.8606,
      "step": 3229
    },
    {
      "epoch": 0.4181635757516911,
      "grad_norm": 2.1246530382197535,
      "learning_rate": 9.026490024082915e-07,
      "loss": 2.3906,
      "step": 3230
    },
    {
      "epoch": 0.4181635757516911,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6206942796707153,
      "eval_runtime": 14.2383,
      "eval_samples_per_second": 3.09,
      "eval_steps_per_second": 0.211,
      "step": 3230
    },
    {
      "epoch": 0.41829303815904456,
      "grad_norm": 1.9113312190773661,
      "learning_rate": 9.023959007682197e-07,
      "loss": 2.3535,
      "step": 3231
    },
    {
      "epoch": 0.418422500566398,
      "grad_norm": 2.314864199281409,
      "learning_rate": 9.021427424644127e-07,
      "loss": 2.6824,
      "step": 3232
    },
    {
      "epoch": 0.4185519629737515,
      "grad_norm": 1.8557837018790813,
      "learning_rate": 9.018895275485828e-07,
      "loss": 2.603,
      "step": 3233
    },
    {
      "epoch": 0.418681425381105,
      "grad_norm": 1.234498423406555,
      "learning_rate": 9.01636256072455e-07,
      "loss": 2.4275,
      "step": 3234
    },
    {
      "epoch": 0.41881088778845843,
      "grad_norm": 2.1039299027435256,
      "learning_rate": 9.013829280877647e-07,
      "loss": 2.4912,
      "step": 3235
    },
    {
      "epoch": 0.41881088778845843,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6212269067764282,
      "eval_runtime": 13.6099,
      "eval_samples_per_second": 3.233,
      "eval_steps_per_second": 0.22,
      "step": 3235
    },
    {
      "epoch": 0.4189403501958119,
      "grad_norm": 3.4168389852644068,
      "learning_rate": 9.011295436462592e-07,
      "loss": 2.7859,
      "step": 3236
    },
    {
      "epoch": 0.41906981260316534,
      "grad_norm": 2.11613200983246,
      "learning_rate": 9.008761027996973e-07,
      "loss": 2.5972,
      "step": 3237
    },
    {
      "epoch": 0.41919927501051885,
      "grad_norm": 3.345025552839822,
      "learning_rate": 9.006226055998498e-07,
      "loss": 2.5752,
      "step": 3238
    },
    {
      "epoch": 0.4193287374178723,
      "grad_norm": 1.633146160231083,
      "learning_rate": 9.003690520984987e-07,
      "loss": 2.6104,
      "step": 3239
    },
    {
      "epoch": 0.41945819982522575,
      "grad_norm": 2.5802724005372504,
      "learning_rate": 9.001154423474372e-07,
      "loss": 2.5286,
      "step": 3240
    },
    {
      "epoch": 0.41945819982522575,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.622025966644287,
      "eval_runtime": 13.7148,
      "eval_samples_per_second": 3.208,
      "eval_steps_per_second": 0.219,
      "step": 3240
    },
    {
      "epoch": 0.4195876622325792,
      "grad_norm": 2.5670736306364423,
      "learning_rate": 8.9986177639847e-07,
      "loss": 2.5991,
      "step": 3241
    },
    {
      "epoch": 0.41971712463993266,
      "grad_norm": 2.7100968884954866,
      "learning_rate": 8.996080543034139e-07,
      "loss": 2.6877,
      "step": 3242
    },
    {
      "epoch": 0.41984658704728617,
      "grad_norm": 2.2197656405346446,
      "learning_rate": 8.993542761140967e-07,
      "loss": 2.7317,
      "step": 3243
    },
    {
      "epoch": 0.4199760494546396,
      "grad_norm": 2.5189336382272995,
      "learning_rate": 8.991004418823576e-07,
      "loss": 2.8982,
      "step": 3244
    },
    {
      "epoch": 0.4201055118619931,
      "grad_norm": 2.2333111446540888,
      "learning_rate": 8.988465516600476e-07,
      "loss": 2.8354,
      "step": 3245
    },
    {
      "epoch": 0.4201055118619931,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6212047338485718,
      "eval_runtime": 14.9291,
      "eval_samples_per_second": 2.947,
      "eval_steps_per_second": 0.201,
      "step": 3245
    },
    {
      "epoch": 0.42023497426934653,
      "grad_norm": 3.573335572794352,
      "learning_rate": 8.985926054990287e-07,
      "loss": 2.5356,
      "step": 3246
    },
    {
      "epoch": 0.4203644366767,
      "grad_norm": 1.8581797741235633,
      "learning_rate": 8.983386034511748e-07,
      "loss": 2.5354,
      "step": 3247
    },
    {
      "epoch": 0.4204938990840535,
      "grad_norm": 1.490822872434631,
      "learning_rate": 8.980845455683706e-07,
      "loss": 2.4744,
      "step": 3248
    },
    {
      "epoch": 0.42062336149140694,
      "grad_norm": 1.433193793646382,
      "learning_rate": 8.978304319025131e-07,
      "loss": 2.3569,
      "step": 3249
    },
    {
      "epoch": 0.4207528238987604,
      "grad_norm": 1.9253861712432947,
      "learning_rate": 8.975762625055096e-07,
      "loss": 2.5391,
      "step": 3250
    },
    {
      "epoch": 0.4207528238987604,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6213600635528564,
      "eval_runtime": 13.9722,
      "eval_samples_per_second": 3.149,
      "eval_steps_per_second": 0.215,
      "step": 3250
    },
    {
      "epoch": 0.42088228630611385,
      "grad_norm": 2.3023110306251238,
      "learning_rate": 8.973220374292799e-07,
      "loss": 2.2634,
      "step": 3251
    },
    {
      "epoch": 0.4210117487134673,
      "grad_norm": 1.6665098237002203,
      "learning_rate": 8.970677567257541e-07,
      "loss": 2.5813,
      "step": 3252
    },
    {
      "epoch": 0.4211412111208208,
      "grad_norm": 1.524505509354147,
      "learning_rate": 8.968134204468744e-07,
      "loss": 2.4338,
      "step": 3253
    },
    {
      "epoch": 0.42127067352817427,
      "grad_norm": 1.4307524908907743,
      "learning_rate": 8.965590286445943e-07,
      "loss": 2.2358,
      "step": 3254
    },
    {
      "epoch": 0.4214001359355277,
      "grad_norm": 2.6827791641067913,
      "learning_rate": 8.963045813708781e-07,
      "loss": 2.7305,
      "step": 3255
    },
    {
      "epoch": 0.4214001359355277,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6197398900985718,
      "eval_runtime": 13.6866,
      "eval_samples_per_second": 3.215,
      "eval_steps_per_second": 0.219,
      "step": 3255
    },
    {
      "epoch": 0.4215295983428812,
      "grad_norm": 3.2374998887108175,
      "learning_rate": 8.960500786777024e-07,
      "loss": 2.6328,
      "step": 3256
    },
    {
      "epoch": 0.4216590607502346,
      "grad_norm": 4.542618127048336,
      "learning_rate": 8.957955206170542e-07,
      "loss": 2.8135,
      "step": 3257
    },
    {
      "epoch": 0.42178852315758814,
      "grad_norm": 1.9912255441854707,
      "learning_rate": 8.95540907240932e-07,
      "loss": 2.7024,
      "step": 3258
    },
    {
      "epoch": 0.4219179855649416,
      "grad_norm": 2.6951357544798276,
      "learning_rate": 8.952862386013458e-07,
      "loss": 2.6274,
      "step": 3259
    },
    {
      "epoch": 0.42204744797229504,
      "grad_norm": 2.316426134433997,
      "learning_rate": 8.950315147503169e-07,
      "loss": 2.4709,
      "step": 3260
    },
    {
      "epoch": 0.42204744797229504,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6160777807235718,
      "eval_runtime": 15.4201,
      "eval_samples_per_second": 2.853,
      "eval_steps_per_second": 0.195,
      "step": 3260
    },
    {
      "epoch": 0.4221769103796485,
      "grad_norm": 1.9462753706412466,
      "learning_rate": 8.947767357398779e-07,
      "loss": 2.7262,
      "step": 3261
    },
    {
      "epoch": 0.42230637278700195,
      "grad_norm": 3.678258936761399,
      "learning_rate": 8.945219016220725e-07,
      "loss": 2.8623,
      "step": 3262
    },
    {
      "epoch": 0.42243583519435546,
      "grad_norm": 2.0758018068650768,
      "learning_rate": 8.94267012448956e-07,
      "loss": 2.6953,
      "step": 3263
    },
    {
      "epoch": 0.4225652976017089,
      "grad_norm": 1.6535743939177947,
      "learning_rate": 8.94012068272594e-07,
      "loss": 2.4502,
      "step": 3264
    },
    {
      "epoch": 0.42269476000906236,
      "grad_norm": 1.668886492377874,
      "learning_rate": 8.937570691450646e-07,
      "loss": 2.5532,
      "step": 3265
    },
    {
      "epoch": 0.42269476000906236,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.619140625,
      "eval_runtime": 14.21,
      "eval_samples_per_second": 3.096,
      "eval_steps_per_second": 0.211,
      "step": 3265
    },
    {
      "epoch": 0.4228242224164158,
      "grad_norm": 2.1617070471768076,
      "learning_rate": 8.935020151184563e-07,
      "loss": 2.5706,
      "step": 3266
    },
    {
      "epoch": 0.42295368482376927,
      "grad_norm": 2.1647015531278186,
      "learning_rate": 8.932469062448693e-07,
      "loss": 2.71,
      "step": 3267
    },
    {
      "epoch": 0.4230831472311228,
      "grad_norm": 1.962576888316768,
      "learning_rate": 8.929917425764146e-07,
      "loss": 2.3623,
      "step": 3268
    },
    {
      "epoch": 0.42321260963847623,
      "grad_norm": 2.019840024360644,
      "learning_rate": 8.927365241652142e-07,
      "loss": 2.592,
      "step": 3269
    },
    {
      "epoch": 0.4233420720458297,
      "grad_norm": 1.9404260348685092,
      "learning_rate": 8.924812510634021e-07,
      "loss": 2.5413,
      "step": 3270
    },
    {
      "epoch": 0.4233420720458297,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6184748411178589,
      "eval_runtime": 15.0263,
      "eval_samples_per_second": 2.928,
      "eval_steps_per_second": 0.2,
      "step": 3270
    },
    {
      "epoch": 0.42347153445318314,
      "grad_norm": 2.1198821501137255,
      "learning_rate": 8.92225923323123e-07,
      "loss": 2.375,
      "step": 3271
    },
    {
      "epoch": 0.42360099686053665,
      "grad_norm": 2.5545400151560305,
      "learning_rate": 8.919705409965326e-07,
      "loss": 2.5815,
      "step": 3272
    },
    {
      "epoch": 0.4237304592678901,
      "grad_norm": 1.8061257156703898,
      "learning_rate": 8.917151041357979e-07,
      "loss": 2.5798,
      "step": 3273
    },
    {
      "epoch": 0.42385992167524356,
      "grad_norm": 1.528558582689481,
      "learning_rate": 8.914596127930972e-07,
      "loss": 2.4153,
      "step": 3274
    },
    {
      "epoch": 0.423989384082597,
      "grad_norm": 2.439987648451466,
      "learning_rate": 8.912040670206196e-07,
      "loss": 2.6223,
      "step": 3275
    },
    {
      "epoch": 0.423989384082597,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6185413599014282,
      "eval_runtime": 13.6,
      "eval_samples_per_second": 3.235,
      "eval_steps_per_second": 0.221,
      "step": 3275
    },
    {
      "epoch": 0.42411884648995046,
      "grad_norm": 3.2820873926307863,
      "learning_rate": 8.909484668705656e-07,
      "loss": 2.8037,
      "step": 3276
    },
    {
      "epoch": 0.42424830889730397,
      "grad_norm": 1.8007726185049484,
      "learning_rate": 8.906928123951468e-07,
      "loss": 2.4868,
      "step": 3277
    },
    {
      "epoch": 0.4243777713046574,
      "grad_norm": 1.3391092303121688,
      "learning_rate": 8.904371036465856e-07,
      "loss": 2.3364,
      "step": 3278
    },
    {
      "epoch": 0.4245072337120109,
      "grad_norm": 4.157233652245847,
      "learning_rate": 8.90181340677116e-07,
      "loss": 3.1763,
      "step": 3279
    },
    {
      "epoch": 0.42463669611936433,
      "grad_norm": 1.4170572778663098,
      "learning_rate": 8.899255235389824e-07,
      "loss": 2.4294,
      "step": 3280
    },
    {
      "epoch": 0.42463669611936433,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6171208620071411,
      "eval_runtime": 13.1351,
      "eval_samples_per_second": 3.35,
      "eval_steps_per_second": 0.228,
      "step": 3280
    },
    {
      "epoch": 0.4247661585267178,
      "grad_norm": 2.130560649596276,
      "learning_rate": 8.896696522844408e-07,
      "loss": 2.4863,
      "step": 3281
    },
    {
      "epoch": 0.4248956209340713,
      "grad_norm": 1.3699061765897067,
      "learning_rate": 8.894137269657582e-07,
      "loss": 2.5444,
      "step": 3282
    },
    {
      "epoch": 0.42502508334142475,
      "grad_norm": 1.5999821864135884,
      "learning_rate": 8.891577476352126e-07,
      "loss": 2.2563,
      "step": 3283
    },
    {
      "epoch": 0.4251545457487782,
      "grad_norm": 1.7772694657872636,
      "learning_rate": 8.889017143450926e-07,
      "loss": 2.8235,
      "step": 3284
    },
    {
      "epoch": 0.42528400815613165,
      "grad_norm": 2.937525886820933,
      "learning_rate": 8.886456271476986e-07,
      "loss": 2.5205,
      "step": 3285
    },
    {
      "epoch": 0.42528400815613165,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6163440942764282,
      "eval_runtime": 13.6553,
      "eval_samples_per_second": 3.222,
      "eval_steps_per_second": 0.22,
      "step": 3285
    },
    {
      "epoch": 0.4254134705634851,
      "grad_norm": 1.5263810748228492,
      "learning_rate": 8.883894860953413e-07,
      "loss": 2.4785,
      "step": 3286
    },
    {
      "epoch": 0.4255429329708386,
      "grad_norm": 1.4393148016264663,
      "learning_rate": 8.881332912403426e-07,
      "loss": 2.4431,
      "step": 3287
    },
    {
      "epoch": 0.42567239537819207,
      "grad_norm": 2.9434749987556366,
      "learning_rate": 8.878770426350358e-07,
      "loss": 2.7056,
      "step": 3288
    },
    {
      "epoch": 0.4258018577855455,
      "grad_norm": 4.360730596043665,
      "learning_rate": 8.87620740331765e-07,
      "loss": 2.6819,
      "step": 3289
    },
    {
      "epoch": 0.425931320192899,
      "grad_norm": 1.674021010529709,
      "learning_rate": 8.873643843828847e-07,
      "loss": 2.5413,
      "step": 3290
    },
    {
      "epoch": 0.425931320192899,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6112614870071411,
      "eval_runtime": 13.7538,
      "eval_samples_per_second": 3.199,
      "eval_steps_per_second": 0.218,
      "step": 3290
    },
    {
      "epoch": 0.42606078260025243,
      "grad_norm": 2.3355657717651868,
      "learning_rate": 8.87107974840761e-07,
      "loss": 2.4731,
      "step": 3291
    },
    {
      "epoch": 0.42619024500760594,
      "grad_norm": 1.525437938770551,
      "learning_rate": 8.868515117577707e-07,
      "loss": 2.5164,
      "step": 3292
    },
    {
      "epoch": 0.4263197074149594,
      "grad_norm": 3.8956202128910533,
      "learning_rate": 8.865949951863019e-07,
      "loss": 2.6938,
      "step": 3293
    },
    {
      "epoch": 0.42644916982231285,
      "grad_norm": 2.5048905859058133,
      "learning_rate": 8.863384251787528e-07,
      "loss": 2.6521,
      "step": 3294
    },
    {
      "epoch": 0.4265786322296663,
      "grad_norm": 1.746924905531486,
      "learning_rate": 8.860818017875334e-07,
      "loss": 2.3784,
      "step": 3295
    },
    {
      "epoch": 0.4265786322296663,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.611328125,
      "eval_runtime": 13.4273,
      "eval_samples_per_second": 3.277,
      "eval_steps_per_second": 0.223,
      "step": 3295
    },
    {
      "epoch": 0.42670809463701975,
      "grad_norm": 3.1409009993341206,
      "learning_rate": 8.85825125065064e-07,
      "loss": 2.4053,
      "step": 3296
    },
    {
      "epoch": 0.42683755704437326,
      "grad_norm": 3.2568328795019545,
      "learning_rate": 8.855683950637761e-07,
      "loss": 2.8015,
      "step": 3297
    },
    {
      "epoch": 0.4269670194517267,
      "grad_norm": 1.9369177785713558,
      "learning_rate": 8.853116118361121e-07,
      "loss": 2.4966,
      "step": 3298
    },
    {
      "epoch": 0.42709648185908017,
      "grad_norm": 2.5107018877847325,
      "learning_rate": 8.850547754345251e-07,
      "loss": 2.415,
      "step": 3299
    },
    {
      "epoch": 0.4272259442664336,
      "grad_norm": 2.0471591360427315,
      "learning_rate": 8.847978859114792e-07,
      "loss": 2.3945,
      "step": 3300
    },
    {
      "epoch": 0.4272259442664336,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6119939088821411,
      "eval_runtime": 13.967,
      "eval_samples_per_second": 3.15,
      "eval_steps_per_second": 0.215,
      "step": 3300
    },
    {
      "epoch": 0.4273554066737871,
      "grad_norm": 2.3780266166788695,
      "learning_rate": 8.845409433194491e-07,
      "loss": 2.9751,
      "step": 3301
    },
    {
      "epoch": 0.4274848690811406,
      "grad_norm": 2.719911936365414,
      "learning_rate": 8.842839477109208e-07,
      "loss": 2.4644,
      "step": 3302
    },
    {
      "epoch": 0.42761433148849404,
      "grad_norm": 2.2835477644242106,
      "learning_rate": 8.840268991383908e-07,
      "loss": 2.6497,
      "step": 3303
    },
    {
      "epoch": 0.4277437938958475,
      "grad_norm": 1.8329933799909754,
      "learning_rate": 8.837697976543663e-07,
      "loss": 2.5146,
      "step": 3304
    },
    {
      "epoch": 0.42787325630320094,
      "grad_norm": 3.0023311875579655,
      "learning_rate": 8.835126433113655e-07,
      "loss": 2.6895,
      "step": 3305
    },
    {
      "epoch": 0.42787325630320094,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6106178760528564,
      "eval_runtime": 14.2485,
      "eval_samples_per_second": 3.088,
      "eval_steps_per_second": 0.211,
      "step": 3305
    },
    {
      "epoch": 0.4280027187105544,
      "grad_norm": 2.471240027748569,
      "learning_rate": 8.832554361619175e-07,
      "loss": 2.6199,
      "step": 3306
    },
    {
      "epoch": 0.4281321811179079,
      "grad_norm": 2.5092914791181524,
      "learning_rate": 8.829981762585624e-07,
      "loss": 2.4937,
      "step": 3307
    },
    {
      "epoch": 0.42826164352526136,
      "grad_norm": 1.8282969143604033,
      "learning_rate": 8.827408636538501e-07,
      "loss": 2.3113,
      "step": 3308
    },
    {
      "epoch": 0.4283911059326148,
      "grad_norm": 2.365926070476587,
      "learning_rate": 8.824834984003422e-07,
      "loss": 2.3408,
      "step": 3309
    },
    {
      "epoch": 0.42852056833996827,
      "grad_norm": 1.681006429764802,
      "learning_rate": 8.82226080550611e-07,
      "loss": 2.6309,
      "step": 3310
    },
    {
      "epoch": 0.42852056833996827,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6099520921707153,
      "eval_runtime": 13.9877,
      "eval_samples_per_second": 3.146,
      "eval_steps_per_second": 0.214,
      "step": 3310
    },
    {
      "epoch": 0.4286500307473217,
      "grad_norm": 1.7909116590903777,
      "learning_rate": 8.819686101572389e-07,
      "loss": 2.4551,
      "step": 3311
    },
    {
      "epoch": 0.42877949315467523,
      "grad_norm": 1.8042349247125555,
      "learning_rate": 8.817110872728199e-07,
      "loss": 2.7805,
      "step": 3312
    },
    {
      "epoch": 0.4289089555620287,
      "grad_norm": 1.5481426122418513,
      "learning_rate": 8.814535119499579e-07,
      "loss": 2.7432,
      "step": 3313
    },
    {
      "epoch": 0.42903841796938214,
      "grad_norm": 1.961620933477069,
      "learning_rate": 8.811958842412679e-07,
      "loss": 2.5015,
      "step": 3314
    },
    {
      "epoch": 0.4291678803767356,
      "grad_norm": 2.231523269839053,
      "learning_rate": 8.80938204199376e-07,
      "loss": 2.584,
      "step": 3315
    },
    {
      "epoch": 0.4291678803767356,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.61279296875,
      "eval_runtime": 13.9822,
      "eval_samples_per_second": 3.147,
      "eval_steps_per_second": 0.215,
      "step": 3315
    },
    {
      "epoch": 0.4292973427840891,
      "grad_norm": 1.923431514811729,
      "learning_rate": 8.80680471876918e-07,
      "loss": 2.5432,
      "step": 3316
    },
    {
      "epoch": 0.42942680519144255,
      "grad_norm": 1.7322720945534793,
      "learning_rate": 8.804226873265412e-07,
      "loss": 2.48,
      "step": 3317
    },
    {
      "epoch": 0.429556267598796,
      "grad_norm": 2.204274486264051,
      "learning_rate": 8.801648506009033e-07,
      "loss": 2.4814,
      "step": 3318
    },
    {
      "epoch": 0.42968573000614946,
      "grad_norm": 1.6943207208619377,
      "learning_rate": 8.799069617526728e-07,
      "loss": 2.5107,
      "step": 3319
    },
    {
      "epoch": 0.4298151924135029,
      "grad_norm": 2.7090515846357803,
      "learning_rate": 8.796490208345288e-07,
      "loss": 2.4072,
      "step": 3320
    },
    {
      "epoch": 0.4298151924135029,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6091974973678589,
      "eval_runtime": 14.5239,
      "eval_samples_per_second": 3.029,
      "eval_steps_per_second": 0.207,
      "step": 3320
    },
    {
      "epoch": 0.4299446548208564,
      "grad_norm": 1.626923485269989,
      "learning_rate": 8.793910278991608e-07,
      "loss": 2.4314,
      "step": 3321
    },
    {
      "epoch": 0.4300741172282099,
      "grad_norm": 3.1800503795137804,
      "learning_rate": 8.791329829992688e-07,
      "loss": 2.9187,
      "step": 3322
    },
    {
      "epoch": 0.4302035796355633,
      "grad_norm": 1.3641566312040845,
      "learning_rate": 8.78874886187564e-07,
      "loss": 2.2957,
      "step": 3323
    },
    {
      "epoch": 0.4303330420429168,
      "grad_norm": 2.7601740894497424,
      "learning_rate": 8.786167375167681e-07,
      "loss": 2.7703,
      "step": 3324
    },
    {
      "epoch": 0.43046250445027023,
      "grad_norm": 2.1503164425255212,
      "learning_rate": 8.783585370396127e-07,
      "loss": 2.7195,
      "step": 3325
    },
    {
      "epoch": 0.43046250445027023,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6126598119735718,
      "eval_runtime": 14.8409,
      "eval_samples_per_second": 2.965,
      "eval_steps_per_second": 0.202,
      "step": 3325
    },
    {
      "epoch": 0.43059196685762374,
      "grad_norm": 1.7714639521996063,
      "learning_rate": 8.781002848088411e-07,
      "loss": 2.5303,
      "step": 3326
    },
    {
      "epoch": 0.4307214292649772,
      "grad_norm": 2.1597747392886877,
      "learning_rate": 8.77841980877206e-07,
      "loss": 2.666,
      "step": 3327
    },
    {
      "epoch": 0.43085089167233065,
      "grad_norm": 1.1357448939177524,
      "learning_rate": 8.775836252974715e-07,
      "loss": 2.1963,
      "step": 3328
    },
    {
      "epoch": 0.4309803540796841,
      "grad_norm": 2.134878756173607,
      "learning_rate": 8.773252181224121e-07,
      "loss": 2.5022,
      "step": 3329
    },
    {
      "epoch": 0.43110981648703756,
      "grad_norm": 1.6217727308247625,
      "learning_rate": 8.770667594048121e-07,
      "loss": 2.2305,
      "step": 3330
    },
    {
      "epoch": 0.43110981648703756,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.612016201019287,
      "eval_runtime": 14.5759,
      "eval_samples_per_second": 3.019,
      "eval_steps_per_second": 0.206,
      "step": 3330
    },
    {
      "epoch": 0.43123927889439106,
      "grad_norm": 4.593393228292561,
      "learning_rate": 8.768082491974675e-07,
      "loss": 2.9822,
      "step": 3331
    },
    {
      "epoch": 0.4313687413017445,
      "grad_norm": 2.543337754708295,
      "learning_rate": 8.765496875531841e-07,
      "loss": 2.6838,
      "step": 3332
    },
    {
      "epoch": 0.43149820370909797,
      "grad_norm": 1.31875583086638,
      "learning_rate": 8.762910745247784e-07,
      "loss": 2.3336,
      "step": 3333
    },
    {
      "epoch": 0.4316276661164514,
      "grad_norm": 1.8618292485888146,
      "learning_rate": 8.760324101650772e-07,
      "loss": 2.4878,
      "step": 3334
    },
    {
      "epoch": 0.4317571285238049,
      "grad_norm": 1.5730626090404043,
      "learning_rate": 8.757736945269183e-07,
      "loss": 2.4136,
      "step": 3335
    },
    {
      "epoch": 0.4317571285238049,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6099299192428589,
      "eval_runtime": 15.1485,
      "eval_samples_per_second": 2.905,
      "eval_steps_per_second": 0.198,
      "step": 3335
    },
    {
      "epoch": 0.4318865909311584,
      "grad_norm": 2.793919684749034,
      "learning_rate": 8.755149276631488e-07,
      "loss": 3.2341,
      "step": 3336
    },
    {
      "epoch": 0.43201605333851184,
      "grad_norm": 3.027613300059181,
      "learning_rate": 8.752561096266279e-07,
      "loss": 2.5266,
      "step": 3337
    },
    {
      "epoch": 0.4321455157458653,
      "grad_norm": 1.698814111646842,
      "learning_rate": 8.749972404702242e-07,
      "loss": 2.3486,
      "step": 3338
    },
    {
      "epoch": 0.43227497815321875,
      "grad_norm": 2.363862095933434,
      "learning_rate": 8.74738320246817e-07,
      "loss": 2.6167,
      "step": 3339
    },
    {
      "epoch": 0.4324044405605722,
      "grad_norm": 1.4915473421090697,
      "learning_rate": 8.74479349009296e-07,
      "loss": 2.4108,
      "step": 3340
    },
    {
      "epoch": 0.4324044405605722,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.604203701019287,
      "eval_runtime": 14.0615,
      "eval_samples_per_second": 3.129,
      "eval_steps_per_second": 0.213,
      "step": 3340
    },
    {
      "epoch": 0.4325339029679257,
      "grad_norm": 2.04943456538855,
      "learning_rate": 8.742203268105607e-07,
      "loss": 2.6826,
      "step": 3341
    },
    {
      "epoch": 0.43266336537527916,
      "grad_norm": 1.533730457714011,
      "learning_rate": 8.739612537035227e-07,
      "loss": 2.4873,
      "step": 3342
    },
    {
      "epoch": 0.4327928277826326,
      "grad_norm": 1.4484341318687055,
      "learning_rate": 8.737021297411024e-07,
      "loss": 2.6475,
      "step": 3343
    },
    {
      "epoch": 0.43292229018998607,
      "grad_norm": 1.702869870301113,
      "learning_rate": 8.734429549762309e-07,
      "loss": 2.6052,
      "step": 3344
    },
    {
      "epoch": 0.4330517525973395,
      "grad_norm": 2.859857165430935,
      "learning_rate": 8.731837294618501e-07,
      "loss": 2.4946,
      "step": 3345
    },
    {
      "epoch": 0.4330517525973395,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6056462526321411,
      "eval_runtime": 13.7967,
      "eval_samples_per_second": 3.189,
      "eval_steps_per_second": 0.217,
      "step": 3345
    },
    {
      "epoch": 0.43318121500469303,
      "grad_norm": 1.4731800231537813,
      "learning_rate": 8.729244532509124e-07,
      "loss": 2.5166,
      "step": 3346
    },
    {
      "epoch": 0.4333106774120465,
      "grad_norm": 2.1589361458003804,
      "learning_rate": 8.726651263963797e-07,
      "loss": 2.7766,
      "step": 3347
    },
    {
      "epoch": 0.43344013981939994,
      "grad_norm": 1.3084918079653816,
      "learning_rate": 8.72405748951225e-07,
      "loss": 2.3594,
      "step": 3348
    },
    {
      "epoch": 0.4335696022267534,
      "grad_norm": 1.812523510963763,
      "learning_rate": 8.721463209684315e-07,
      "loss": 2.6387,
      "step": 3349
    },
    {
      "epoch": 0.43369906463410685,
      "grad_norm": 3.1303527253118726,
      "learning_rate": 8.718868425009924e-07,
      "loss": 2.6023,
      "step": 3350
    },
    {
      "epoch": 0.43369906463410685,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6034934520721436,
      "eval_runtime": 13.6155,
      "eval_samples_per_second": 3.232,
      "eval_steps_per_second": 0.22,
      "step": 3350
    },
    {
      "epoch": 0.43382852704146035,
      "grad_norm": 1.8553547256261265,
      "learning_rate": 8.716273136019116e-07,
      "loss": 2.6775,
      "step": 3351
    },
    {
      "epoch": 0.4339579894488138,
      "grad_norm": 2.547847038990347,
      "learning_rate": 8.713677343242031e-07,
      "loss": 2.4614,
      "step": 3352
    },
    {
      "epoch": 0.43408745185616726,
      "grad_norm": 3.75109330221715,
      "learning_rate": 8.711081047208912e-07,
      "loss": 2.6025,
      "step": 3353
    },
    {
      "epoch": 0.4342169142635207,
      "grad_norm": 1.9635393930900862,
      "learning_rate": 8.708484248450105e-07,
      "loss": 2.6328,
      "step": 3354
    },
    {
      "epoch": 0.43434637667087417,
      "grad_norm": 1.7949588842666337,
      "learning_rate": 8.70588694749606e-07,
      "loss": 2.6301,
      "step": 3355
    },
    {
      "epoch": 0.43434637667087417,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.600874423980713,
      "eval_runtime": 14.369,
      "eval_samples_per_second": 3.062,
      "eval_steps_per_second": 0.209,
      "step": 3355
    },
    {
      "epoch": 0.4344758390782277,
      "grad_norm": 1.9656718842338972,
      "learning_rate": 8.703289144877326e-07,
      "loss": 2.7419,
      "step": 3356
    },
    {
      "epoch": 0.43460530148558113,
      "grad_norm": 2.6642528906401544,
      "learning_rate": 8.70069084112456e-07,
      "loss": 2.5046,
      "step": 3357
    },
    {
      "epoch": 0.4347347638929346,
      "grad_norm": 2.8585548146962867,
      "learning_rate": 8.698092036768516e-07,
      "loss": 2.8552,
      "step": 3358
    },
    {
      "epoch": 0.43486422630028804,
      "grad_norm": 1.1999466163898747,
      "learning_rate": 8.695492732340053e-07,
      "loss": 2.3899,
      "step": 3359
    },
    {
      "epoch": 0.43499368870764155,
      "grad_norm": 3.3602213170747643,
      "learning_rate": 8.692892928370131e-07,
      "loss": 2.5068,
      "step": 3360
    },
    {
      "epoch": 0.43499368870764155,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6021617650985718,
      "eval_runtime": 15.3449,
      "eval_samples_per_second": 2.867,
      "eval_steps_per_second": 0.196,
      "step": 3360
    },
    {
      "epoch": 0.435123151114995,
      "grad_norm": 1.5053696365724347,
      "learning_rate": 8.690292625389813e-07,
      "loss": 2.3132,
      "step": 3361
    },
    {
      "epoch": 0.43525261352234845,
      "grad_norm": 2.0258490903535473,
      "learning_rate": 8.687691823930264e-07,
      "loss": 2.3914,
      "step": 3362
    },
    {
      "epoch": 0.4353820759297019,
      "grad_norm": 1.6941609582033224,
      "learning_rate": 8.685090524522751e-07,
      "loss": 2.2051,
      "step": 3363
    },
    {
      "epoch": 0.43551153833705536,
      "grad_norm": 2.9855637013370764,
      "learning_rate": 8.682488727698638e-07,
      "loss": 2.8433,
      "step": 3364
    },
    {
      "epoch": 0.43564100074440887,
      "grad_norm": 4.064327015181763,
      "learning_rate": 8.679886433989401e-07,
      "loss": 2.8965,
      "step": 3365
    },
    {
      "epoch": 0.43564100074440887,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.600053310394287,
      "eval_runtime": 16.8051,
      "eval_samples_per_second": 2.618,
      "eval_steps_per_second": 0.179,
      "step": 3365
    },
    {
      "epoch": 0.4357704631517623,
      "grad_norm": 1.4337001854329114,
      "learning_rate": 8.677283643926609e-07,
      "loss": 2.5332,
      "step": 3366
    },
    {
      "epoch": 0.4358999255591158,
      "grad_norm": 1.4481161669105034,
      "learning_rate": 8.674680358041932e-07,
      "loss": 2.3538,
      "step": 3367
    },
    {
      "epoch": 0.43602938796646923,
      "grad_norm": 2.7396680621904284,
      "learning_rate": 8.672076576867147e-07,
      "loss": 2.6411,
      "step": 3368
    },
    {
      "epoch": 0.4361588503738227,
      "grad_norm": 1.956391892926346,
      "learning_rate": 8.669472300934128e-07,
      "loss": 2.4097,
      "step": 3369
    },
    {
      "epoch": 0.4362883127811762,
      "grad_norm": 1.622880309561545,
      "learning_rate": 8.66686753077485e-07,
      "loss": 2.5227,
      "step": 3370
    },
    {
      "epoch": 0.4362883127811762,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.597212314605713,
      "eval_runtime": 14.3505,
      "eval_samples_per_second": 3.066,
      "eval_steps_per_second": 0.209,
      "step": 3370
    },
    {
      "epoch": 0.43641777518852964,
      "grad_norm": 3.9637736274517943,
      "learning_rate": 8.664262266921395e-07,
      "loss": 3.2598,
      "step": 3371
    },
    {
      "epoch": 0.4365472375958831,
      "grad_norm": 2.2328216453439893,
      "learning_rate": 8.661656509905933e-07,
      "loss": 2.6431,
      "step": 3372
    },
    {
      "epoch": 0.43667670000323655,
      "grad_norm": 1.7118936859919116,
      "learning_rate": 8.659050260260746e-07,
      "loss": 2.2583,
      "step": 3373
    },
    {
      "epoch": 0.43680616241059,
      "grad_norm": 1.8773939494511314,
      "learning_rate": 8.656443518518216e-07,
      "loss": 2.3059,
      "step": 3374
    },
    {
      "epoch": 0.4369356248179435,
      "grad_norm": 1.3219131010165073,
      "learning_rate": 8.653836285210822e-07,
      "loss": 2.4644,
      "step": 3375
    },
    {
      "epoch": 0.4369356248179435,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5993430614471436,
      "eval_runtime": 13.9219,
      "eval_samples_per_second": 3.16,
      "eval_steps_per_second": 0.215,
      "step": 3375
    },
    {
      "epoch": 0.43706508722529697,
      "grad_norm": 1.2464528363036336,
      "learning_rate": 8.651228560871142e-07,
      "loss": 2.4553,
      "step": 3376
    },
    {
      "epoch": 0.4371945496326504,
      "grad_norm": 2.9549599624428957,
      "learning_rate": 8.648620346031856e-07,
      "loss": 2.7131,
      "step": 3377
    },
    {
      "epoch": 0.4373240120400039,
      "grad_norm": 1.6168513729271259,
      "learning_rate": 8.646011641225749e-07,
      "loss": 2.551,
      "step": 3378
    },
    {
      "epoch": 0.4374534744473573,
      "grad_norm": 2.503684851559407,
      "learning_rate": 8.643402446985699e-07,
      "loss": 2.3838,
      "step": 3379
    },
    {
      "epoch": 0.43758293685471084,
      "grad_norm": 5.204177313655293,
      "learning_rate": 8.640792763844688e-07,
      "loss": 3.0251,
      "step": 3380
    },
    {
      "epoch": 0.43758293685471084,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6007634401321411,
      "eval_runtime": 14.1301,
      "eval_samples_per_second": 3.114,
      "eval_steps_per_second": 0.212,
      "step": 3380
    },
    {
      "epoch": 0.4377123992620643,
      "grad_norm": 1.6433985321588829,
      "learning_rate": 8.638182592335793e-07,
      "loss": 2.5015,
      "step": 3381
    },
    {
      "epoch": 0.43784186166941774,
      "grad_norm": 1.3609592400905013,
      "learning_rate": 8.6355719329922e-07,
      "loss": 2.4255,
      "step": 3382
    },
    {
      "epoch": 0.4379713240767712,
      "grad_norm": 1.8615098510039647,
      "learning_rate": 8.632960786347184e-07,
      "loss": 2.4766,
      "step": 3383
    },
    {
      "epoch": 0.43810078648412465,
      "grad_norm": 2.1488978232549893,
      "learning_rate": 8.630349152934128e-07,
      "loss": 2.4688,
      "step": 3384
    },
    {
      "epoch": 0.43823024889147816,
      "grad_norm": 3.941986928649056,
      "learning_rate": 8.627737033286512e-07,
      "loss": 2.1689,
      "step": 3385
    },
    {
      "epoch": 0.43823024889147816,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.59716796875,
      "eval_runtime": 15.117,
      "eval_samples_per_second": 2.911,
      "eval_steps_per_second": 0.198,
      "step": 3385
    },
    {
      "epoch": 0.4383597112988316,
      "grad_norm": 1.4453550552085235,
      "learning_rate": 8.625124427937909e-07,
      "loss": 2.532,
      "step": 3386
    },
    {
      "epoch": 0.43848917370618506,
      "grad_norm": 2.016114918139788,
      "learning_rate": 8.622511337422002e-07,
      "loss": 2.2383,
      "step": 3387
    },
    {
      "epoch": 0.4386186361135385,
      "grad_norm": 1.4919377310131778,
      "learning_rate": 8.619897762272566e-07,
      "loss": 2.4177,
      "step": 3388
    },
    {
      "epoch": 0.43874809852089197,
      "grad_norm": 2.6028617708322934,
      "learning_rate": 8.617283703023475e-07,
      "loss": 2.3193,
      "step": 3389
    },
    {
      "epoch": 0.4388775609282455,
      "grad_norm": 2.2683466424149334,
      "learning_rate": 8.614669160208707e-07,
      "loss": 2.7856,
      "step": 3390
    },
    {
      "epoch": 0.4388775609282455,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.597212314605713,
      "eval_runtime": 15.0193,
      "eval_samples_per_second": 2.93,
      "eval_steps_per_second": 0.2,
      "step": 3390
    },
    {
      "epoch": 0.43900702333559893,
      "grad_norm": 1.9721875224024048,
      "learning_rate": 8.612054134362331e-07,
      "loss": 2.408,
      "step": 3391
    },
    {
      "epoch": 0.4391364857429524,
      "grad_norm": 3.7510369403129245,
      "learning_rate": 8.609438626018522e-07,
      "loss": 2.6289,
      "step": 3392
    },
    {
      "epoch": 0.43926594815030584,
      "grad_norm": 1.2007011663733482,
      "learning_rate": 8.606822635711553e-07,
      "loss": 2.2271,
      "step": 3393
    },
    {
      "epoch": 0.4393954105576593,
      "grad_norm": 2.2140861932290323,
      "learning_rate": 8.604206163975787e-07,
      "loss": 2.8462,
      "step": 3394
    },
    {
      "epoch": 0.4395248729650128,
      "grad_norm": 1.872874280001812,
      "learning_rate": 8.601589211345694e-07,
      "loss": 2.366,
      "step": 3395
    },
    {
      "epoch": 0.4395248729650128,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.59716796875,
      "eval_runtime": 13.7835,
      "eval_samples_per_second": 3.192,
      "eval_steps_per_second": 0.218,
      "step": 3395
    },
    {
      "epoch": 0.43965433537236626,
      "grad_norm": 2.270875388489974,
      "learning_rate": 8.598971778355842e-07,
      "loss": 2.3323,
      "step": 3396
    },
    {
      "epoch": 0.4397837977797197,
      "grad_norm": 1.8802560539137303,
      "learning_rate": 8.596353865540891e-07,
      "loss": 2.6616,
      "step": 3397
    },
    {
      "epoch": 0.43991326018707316,
      "grad_norm": 1.5463882268775082,
      "learning_rate": 8.593735473435605e-07,
      "loss": 2.4248,
      "step": 3398
    },
    {
      "epoch": 0.4400427225944266,
      "grad_norm": 2.5888801357490405,
      "learning_rate": 8.591116602574845e-07,
      "loss": 2.7466,
      "step": 3399
    },
    {
      "epoch": 0.4401721850017801,
      "grad_norm": 2.463832441045507,
      "learning_rate": 8.588497253493563e-07,
      "loss": 2.604,
      "step": 3400
    },
    {
      "epoch": 0.4401721850017801,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.602095127105713,
      "eval_runtime": 15.6022,
      "eval_samples_per_second": 2.82,
      "eval_steps_per_second": 0.192,
      "step": 3400
    },
    {
      "epoch": 0.4403016474091336,
      "grad_norm": 4.191232338270609,
      "learning_rate": 8.585877426726821e-07,
      "loss": 3.2261,
      "step": 3401
    },
    {
      "epoch": 0.44043110981648703,
      "grad_norm": 2.5971509161605466,
      "learning_rate": 8.583257122809768e-07,
      "loss": 2.5598,
      "step": 3402
    },
    {
      "epoch": 0.4405605722238405,
      "grad_norm": 2.7262606605151185,
      "learning_rate": 8.580636342277654e-07,
      "loss": 2.3704,
      "step": 3403
    },
    {
      "epoch": 0.440690034631194,
      "grad_norm": 1.575631610522882,
      "learning_rate": 8.578015085665828e-07,
      "loss": 2.4663,
      "step": 3404
    },
    {
      "epoch": 0.44081949703854745,
      "grad_norm": 1.7416078489275513,
      "learning_rate": 8.575393353509733e-07,
      "loss": 2.437,
      "step": 3405
    },
    {
      "epoch": 0.44081949703854745,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5992764234542847,
      "eval_runtime": 16.736,
      "eval_samples_per_second": 2.629,
      "eval_steps_per_second": 0.179,
      "step": 3405
    },
    {
      "epoch": 0.4409489594459009,
      "grad_norm": 1.858486069236183,
      "learning_rate": 8.572771146344911e-07,
      "loss": 2.4883,
      "step": 3406
    },
    {
      "epoch": 0.44107842185325435,
      "grad_norm": 1.449590755098125,
      "learning_rate": 8.570148464707005e-07,
      "loss": 2.3585,
      "step": 3407
    },
    {
      "epoch": 0.4412078842606078,
      "grad_norm": 1.9889072502150644,
      "learning_rate": 8.567525309131746e-07,
      "loss": 2.6187,
      "step": 3408
    },
    {
      "epoch": 0.4413373466679613,
      "grad_norm": 1.7210628402362425,
      "learning_rate": 8.564901680154968e-07,
      "loss": 2.2261,
      "step": 3409
    },
    {
      "epoch": 0.44146680907531477,
      "grad_norm": 2.163720413595985,
      "learning_rate": 8.562277578312603e-07,
      "loss": 2.3452,
      "step": 3410
    },
    {
      "epoch": 0.44146680907531477,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6006969213485718,
      "eval_runtime": 14.8723,
      "eval_samples_per_second": 2.959,
      "eval_steps_per_second": 0.202,
      "step": 3410
    },
    {
      "epoch": 0.4415962714826682,
      "grad_norm": 2.4243626518666135,
      "learning_rate": 8.559653004140672e-07,
      "loss": 2.7583,
      "step": 3411
    },
    {
      "epoch": 0.4417257338900217,
      "grad_norm": 2.0677474814978734,
      "learning_rate": 8.557027958175303e-07,
      "loss": 2.4653,
      "step": 3412
    },
    {
      "epoch": 0.44185519629737513,
      "grad_norm": 2.1336206226226317,
      "learning_rate": 8.554402440952711e-07,
      "loss": 2.7041,
      "step": 3413
    },
    {
      "epoch": 0.44198465870472864,
      "grad_norm": 1.2137327826622863,
      "learning_rate": 8.551776453009212e-07,
      "loss": 2.4465,
      "step": 3414
    },
    {
      "epoch": 0.4421141211120821,
      "grad_norm": 1.6909561260831407,
      "learning_rate": 8.549149994881217e-07,
      "loss": 2.4873,
      "step": 3415
    },
    {
      "epoch": 0.4421141211120821,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5963245630264282,
      "eval_runtime": 14.0823,
      "eval_samples_per_second": 3.124,
      "eval_steps_per_second": 0.213,
      "step": 3415
    },
    {
      "epoch": 0.44224358351943555,
      "grad_norm": 1.4637293706623524,
      "learning_rate": 8.546523067105236e-07,
      "loss": 2.5273,
      "step": 3416
    },
    {
      "epoch": 0.442373045926789,
      "grad_norm": 1.9654887845983764,
      "learning_rate": 8.543895670217869e-07,
      "loss": 2.5505,
      "step": 3417
    },
    {
      "epoch": 0.44250250833414245,
      "grad_norm": 2.0927922120530504,
      "learning_rate": 8.541267804755817e-07,
      "loss": 2.4158,
      "step": 3418
    },
    {
      "epoch": 0.44263197074149596,
      "grad_norm": 1.7630548540773947,
      "learning_rate": 8.538639471255873e-07,
      "loss": 2.373,
      "step": 3419
    },
    {
      "epoch": 0.4427614331488494,
      "grad_norm": 2.3750621974147834,
      "learning_rate": 8.53601067025493e-07,
      "loss": 2.7734,
      "step": 3420
    },
    {
      "epoch": 0.4427614331488494,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.6000088453292847,
      "eval_runtime": 14.1695,
      "eval_samples_per_second": 3.105,
      "eval_steps_per_second": 0.212,
      "step": 3420
    },
    {
      "epoch": 0.44289089555620287,
      "grad_norm": 2.0183413821139617,
      "learning_rate": 8.533381402289971e-07,
      "loss": 2.5857,
      "step": 3421
    },
    {
      "epoch": 0.4430203579635563,
      "grad_norm": 2.25435004222512,
      "learning_rate": 8.530751667898077e-07,
      "loss": 2.6216,
      "step": 3422
    },
    {
      "epoch": 0.4431498203709098,
      "grad_norm": 2.356841055509401,
      "learning_rate": 8.528121467616426e-07,
      "loss": 3.0442,
      "step": 3423
    },
    {
      "epoch": 0.4432792827782633,
      "grad_norm": 1.4931724895600031,
      "learning_rate": 8.525490801982294e-07,
      "loss": 2.3027,
      "step": 3424
    },
    {
      "epoch": 0.44340874518561674,
      "grad_norm": 2.283011298172669,
      "learning_rate": 8.522859671533041e-07,
      "loss": 2.4888,
      "step": 3425
    },
    {
      "epoch": 0.44340874518561674,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.597212314605713,
      "eval_runtime": 13.8367,
      "eval_samples_per_second": 3.18,
      "eval_steps_per_second": 0.217,
      "step": 3425
    },
    {
      "epoch": 0.4435382075929702,
      "grad_norm": 1.974450548992583,
      "learning_rate": 8.520228076806131e-07,
      "loss": 2.4941,
      "step": 3426
    },
    {
      "epoch": 0.44366767000032364,
      "grad_norm": 2.048760438948378,
      "learning_rate": 8.517596018339121e-07,
      "loss": 2.2754,
      "step": 3427
    },
    {
      "epoch": 0.4437971324076771,
      "grad_norm": 2.0463369374222444,
      "learning_rate": 8.514963496669662e-07,
      "loss": 2.3474,
      "step": 3428
    },
    {
      "epoch": 0.4439265948150306,
      "grad_norm": 2.1441024974320495,
      "learning_rate": 8.512330512335501e-07,
      "loss": 2.541,
      "step": 3429
    },
    {
      "epoch": 0.44405605722238406,
      "grad_norm": 2.274939591214134,
      "learning_rate": 8.509697065874478e-07,
      "loss": 2.5029,
      "step": 3430
    },
    {
      "epoch": 0.44405605722238406,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5963689088821411,
      "eval_runtime": 14.1581,
      "eval_samples_per_second": 3.108,
      "eval_steps_per_second": 0.212,
      "step": 3430
    },
    {
      "epoch": 0.4441855196297375,
      "grad_norm": 2.0364956373765364,
      "learning_rate": 8.507063157824526e-07,
      "loss": 2.4841,
      "step": 3431
    },
    {
      "epoch": 0.44431498203709097,
      "grad_norm": 3.682776302771488,
      "learning_rate": 8.504428788723676e-07,
      "loss": 2.8289,
      "step": 3432
    },
    {
      "epoch": 0.4444444444444444,
      "grad_norm": 1.4349522246313382,
      "learning_rate": 8.501793959110049e-07,
      "loss": 2.6213,
      "step": 3433
    },
    {
      "epoch": 0.44457390685179793,
      "grad_norm": 2.0647479134090037,
      "learning_rate": 8.499158669521867e-07,
      "loss": 2.6248,
      "step": 3434
    },
    {
      "epoch": 0.4447033692591514,
      "grad_norm": 1.9191416166687798,
      "learning_rate": 8.496522920497436e-07,
      "loss": 2.4507,
      "step": 3435
    },
    {
      "epoch": 0.4447033692591514,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.595015048980713,
      "eval_runtime": 14.0327,
      "eval_samples_per_second": 3.136,
      "eval_steps_per_second": 0.214,
      "step": 3435
    },
    {
      "epoch": 0.44483283166650484,
      "grad_norm": 1.827599335538608,
      "learning_rate": 8.493886712575164e-07,
      "loss": 2.4219,
      "step": 3436
    },
    {
      "epoch": 0.4449622940738583,
      "grad_norm": 1.7583880447580256,
      "learning_rate": 8.491250046293549e-07,
      "loss": 2.5254,
      "step": 3437
    },
    {
      "epoch": 0.44509175648121174,
      "grad_norm": 3.20864614085216,
      "learning_rate": 8.488612922191187e-07,
      "loss": 2.5859,
      "step": 3438
    },
    {
      "epoch": 0.44522121888856525,
      "grad_norm": 2.329617569954473,
      "learning_rate": 8.485975340806758e-07,
      "loss": 2.405,
      "step": 3439
    },
    {
      "epoch": 0.4453506812959187,
      "grad_norm": 2.541178163993142,
      "learning_rate": 8.483337302679045e-07,
      "loss": 2.5632,
      "step": 3440
    },
    {
      "epoch": 0.4453506812959187,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.592240810394287,
      "eval_runtime": 14.1902,
      "eval_samples_per_second": 3.101,
      "eval_steps_per_second": 0.211,
      "step": 3440
    },
    {
      "epoch": 0.44548014370327216,
      "grad_norm": 9.755125786204166,
      "learning_rate": 8.480698808346921e-07,
      "loss": 3.2383,
      "step": 3441
    },
    {
      "epoch": 0.4456096061106256,
      "grad_norm": 2.5351495088939333,
      "learning_rate": 8.478059858349351e-07,
      "loss": 2.3838,
      "step": 3442
    },
    {
      "epoch": 0.44573906851797906,
      "grad_norm": 2.5097741971867453,
      "learning_rate": 8.475420453225394e-07,
      "loss": 2.6748,
      "step": 3443
    },
    {
      "epoch": 0.4458685309253326,
      "grad_norm": 1.5619751969655316,
      "learning_rate": 8.472780593514203e-07,
      "loss": 2.4473,
      "step": 3444
    },
    {
      "epoch": 0.445997993332686,
      "grad_norm": 2.106104250260571,
      "learning_rate": 8.47014027975502e-07,
      "loss": 2.9629,
      "step": 3445
    },
    {
      "epoch": 0.445997993332686,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5936611890792847,
      "eval_runtime": 14.5556,
      "eval_samples_per_second": 3.023,
      "eval_steps_per_second": 0.206,
      "step": 3445
    },
    {
      "epoch": 0.4461274557400395,
      "grad_norm": 4.360322621023151,
      "learning_rate": 8.467499512487187e-07,
      "loss": 2.5847,
      "step": 3446
    },
    {
      "epoch": 0.44625691814739293,
      "grad_norm": 2.042969660972691,
      "learning_rate": 8.46485829225013e-07,
      "loss": 2.6223,
      "step": 3447
    },
    {
      "epoch": 0.44638638055474644,
      "grad_norm": 1.65399542988463,
      "learning_rate": 8.462216619583375e-07,
      "loss": 2.4766,
      "step": 3448
    },
    {
      "epoch": 0.4465158429620999,
      "grad_norm": 1.8625609503245593,
      "learning_rate": 8.459574495026534e-07,
      "loss": 2.3845,
      "step": 3449
    },
    {
      "epoch": 0.44664530536945335,
      "grad_norm": 1.543416912225978,
      "learning_rate": 8.456931919119318e-07,
      "loss": 2.5769,
      "step": 3450
    },
    {
      "epoch": 0.44664530536945335,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5943936109542847,
      "eval_runtime": 14.9721,
      "eval_samples_per_second": 2.939,
      "eval_steps_per_second": 0.2,
      "step": 3450
    },
    {
      "epoch": 0.4467747677768068,
      "grad_norm": 1.7372836516407466,
      "learning_rate": 8.454288892401525e-07,
      "loss": 2.5654,
      "step": 3451
    },
    {
      "epoch": 0.44690423018416026,
      "grad_norm": 1.7735550649244403,
      "learning_rate": 8.451645415413049e-07,
      "loss": 2.6235,
      "step": 3452
    },
    {
      "epoch": 0.44703369259151376,
      "grad_norm": 1.8790458134274919,
      "learning_rate": 8.449001488693868e-07,
      "loss": 2.4988,
      "step": 3453
    },
    {
      "epoch": 0.4471631549988672,
      "grad_norm": 2.152908518765958,
      "learning_rate": 8.446357112784063e-07,
      "loss": 2.8228,
      "step": 3454
    },
    {
      "epoch": 0.44729261740622067,
      "grad_norm": 2.7746991861878727,
      "learning_rate": 8.443712288223799e-07,
      "loss": 2.8479,
      "step": 3455
    },
    {
      "epoch": 0.44729261740622067,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5908424854278564,
      "eval_runtime": 14.1587,
      "eval_samples_per_second": 3.108,
      "eval_steps_per_second": 0.212,
      "step": 3455
    },
    {
      "epoch": 0.4474220798135741,
      "grad_norm": 1.212695881199798,
      "learning_rate": 8.441067015553335e-07,
      "loss": 2.4155,
      "step": 3456
    },
    {
      "epoch": 0.4475515422209276,
      "grad_norm": 2.0977987197324413,
      "learning_rate": 8.438421295313024e-07,
      "loss": 2.5911,
      "step": 3457
    },
    {
      "epoch": 0.4476810046282811,
      "grad_norm": 4.683650449412505,
      "learning_rate": 8.435775128043304e-07,
      "loss": 2.4507,
      "step": 3458
    },
    {
      "epoch": 0.44781046703563454,
      "grad_norm": 1.7245941876593716,
      "learning_rate": 8.43312851428471e-07,
      "loss": 2.5325,
      "step": 3459
    },
    {
      "epoch": 0.447939929442988,
      "grad_norm": 1.8826085030779252,
      "learning_rate": 8.43048145457787e-07,
      "loss": 2.4119,
      "step": 3460
    },
    {
      "epoch": 0.447939929442988,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5928400754928589,
      "eval_runtime": 14.1723,
      "eval_samples_per_second": 3.105,
      "eval_steps_per_second": 0.212,
      "step": 3460
    },
    {
      "epoch": 0.44806939185034145,
      "grad_norm": 4.8805519417513334,
      "learning_rate": 8.427833949463495e-07,
      "loss": 2.9414,
      "step": 3461
    },
    {
      "epoch": 0.4481988542576949,
      "grad_norm": 1.4515816772874621,
      "learning_rate": 8.425185999482392e-07,
      "loss": 2.4395,
      "step": 3462
    },
    {
      "epoch": 0.4483283166650484,
      "grad_norm": 1.768673429111786,
      "learning_rate": 8.422537605175461e-07,
      "loss": 2.3357,
      "step": 3463
    },
    {
      "epoch": 0.44845777907240186,
      "grad_norm": 4.9546667520887695,
      "learning_rate": 8.419888767083688e-07,
      "loss": 2.5649,
      "step": 3464
    },
    {
      "epoch": 0.4485872414797553,
      "grad_norm": 1.5034713072550403,
      "learning_rate": 8.417239485748154e-07,
      "loss": 2.582,
      "step": 3465
    },
    {
      "epoch": 0.4485872414797553,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.593550205230713,
      "eval_runtime": 14.4344,
      "eval_samples_per_second": 3.048,
      "eval_steps_per_second": 0.208,
      "step": 3465
    },
    {
      "epoch": 0.44871670388710877,
      "grad_norm": 2.4904044082507673,
      "learning_rate": 8.414589761710027e-07,
      "loss": 2.2974,
      "step": 3466
    },
    {
      "epoch": 0.4488461662944622,
      "grad_norm": 1.4453400295254624,
      "learning_rate": 8.411939595510565e-07,
      "loss": 2.3704,
      "step": 3467
    },
    {
      "epoch": 0.44897562870181573,
      "grad_norm": 2.354240400747779,
      "learning_rate": 8.409288987691122e-07,
      "loss": 2.573,
      "step": 3468
    },
    {
      "epoch": 0.4491050911091692,
      "grad_norm": 2.125029297268219,
      "learning_rate": 8.406637938793136e-07,
      "loss": 2.4602,
      "step": 3469
    },
    {
      "epoch": 0.44923455351652264,
      "grad_norm": 2.653936148520379,
      "learning_rate": 8.403986449358137e-07,
      "loss": 2.6267,
      "step": 3470
    },
    {
      "epoch": 0.44923455351652264,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.587202548980713,
      "eval_runtime": 14.2485,
      "eval_samples_per_second": 3.088,
      "eval_steps_per_second": 0.211,
      "step": 3470
    },
    {
      "epoch": 0.4493640159238761,
      "grad_norm": 1.542150110411501,
      "learning_rate": 8.401334519927748e-07,
      "loss": 2.6687,
      "step": 3471
    },
    {
      "epoch": 0.44949347833122955,
      "grad_norm": 1.970404376171036,
      "learning_rate": 8.398682151043675e-07,
      "loss": 2.5945,
      "step": 3472
    },
    {
      "epoch": 0.44962294073858305,
      "grad_norm": 2.75045176495899,
      "learning_rate": 8.396029343247722e-07,
      "loss": 2.3618,
      "step": 3473
    },
    {
      "epoch": 0.4497524031459365,
      "grad_norm": 1.8640981826818483,
      "learning_rate": 8.39337609708178e-07,
      "loss": 2.4033,
      "step": 3474
    },
    {
      "epoch": 0.44988186555328996,
      "grad_norm": 1.8051303242980334,
      "learning_rate": 8.390722413087823e-07,
      "loss": 2.4114,
      "step": 3475
    },
    {
      "epoch": 0.44988186555328996,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.590775966644287,
      "eval_runtime": 14.2357,
      "eval_samples_per_second": 3.091,
      "eval_steps_per_second": 0.211,
      "step": 3475
    },
    {
      "epoch": 0.4500113279606434,
      "grad_norm": 1.5747446483541916,
      "learning_rate": 8.388068291807924e-07,
      "loss": 2.4559,
      "step": 3476
    },
    {
      "epoch": 0.45014079036799687,
      "grad_norm": 1.8801924727738788,
      "learning_rate": 8.38541373378424e-07,
      "loss": 2.3982,
      "step": 3477
    },
    {
      "epoch": 0.4502702527753504,
      "grad_norm": 3.0946284691450274,
      "learning_rate": 8.382758739559017e-07,
      "loss": 2.4312,
      "step": 3478
    },
    {
      "epoch": 0.45039971518270383,
      "grad_norm": 1.9475762253494309,
      "learning_rate": 8.380103309674594e-07,
      "loss": 2.2505,
      "step": 3479
    },
    {
      "epoch": 0.4505291775900573,
      "grad_norm": 1.952188856399258,
      "learning_rate": 8.377447444673395e-07,
      "loss": 2.3733,
      "step": 3480
    },
    {
      "epoch": 0.4505291775900573,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.592817783355713,
      "eval_runtime": 13.6523,
      "eval_samples_per_second": 3.223,
      "eval_steps_per_second": 0.22,
      "step": 3480
    },
    {
      "epoch": 0.45065863999741074,
      "grad_norm": 1.6809954356463555,
      "learning_rate": 8.374791145097931e-07,
      "loss": 2.5569,
      "step": 3481
    },
    {
      "epoch": 0.4507881024047642,
      "grad_norm": 1.983387145091062,
      "learning_rate": 8.37213441149081e-07,
      "loss": 2.4927,
      "step": 3482
    },
    {
      "epoch": 0.4509175648121177,
      "grad_norm": 1.4753342091772366,
      "learning_rate": 8.369477244394722e-07,
      "loss": 2.4827,
      "step": 3483
    },
    {
      "epoch": 0.45104702721947115,
      "grad_norm": 1.7076497338199936,
      "learning_rate": 8.366819644352446e-07,
      "loss": 2.8013,
      "step": 3484
    },
    {
      "epoch": 0.4511764896268246,
      "grad_norm": 1.5460583439837456,
      "learning_rate": 8.364161611906852e-07,
      "loss": 2.3591,
      "step": 3485
    },
    {
      "epoch": 0.4511764896268246,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.591352939605713,
      "eval_runtime": 14.0091,
      "eval_samples_per_second": 3.141,
      "eval_steps_per_second": 0.214,
      "step": 3485
    },
    {
      "epoch": 0.45130595203417806,
      "grad_norm": 1.8335786141903636,
      "learning_rate": 8.361503147600897e-07,
      "loss": 2.4473,
      "step": 3486
    },
    {
      "epoch": 0.4514354144415315,
      "grad_norm": 2.754884804185985,
      "learning_rate": 8.358844251977626e-07,
      "loss": 2.4214,
      "step": 3487
    },
    {
      "epoch": 0.451564876848885,
      "grad_norm": 3.5112997594448667,
      "learning_rate": 8.356184925580173e-07,
      "loss": 2.4131,
      "step": 3488
    },
    {
      "epoch": 0.4516943392562385,
      "grad_norm": 1.6092942665609502,
      "learning_rate": 8.353525168951756e-07,
      "loss": 2.4714,
      "step": 3489
    },
    {
      "epoch": 0.45182380166359193,
      "grad_norm": 2.125823252390096,
      "learning_rate": 8.35086498263569e-07,
      "loss": 2.4609,
      "step": 3490
    },
    {
      "epoch": 0.45182380166359193,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5899990797042847,
      "eval_runtime": 14.5263,
      "eval_samples_per_second": 3.029,
      "eval_steps_per_second": 0.207,
      "step": 3490
    },
    {
      "epoch": 0.4519532640709454,
      "grad_norm": 2.159282465658603,
      "learning_rate": 8.348204367175367e-07,
      "loss": 2.4088,
      "step": 3491
    },
    {
      "epoch": 0.4520827264782989,
      "grad_norm": 1.7870489868656576,
      "learning_rate": 8.345543323114272e-07,
      "loss": 2.3232,
      "step": 3492
    },
    {
      "epoch": 0.45221218888565234,
      "grad_norm": 1.422530218838628,
      "learning_rate": 8.342881850995979e-07,
      "loss": 2.5317,
      "step": 3493
    },
    {
      "epoch": 0.4523416512930058,
      "grad_norm": 1.7856068188226795,
      "learning_rate": 8.340219951364151e-07,
      "loss": 2.4795,
      "step": 3494
    },
    {
      "epoch": 0.45247111370035925,
      "grad_norm": 1.5030901226106912,
      "learning_rate": 8.337557624762525e-07,
      "loss": 2.4221,
      "step": 3495
    },
    {
      "epoch": 0.45247111370035925,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5936168432235718,
      "eval_runtime": 14.9019,
      "eval_samples_per_second": 2.953,
      "eval_steps_per_second": 0.201,
      "step": 3495
    },
    {
      "epoch": 0.4526005761077127,
      "grad_norm": 1.5523475542841818,
      "learning_rate": 8.334894871734946e-07,
      "loss": 2.4216,
      "step": 3496
    },
    {
      "epoch": 0.4527300385150662,
      "grad_norm": 1.960794740924934,
      "learning_rate": 8.332231692825329e-07,
      "loss": 2.3936,
      "step": 3497
    },
    {
      "epoch": 0.45285950092241967,
      "grad_norm": 2.5424394583593775,
      "learning_rate": 8.329568088577683e-07,
      "loss": 2.418,
      "step": 3498
    },
    {
      "epoch": 0.4529889633297731,
      "grad_norm": 1.691201976792941,
      "learning_rate": 8.326904059536105e-07,
      "loss": 2.5312,
      "step": 3499
    },
    {
      "epoch": 0.4531184257371266,
      "grad_norm": 1.3963541506015975,
      "learning_rate": 8.324239606244775e-07,
      "loss": 2.2056,
      "step": 3500
    },
    {
      "epoch": 0.4531184257371266,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5943492650985718,
      "eval_runtime": 14.3213,
      "eval_samples_per_second": 3.072,
      "eval_steps_per_second": 0.209,
      "step": 3500
    },
    {
      "epoch": 0.45324788814448,
      "grad_norm": 4.037712625281316,
      "learning_rate": 8.321574729247963e-07,
      "loss": 3.0259,
      "step": 3501
    },
    {
      "epoch": 0.45337735055183354,
      "grad_norm": 2.7086904222357995,
      "learning_rate": 8.318909429090025e-07,
      "loss": 2.6284,
      "step": 3502
    },
    {
      "epoch": 0.453506812959187,
      "grad_norm": 2.138947189395762,
      "learning_rate": 8.316243706315399e-07,
      "loss": 2.6094,
      "step": 3503
    },
    {
      "epoch": 0.45363627536654044,
      "grad_norm": 2.052154068345744,
      "learning_rate": 8.313577561468616e-07,
      "loss": 2.7842,
      "step": 3504
    },
    {
      "epoch": 0.4537657377738939,
      "grad_norm": 1.5251842087983352,
      "learning_rate": 8.31091099509429e-07,
      "loss": 2.4153,
      "step": 3505
    },
    {
      "epoch": 0.4537657377738939,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.589888095855713,
      "eval_runtime": 13.6086,
      "eval_samples_per_second": 3.233,
      "eval_steps_per_second": 0.22,
      "step": 3505
    },
    {
      "epoch": 0.45389520018124735,
      "grad_norm": 1.8900255833797497,
      "learning_rate": 8.30824400773712e-07,
      "loss": 2.7739,
      "step": 3506
    },
    {
      "epoch": 0.45402466258860086,
      "grad_norm": 1.3044646331954206,
      "learning_rate": 8.305576599941893e-07,
      "loss": 2.384,
      "step": 3507
    },
    {
      "epoch": 0.4541541249959543,
      "grad_norm": 3.646796258599669,
      "learning_rate": 8.302908772253479e-07,
      "loss": 2.5906,
      "step": 3508
    },
    {
      "epoch": 0.45428358740330776,
      "grad_norm": 2.5083107607117356,
      "learning_rate": 8.300240525216839e-07,
      "loss": 2.3652,
      "step": 3509
    },
    {
      "epoch": 0.4544130498106612,
      "grad_norm": 1.7266092241583402,
      "learning_rate": 8.297571859377017e-07,
      "loss": 2.4023,
      "step": 3510
    },
    {
      "epoch": 0.4544130498106612,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.592041015625,
      "eval_runtime": 14.584,
      "eval_samples_per_second": 3.017,
      "eval_steps_per_second": 0.206,
      "step": 3510
    },
    {
      "epoch": 0.45454251221801467,
      "grad_norm": 1.755901879661211,
      "learning_rate": 8.294902775279139e-07,
      "loss": 2.5237,
      "step": 3511
    },
    {
      "epoch": 0.4546719746253682,
      "grad_norm": 4.733184420278606,
      "learning_rate": 8.29223327346842e-07,
      "loss": 2.9595,
      "step": 3512
    },
    {
      "epoch": 0.45480143703272163,
      "grad_norm": 2.0147043517710794,
      "learning_rate": 8.289563354490161e-07,
      "loss": 2.5232,
      "step": 3513
    },
    {
      "epoch": 0.4549308994400751,
      "grad_norm": 2.9454545173513087,
      "learning_rate": 8.286893018889747e-07,
      "loss": 2.6023,
      "step": 3514
    },
    {
      "epoch": 0.45506036184742854,
      "grad_norm": 1.9390849326008988,
      "learning_rate": 8.284222267212648e-07,
      "loss": 2.3904,
      "step": 3515
    },
    {
      "epoch": 0.45506036184742854,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5878239870071411,
      "eval_runtime": 14.4249,
      "eval_samples_per_second": 3.05,
      "eval_steps_per_second": 0.208,
      "step": 3515
    },
    {
      "epoch": 0.455189824254782,
      "grad_norm": 2.407469520826565,
      "learning_rate": 8.28155110000442e-07,
      "loss": 2.4973,
      "step": 3516
    },
    {
      "epoch": 0.4553192866621355,
      "grad_norm": 2.9037194799249355,
      "learning_rate": 8.278879517810701e-07,
      "loss": 2.7073,
      "step": 3517
    },
    {
      "epoch": 0.45544874906948896,
      "grad_norm": 2.0242189250031073,
      "learning_rate": 8.276207521177215e-07,
      "loss": 2.5239,
      "step": 3518
    },
    {
      "epoch": 0.4555782114768424,
      "grad_norm": 1.620279429623691,
      "learning_rate": 8.273535110649777e-07,
      "loss": 2.4004,
      "step": 3519
    },
    {
      "epoch": 0.45570767388419586,
      "grad_norm": 1.4933697348122157,
      "learning_rate": 8.270862286774277e-07,
      "loss": 2.624,
      "step": 3520
    },
    {
      "epoch": 0.45570767388419586,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.592729091644287,
      "eval_runtime": 13.7174,
      "eval_samples_per_second": 3.208,
      "eval_steps_per_second": 0.219,
      "step": 3520
    },
    {
      "epoch": 0.4558371362915493,
      "grad_norm": 2.2558784021244076,
      "learning_rate": 8.268189050096692e-07,
      "loss": 2.3595,
      "step": 3521
    },
    {
      "epoch": 0.4559665986989028,
      "grad_norm": 1.8697621736823407,
      "learning_rate": 8.265515401163088e-07,
      "loss": 2.5815,
      "step": 3522
    },
    {
      "epoch": 0.4560960611062563,
      "grad_norm": 2.6995572062800197,
      "learning_rate": 8.26284134051961e-07,
      "loss": 2.9189,
      "step": 3523
    },
    {
      "epoch": 0.45622552351360973,
      "grad_norm": 1.6164022513948326,
      "learning_rate": 8.260166868712492e-07,
      "loss": 2.126,
      "step": 3524
    },
    {
      "epoch": 0.4563549859209632,
      "grad_norm": 1.6266596175980144,
      "learning_rate": 8.257491986288045e-07,
      "loss": 2.6719,
      "step": 3525
    },
    {
      "epoch": 0.4563549859209632,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5919743776321411,
      "eval_runtime": 13.5262,
      "eval_samples_per_second": 3.253,
      "eval_steps_per_second": 0.222,
      "step": 3525
    },
    {
      "epoch": 0.45648444832831664,
      "grad_norm": 2.2481635107459734,
      "learning_rate": 8.254816693792669e-07,
      "loss": 2.6448,
      "step": 3526
    },
    {
      "epoch": 0.45661391073567015,
      "grad_norm": 1.264468241384177,
      "learning_rate": 8.252140991772849e-07,
      "loss": 2.4055,
      "step": 3527
    },
    {
      "epoch": 0.4567433731430236,
      "grad_norm": 1.3532541377530152,
      "learning_rate": 8.249464880775148e-07,
      "loss": 2.3101,
      "step": 3528
    },
    {
      "epoch": 0.45687283555037705,
      "grad_norm": 2.644552174923601,
      "learning_rate": 8.246788361346219e-07,
      "loss": 2.5945,
      "step": 3529
    },
    {
      "epoch": 0.4570022979577305,
      "grad_norm": 1.5353137928902914,
      "learning_rate": 8.244111434032795e-07,
      "loss": 2.6028,
      "step": 3530
    },
    {
      "epoch": 0.4570022979577305,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5955921411514282,
      "eval_runtime": 14.0368,
      "eval_samples_per_second": 3.135,
      "eval_steps_per_second": 0.214,
      "step": 3530
    },
    {
      "epoch": 0.45713176036508396,
      "grad_norm": 1.817363881152656,
      "learning_rate": 8.24143409938169e-07,
      "loss": 2.5415,
      "step": 3531
    },
    {
      "epoch": 0.45726122277243747,
      "grad_norm": 1.6082781305740184,
      "learning_rate": 8.238756357939807e-07,
      "loss": 2.4641,
      "step": 3532
    },
    {
      "epoch": 0.4573906851797909,
      "grad_norm": 1.235996569190046,
      "learning_rate": 8.236078210254128e-07,
      "loss": 2.353,
      "step": 3533
    },
    {
      "epoch": 0.4575201475871444,
      "grad_norm": 1.7162396031100193,
      "learning_rate": 8.233399656871718e-07,
      "loss": 2.6216,
      "step": 3534
    },
    {
      "epoch": 0.45764960999449783,
      "grad_norm": 2.8453265415094204,
      "learning_rate": 8.230720698339726e-07,
      "loss": 2.4707,
      "step": 3535
    },
    {
      "epoch": 0.45764960999449783,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.593461513519287,
      "eval_runtime": 14.0636,
      "eval_samples_per_second": 3.129,
      "eval_steps_per_second": 0.213,
      "step": 3535
    },
    {
      "epoch": 0.45777907240185134,
      "grad_norm": 2.378157097105098,
      "learning_rate": 8.228041335205383e-07,
      "loss": 2.3474,
      "step": 3536
    },
    {
      "epoch": 0.4579085348092048,
      "grad_norm": 3.043788773503862,
      "learning_rate": 8.225361568016006e-07,
      "loss": 2.6514,
      "step": 3537
    },
    {
      "epoch": 0.45803799721655825,
      "grad_norm": 2.0298226832508086,
      "learning_rate": 8.22268139731899e-07,
      "loss": 2.4148,
      "step": 3538
    },
    {
      "epoch": 0.4581674596239117,
      "grad_norm": 1.951288476113402,
      "learning_rate": 8.220000823661814e-07,
      "loss": 2.2568,
      "step": 3539
    },
    {
      "epoch": 0.45829692203126515,
      "grad_norm": 1.7093127850108965,
      "learning_rate": 8.217319847592039e-07,
      "loss": 2.4971,
      "step": 3540
    },
    {
      "epoch": 0.45829692203126515,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5949040651321411,
      "eval_runtime": 13.3483,
      "eval_samples_per_second": 3.296,
      "eval_steps_per_second": 0.225,
      "step": 3540
    },
    {
      "epoch": 0.45842638443861866,
      "grad_norm": 2.840784446007513,
      "learning_rate": 8.21463846965731e-07,
      "loss": 2.9341,
      "step": 3541
    },
    {
      "epoch": 0.4585558468459721,
      "grad_norm": 1.5335644139393236,
      "learning_rate": 8.211956690405351e-07,
      "loss": 2.5134,
      "step": 3542
    },
    {
      "epoch": 0.45868530925332557,
      "grad_norm": 1.3739263786173637,
      "learning_rate": 8.209274510383972e-07,
      "loss": 2.4514,
      "step": 3543
    },
    {
      "epoch": 0.458814771660679,
      "grad_norm": 1.2930812102965403,
      "learning_rate": 8.206591930141061e-07,
      "loss": 2.3064,
      "step": 3544
    },
    {
      "epoch": 0.4589442340680325,
      "grad_norm": 1.9304972048455111,
      "learning_rate": 8.20390895022459e-07,
      "loss": 2.6584,
      "step": 3545
    },
    {
      "epoch": 0.4589442340680325,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5948818922042847,
      "eval_runtime": 13.8604,
      "eval_samples_per_second": 3.175,
      "eval_steps_per_second": 0.216,
      "step": 3545
    },
    {
      "epoch": 0.459073696475386,
      "grad_norm": 1.7601059949576592,
      "learning_rate": 8.201225571182613e-07,
      "loss": 2.3164,
      "step": 3546
    },
    {
      "epoch": 0.45920315888273944,
      "grad_norm": 2.182314595338066,
      "learning_rate": 8.198541793563262e-07,
      "loss": 2.6887,
      "step": 3547
    },
    {
      "epoch": 0.4593326212900929,
      "grad_norm": 1.9765973941016117,
      "learning_rate": 8.195857617914756e-07,
      "loss": 2.3545,
      "step": 3548
    },
    {
      "epoch": 0.45946208369744634,
      "grad_norm": 1.808539425278726,
      "learning_rate": 8.193173044785389e-07,
      "loss": 2.5796,
      "step": 3549
    },
    {
      "epoch": 0.4595915461047998,
      "grad_norm": 2.3527738038771657,
      "learning_rate": 8.190488074723543e-07,
      "loss": 2.562,
      "step": 3550
    },
    {
      "epoch": 0.4595915461047998,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5921741724014282,
      "eval_runtime": 14.4279,
      "eval_samples_per_second": 3.05,
      "eval_steps_per_second": 0.208,
      "step": 3550
    },
    {
      "epoch": 0.4597210085121533,
      "grad_norm": 2.538172015425518,
      "learning_rate": 8.187802708277676e-07,
      "loss": 2.6057,
      "step": 3551
    },
    {
      "epoch": 0.45985047091950676,
      "grad_norm": 1.458300915715859,
      "learning_rate": 8.185116945996329e-07,
      "loss": 2.4456,
      "step": 3552
    },
    {
      "epoch": 0.4599799333268602,
      "grad_norm": 1.541523273447448,
      "learning_rate": 8.182430788428119e-07,
      "loss": 2.3953,
      "step": 3553
    },
    {
      "epoch": 0.46010939573421367,
      "grad_norm": 2.1204613407306554,
      "learning_rate": 8.179744236121753e-07,
      "loss": 2.5063,
      "step": 3554
    },
    {
      "epoch": 0.4602388581415671,
      "grad_norm": 1.857757041972434,
      "learning_rate": 8.177057289626015e-07,
      "loss": 2.5691,
      "step": 3555
    },
    {
      "epoch": 0.4602388581415671,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5920188426971436,
      "eval_runtime": 14.1103,
      "eval_samples_per_second": 3.118,
      "eval_steps_per_second": 0.213,
      "step": 3555
    },
    {
      "epoch": 0.46036832054892063,
      "grad_norm": 1.6251962162355167,
      "learning_rate": 8.174369949489766e-07,
      "loss": 2.7029,
      "step": 3556
    },
    {
      "epoch": 0.4604977829562741,
      "grad_norm": 1.902531341562306,
      "learning_rate": 8.171682216261947e-07,
      "loss": 2.6323,
      "step": 3557
    },
    {
      "epoch": 0.46062724536362754,
      "grad_norm": 4.6452138828166145,
      "learning_rate": 8.168994090491586e-07,
      "loss": 2.6743,
      "step": 3558
    },
    {
      "epoch": 0.460756707770981,
      "grad_norm": 3.2342474410885282,
      "learning_rate": 8.166305572727786e-07,
      "loss": 2.6072,
      "step": 3559
    },
    {
      "epoch": 0.46088617017833444,
      "grad_norm": 1.6112562974224787,
      "learning_rate": 8.163616663519731e-07,
      "loss": 2.5098,
      "step": 3560
    },
    {
      "epoch": 0.46088617017833444,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5934836864471436,
      "eval_runtime": 14.1883,
      "eval_samples_per_second": 3.101,
      "eval_steps_per_second": 0.211,
      "step": 3560
    },
    {
      "epoch": 0.46101563258568795,
      "grad_norm": 4.058965869435389,
      "learning_rate": 8.160927363416687e-07,
      "loss": 2.6421,
      "step": 3561
    },
    {
      "epoch": 0.4611450949930414,
      "grad_norm": 1.9515517112472651,
      "learning_rate": 8.158237672967993e-07,
      "loss": 2.4978,
      "step": 3562
    },
    {
      "epoch": 0.46127455740039486,
      "grad_norm": 1.692602156188153,
      "learning_rate": 8.155547592723077e-07,
      "loss": 2.505,
      "step": 3563
    },
    {
      "epoch": 0.4614040198077483,
      "grad_norm": 2.112893395470927,
      "learning_rate": 8.152857123231442e-07,
      "loss": 2.7878,
      "step": 3564
    },
    {
      "epoch": 0.46153348221510176,
      "grad_norm": 2.1354769230516566,
      "learning_rate": 8.15016626504267e-07,
      "loss": 2.7136,
      "step": 3565
    },
    {
      "epoch": 0.46153348221510176,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5921076536178589,
      "eval_runtime": 14.1796,
      "eval_samples_per_second": 3.103,
      "eval_steps_per_second": 0.212,
      "step": 3565
    },
    {
      "epoch": 0.4616629446224553,
      "grad_norm": 2.462216370965521,
      "learning_rate": 8.147475018706425e-07,
      "loss": 2.5752,
      "step": 3566
    },
    {
      "epoch": 0.4617924070298087,
      "grad_norm": 3.175714550415163,
      "learning_rate": 8.144783384772445e-07,
      "loss": 2.8098,
      "step": 3567
    },
    {
      "epoch": 0.4619218694371622,
      "grad_norm": 1.820835124439004,
      "learning_rate": 8.142091363790555e-07,
      "loss": 2.4871,
      "step": 3568
    },
    {
      "epoch": 0.46205133184451563,
      "grad_norm": 2.433789894150037,
      "learning_rate": 8.139398956310654e-07,
      "loss": 2.4468,
      "step": 3569
    },
    {
      "epoch": 0.4621807942518691,
      "grad_norm": 3.9808301447254513,
      "learning_rate": 8.136706162882719e-07,
      "loss": 2.6421,
      "step": 3570
    },
    {
      "epoch": 0.4621807942518691,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.589888095855713,
      "eval_runtime": 13.6632,
      "eval_samples_per_second": 3.22,
      "eval_steps_per_second": 0.22,
      "step": 3570
    },
    {
      "epoch": 0.4623102566592226,
      "grad_norm": 2.3166104811786936,
      "learning_rate": 8.134012984056807e-07,
      "loss": 2.6492,
      "step": 3571
    },
    {
      "epoch": 0.46243971906657605,
      "grad_norm": 1.5633254614133703,
      "learning_rate": 8.131319420383055e-07,
      "loss": 2.5996,
      "step": 3572
    },
    {
      "epoch": 0.4625691814739295,
      "grad_norm": 2.5436967709171507,
      "learning_rate": 8.12862547241168e-07,
      "loss": 2.6721,
      "step": 3573
    },
    {
      "epoch": 0.46269864388128296,
      "grad_norm": 1.336551219749576,
      "learning_rate": 8.125931140692976e-07,
      "loss": 2.2827,
      "step": 3574
    },
    {
      "epoch": 0.4628281062886364,
      "grad_norm": 3.6220279251908925,
      "learning_rate": 8.123236425777311e-07,
      "loss": 2.6694,
      "step": 3575
    },
    {
      "epoch": 0.4628281062886364,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5941494703292847,
      "eval_runtime": 14.3569,
      "eval_samples_per_second": 3.065,
      "eval_steps_per_second": 0.209,
      "step": 3575
    },
    {
      "epoch": 0.4629575686959899,
      "grad_norm": 1.6404182937055733,
      "learning_rate": 8.120541328215138e-07,
      "loss": 2.5371,
      "step": 3576
    },
    {
      "epoch": 0.46308703110334337,
      "grad_norm": 1.316217332753309,
      "learning_rate": 8.117845848556985e-07,
      "loss": 2.5742,
      "step": 3577
    },
    {
      "epoch": 0.4632164935106968,
      "grad_norm": 3.248335397277242,
      "learning_rate": 8.115149987353457e-07,
      "loss": 2.6111,
      "step": 3578
    },
    {
      "epoch": 0.4633459559180503,
      "grad_norm": 2.858708893883924,
      "learning_rate": 8.112453745155241e-07,
      "loss": 2.3638,
      "step": 3579
    },
    {
      "epoch": 0.4634754183254038,
      "grad_norm": 1.523999816902224,
      "learning_rate": 8.109757122513095e-07,
      "loss": 2.5203,
      "step": 3580
    },
    {
      "epoch": 0.4634754183254038,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5940829515457153,
      "eval_runtime": 14.5063,
      "eval_samples_per_second": 3.033,
      "eval_steps_per_second": 0.207,
      "step": 3580
    },
    {
      "epoch": 0.46360488073275724,
      "grad_norm": 2.0406504523822786,
      "learning_rate": 8.107060119977862e-07,
      "loss": 2.4436,
      "step": 3581
    },
    {
      "epoch": 0.4637343431401107,
      "grad_norm": 2.431364476884996,
      "learning_rate": 8.104362738100459e-07,
      "loss": 2.7263,
      "step": 3582
    },
    {
      "epoch": 0.46386380554746415,
      "grad_norm": 2.2838751982511174,
      "learning_rate": 8.101664977431882e-07,
      "loss": 2.4674,
      "step": 3583
    },
    {
      "epoch": 0.4639932679548176,
      "grad_norm": 1.9316995644600121,
      "learning_rate": 8.098966838523199e-07,
      "loss": 2.6709,
      "step": 3584
    },
    {
      "epoch": 0.4641227303621711,
      "grad_norm": 1.4050379286807781,
      "learning_rate": 8.096268321925564e-07,
      "loss": 2.4502,
      "step": 3585
    },
    {
      "epoch": 0.4641227303621711,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5940829515457153,
      "eval_runtime": 13.3208,
      "eval_samples_per_second": 3.303,
      "eval_steps_per_second": 0.225,
      "step": 3585
    },
    {
      "epoch": 0.46425219276952456,
      "grad_norm": 1.66131694065598,
      "learning_rate": 8.093569428190201e-07,
      "loss": 2.3375,
      "step": 3586
    },
    {
      "epoch": 0.464381655176878,
      "grad_norm": 1.3022443891589117,
      "learning_rate": 8.090870157868416e-07,
      "loss": 2.2363,
      "step": 3587
    },
    {
      "epoch": 0.46451111758423147,
      "grad_norm": 1.448573845830862,
      "learning_rate": 8.08817051151159e-07,
      "loss": 2.4866,
      "step": 3588
    },
    {
      "epoch": 0.4646405799915849,
      "grad_norm": 1.4590393426530284,
      "learning_rate": 8.085470489671177e-07,
      "loss": 2.4417,
      "step": 3589
    },
    {
      "epoch": 0.46477004239893843,
      "grad_norm": 2.42888655972702,
      "learning_rate": 8.082770092898711e-07,
      "loss": 2.3853,
      "step": 3590
    },
    {
      "epoch": 0.46477004239893843,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5933283567428589,
      "eval_runtime": 13.7539,
      "eval_samples_per_second": 3.199,
      "eval_steps_per_second": 0.218,
      "step": 3590
    },
    {
      "epoch": 0.4648995048062919,
      "grad_norm": 1.9350854766805095,
      "learning_rate": 8.080069321745808e-07,
      "loss": 2.5454,
      "step": 3591
    },
    {
      "epoch": 0.46502896721364534,
      "grad_norm": 1.6349943734600694,
      "learning_rate": 8.077368176764152e-07,
      "loss": 2.738,
      "step": 3592
    },
    {
      "epoch": 0.4651584296209988,
      "grad_norm": 1.8505707847809654,
      "learning_rate": 8.074666658505505e-07,
      "loss": 2.5532,
      "step": 3593
    },
    {
      "epoch": 0.46528789202835225,
      "grad_norm": 3.3952495839951364,
      "learning_rate": 8.07196476752171e-07,
      "loss": 2.7383,
      "step": 3594
    },
    {
      "epoch": 0.46541735443570575,
      "grad_norm": 2.8892647458860887,
      "learning_rate": 8.069262504364682e-07,
      "loss": 2.4192,
      "step": 3595
    },
    {
      "epoch": 0.46541735443570575,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5948597192764282,
      "eval_runtime": 13.6196,
      "eval_samples_per_second": 3.231,
      "eval_steps_per_second": 0.22,
      "step": 3595
    },
    {
      "epoch": 0.4655468168430592,
      "grad_norm": 2.133267236864094,
      "learning_rate": 8.066559869586412e-07,
      "loss": 2.7998,
      "step": 3596
    },
    {
      "epoch": 0.46567627925041266,
      "grad_norm": 3.3718440265536542,
      "learning_rate": 8.063856863738969e-07,
      "loss": 2.6304,
      "step": 3597
    },
    {
      "epoch": 0.4658057416577661,
      "grad_norm": 1.9952924172187647,
      "learning_rate": 8.061153487374497e-07,
      "loss": 2.7131,
      "step": 3598
    },
    {
      "epoch": 0.46593520406511957,
      "grad_norm": 1.5456239466364916,
      "learning_rate": 8.058449741045214e-07,
      "loss": 2.4812,
      "step": 3599
    },
    {
      "epoch": 0.4660646664724731,
      "grad_norm": 2.583115023447738,
      "learning_rate": 8.055745625303415e-07,
      "loss": 2.4417,
      "step": 3600
    },
    {
      "epoch": 0.4660646664724731,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5906428098678589,
      "eval_runtime": 14.186,
      "eval_samples_per_second": 3.102,
      "eval_steps_per_second": 0.211,
      "step": 3600
    },
    {
      "epoch": 0.46619412887982653,
      "grad_norm": 1.1488507797715302,
      "learning_rate": 8.053041140701473e-07,
      "loss": 2.2869,
      "step": 3601
    },
    {
      "epoch": 0.46632359128718,
      "grad_norm": 1.7556191640489143,
      "learning_rate": 8.050336287791832e-07,
      "loss": 2.3716,
      "step": 3602
    },
    {
      "epoch": 0.46645305369453344,
      "grad_norm": 1.8969412618488941,
      "learning_rate": 8.047631067127012e-07,
      "loss": 2.5596,
      "step": 3603
    },
    {
      "epoch": 0.4665825161018869,
      "grad_norm": 1.5199006891523428,
      "learning_rate": 8.044925479259609e-07,
      "loss": 2.4817,
      "step": 3604
    },
    {
      "epoch": 0.4667119785092404,
      "grad_norm": 1.9122988107655003,
      "learning_rate": 8.042219524742298e-07,
      "loss": 2.4094,
      "step": 3605
    },
    {
      "epoch": 0.4667119785092404,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.591264247894287,
      "eval_runtime": 14.6212,
      "eval_samples_per_second": 3.009,
      "eval_steps_per_second": 0.205,
      "step": 3605
    },
    {
      "epoch": 0.46684144091659385,
      "grad_norm": 1.6565579502051928,
      "learning_rate": 8.039513204127822e-07,
      "loss": 2.5667,
      "step": 3606
    },
    {
      "epoch": 0.4669709033239473,
      "grad_norm": 1.5887212603423024,
      "learning_rate": 8.036806517969001e-07,
      "loss": 2.7041,
      "step": 3607
    },
    {
      "epoch": 0.46710036573130076,
      "grad_norm": 1.6585018108929768,
      "learning_rate": 8.034099466818732e-07,
      "loss": 2.5796,
      "step": 3608
    },
    {
      "epoch": 0.4672298281386542,
      "grad_norm": 1.414804864498699,
      "learning_rate": 8.031392051229984e-07,
      "loss": 2.4985,
      "step": 3609
    },
    {
      "epoch": 0.4673592905460077,
      "grad_norm": 2.34889392652913,
      "learning_rate": 8.028684271755806e-07,
      "loss": 2.2983,
      "step": 3610
    },
    {
      "epoch": 0.4673592905460077,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5906649827957153,
      "eval_runtime": 13.5233,
      "eval_samples_per_second": 3.254,
      "eval_steps_per_second": 0.222,
      "step": 3610
    },
    {
      "epoch": 0.4674887529533612,
      "grad_norm": 1.8952544735961896,
      "learning_rate": 8.02597612894931e-07,
      "loss": 2.3274,
      "step": 3611
    },
    {
      "epoch": 0.46761821536071463,
      "grad_norm": 2.328226795461395,
      "learning_rate": 8.023267623363693e-07,
      "loss": 2.697,
      "step": 3612
    },
    {
      "epoch": 0.4677476777680681,
      "grad_norm": 1.575542729139767,
      "learning_rate": 8.020558755552221e-07,
      "loss": 2.6398,
      "step": 3613
    },
    {
      "epoch": 0.46787714017542154,
      "grad_norm": 2.1777157593498475,
      "learning_rate": 8.017849526068235e-07,
      "loss": 2.4082,
      "step": 3614
    },
    {
      "epoch": 0.46800660258277504,
      "grad_norm": 2.2741259184806144,
      "learning_rate": 8.015139935465152e-07,
      "loss": 2.4932,
      "step": 3615
    },
    {
      "epoch": 0.46800660258277504,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.591264247894287,
      "eval_runtime": 13.5983,
      "eval_samples_per_second": 3.236,
      "eval_steps_per_second": 0.221,
      "step": 3615
    },
    {
      "epoch": 0.4681360649901285,
      "grad_norm": 2.939417811929448,
      "learning_rate": 8.01242998429646e-07,
      "loss": 2.3882,
      "step": 3616
    },
    {
      "epoch": 0.46826552739748195,
      "grad_norm": 4.701175417229919,
      "learning_rate": 8.00971967311572e-07,
      "loss": 2.8022,
      "step": 3617
    },
    {
      "epoch": 0.4683949898048354,
      "grad_norm": 2.5364959994065366,
      "learning_rate": 8.007009002476568e-07,
      "loss": 2.5098,
      "step": 3618
    },
    {
      "epoch": 0.4685244522121889,
      "grad_norm": 3.5895201998707016,
      "learning_rate": 8.004297972932716e-07,
      "loss": 2.5964,
      "step": 3619
    },
    {
      "epoch": 0.46865391461954237,
      "grad_norm": 1.3531610229149031,
      "learning_rate": 8.001586585037945e-07,
      "loss": 2.3726,
      "step": 3620
    },
    {
      "epoch": 0.46865391461954237,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.594038486480713,
      "eval_runtime": 13.5648,
      "eval_samples_per_second": 3.244,
      "eval_steps_per_second": 0.221,
      "step": 3620
    },
    {
      "epoch": 0.4687833770268958,
      "grad_norm": 1.217856566210872,
      "learning_rate": 7.99887483934611e-07,
      "loss": 2.4648,
      "step": 3621
    },
    {
      "epoch": 0.4689128394342493,
      "grad_norm": 2.4974573885477716,
      "learning_rate": 7.99616273641114e-07,
      "loss": 2.5525,
      "step": 3622
    },
    {
      "epoch": 0.4690423018416027,
      "grad_norm": 1.5620724118722353,
      "learning_rate": 7.993450276787041e-07,
      "loss": 2.2842,
      "step": 3623
    },
    {
      "epoch": 0.46917176424895624,
      "grad_norm": 1.8211379853260594,
      "learning_rate": 7.990737461027882e-07,
      "loss": 2.6074,
      "step": 3624
    },
    {
      "epoch": 0.4693012266563097,
      "grad_norm": 1.8456022956077363,
      "learning_rate": 7.988024289687815e-07,
      "loss": 2.3943,
      "step": 3625
    },
    {
      "epoch": 0.4693012266563097,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5940607786178589,
      "eval_runtime": 14.5194,
      "eval_samples_per_second": 3.03,
      "eval_steps_per_second": 0.207,
      "step": 3625
    },
    {
      "epoch": 0.46943068906366314,
      "grad_norm": 1.135654349596474,
      "learning_rate": 7.985310763321057e-07,
      "loss": 2.2441,
      "step": 3626
    },
    {
      "epoch": 0.4695601514710166,
      "grad_norm": 1.8172491529857226,
      "learning_rate": 7.982596882481905e-07,
      "loss": 2.5605,
      "step": 3627
    },
    {
      "epoch": 0.46968961387837005,
      "grad_norm": 1.547347121707995,
      "learning_rate": 7.979882647724719e-07,
      "loss": 2.5198,
      "step": 3628
    },
    {
      "epoch": 0.46981907628572356,
      "grad_norm": 2.1268555572499714,
      "learning_rate": 7.97716805960394e-07,
      "loss": 2.4683,
      "step": 3629
    },
    {
      "epoch": 0.469948538693077,
      "grad_norm": 3.692927730484567,
      "learning_rate": 7.974453118674077e-07,
      "loss": 2.6782,
      "step": 3630
    },
    {
      "epoch": 0.469948538693077,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5940607786178589,
      "eval_runtime": 14.06,
      "eval_samples_per_second": 3.129,
      "eval_steps_per_second": 0.213,
      "step": 3630
    },
    {
      "epoch": 0.47007800110043046,
      "grad_norm": 1.486069011000907,
      "learning_rate": 7.971737825489712e-07,
      "loss": 2.4634,
      "step": 3631
    },
    {
      "epoch": 0.4702074635077839,
      "grad_norm": 4.336146760444173,
      "learning_rate": 7.969022180605496e-07,
      "loss": 2.748,
      "step": 3632
    },
    {
      "epoch": 0.47033692591513737,
      "grad_norm": 1.6069502772995092,
      "learning_rate": 7.96630618457616e-07,
      "loss": 2.4838,
      "step": 3633
    },
    {
      "epoch": 0.4704663883224909,
      "grad_norm": 1.5285935016895567,
      "learning_rate": 7.963589837956495e-07,
      "loss": 2.2715,
      "step": 3634
    },
    {
      "epoch": 0.47059585072984433,
      "grad_norm": 3.724211678255899,
      "learning_rate": 7.960873141301374e-07,
      "loss": 2.5835,
      "step": 3635
    },
    {
      "epoch": 0.47059585072984433,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.591264247894287,
      "eval_runtime": 14.1703,
      "eval_samples_per_second": 3.105,
      "eval_steps_per_second": 0.212,
      "step": 3635
    },
    {
      "epoch": 0.4707253131371978,
      "grad_norm": 1.8627103727513723,
      "learning_rate": 7.958156095165736e-07,
      "loss": 2.5374,
      "step": 3636
    },
    {
      "epoch": 0.47085477554455124,
      "grad_norm": 1.8171282842768852,
      "learning_rate": 7.955438700104592e-07,
      "loss": 2.7075,
      "step": 3637
    },
    {
      "epoch": 0.4709842379519047,
      "grad_norm": 1.973031343685012,
      "learning_rate": 7.952720956673026e-07,
      "loss": 2.4497,
      "step": 3638
    },
    {
      "epoch": 0.4711137003592582,
      "grad_norm": 4.783104249776493,
      "learning_rate": 7.950002865426192e-07,
      "loss": 3.0967,
      "step": 3639
    },
    {
      "epoch": 0.47124316276661166,
      "grad_norm": 2.9507778313227306,
      "learning_rate": 7.947284426919316e-07,
      "loss": 2.6985,
      "step": 3640
    },
    {
      "epoch": 0.47124316276661166,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5920188426971436,
      "eval_runtime": 14.0808,
      "eval_samples_per_second": 3.125,
      "eval_steps_per_second": 0.213,
      "step": 3640
    },
    {
      "epoch": 0.4713726251739651,
      "grad_norm": 2.8097235245936583,
      "learning_rate": 7.944565641707694e-07,
      "loss": 2.4431,
      "step": 3641
    },
    {
      "epoch": 0.47150208758131856,
      "grad_norm": 2.1359216751180576,
      "learning_rate": 7.941846510346691e-07,
      "loss": 2.3904,
      "step": 3642
    },
    {
      "epoch": 0.471631549988672,
      "grad_norm": 2.3195741311969353,
      "learning_rate": 7.939127033391747e-07,
      "loss": 2.5706,
      "step": 3643
    },
    {
      "epoch": 0.4717610123960255,
      "grad_norm": 1.4403763591503504,
      "learning_rate": 7.936407211398368e-07,
      "loss": 2.3937,
      "step": 3644
    },
    {
      "epoch": 0.471890474803379,
      "grad_norm": 1.6442525806178,
      "learning_rate": 7.933687044922137e-07,
      "loss": 2.4302,
      "step": 3645
    },
    {
      "epoch": 0.471890474803379,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.592729091644287,
      "eval_runtime": 13.7173,
      "eval_samples_per_second": 3.208,
      "eval_steps_per_second": 0.219,
      "step": 3645
    },
    {
      "epoch": 0.47201993721073243,
      "grad_norm": 2.425007683694663,
      "learning_rate": 7.930966534518697e-07,
      "loss": 2.3091,
      "step": 3646
    },
    {
      "epoch": 0.4721493996180859,
      "grad_norm": 2.0655610388583048,
      "learning_rate": 7.928245680743774e-07,
      "loss": 2.4888,
      "step": 3647
    },
    {
      "epoch": 0.47227886202543934,
      "grad_norm": 1.5701088927262847,
      "learning_rate": 7.925524484153151e-07,
      "loss": 2.4421,
      "step": 3648
    },
    {
      "epoch": 0.47240832443279285,
      "grad_norm": 2.3285068063372854,
      "learning_rate": 7.922802945302692e-07,
      "loss": 2.4827,
      "step": 3649
    },
    {
      "epoch": 0.4725377868401463,
      "grad_norm": 3.840222524902766,
      "learning_rate": 7.920081064748326e-07,
      "loss": 2.573,
      "step": 3650
    },
    {
      "epoch": 0.4725377868401463,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5927512645721436,
      "eval_runtime": 13.0948,
      "eval_samples_per_second": 3.36,
      "eval_steps_per_second": 0.229,
      "step": 3650
    },
    {
      "epoch": 0.47266724924749975,
      "grad_norm": 4.430643802202797,
      "learning_rate": 7.917358843046049e-07,
      "loss": 2.8379,
      "step": 3651
    },
    {
      "epoch": 0.4727967116548532,
      "grad_norm": 1.7401327511814533,
      "learning_rate": 7.914636280751936e-07,
      "loss": 2.3899,
      "step": 3652
    },
    {
      "epoch": 0.47292617406220666,
      "grad_norm": 3.4415834060918784,
      "learning_rate": 7.91191337842212e-07,
      "loss": 2.4988,
      "step": 3653
    },
    {
      "epoch": 0.47305563646956017,
      "grad_norm": 1.608745771038558,
      "learning_rate": 7.909190136612811e-07,
      "loss": 2.5613,
      "step": 3654
    },
    {
      "epoch": 0.4731850988769136,
      "grad_norm": 3.193515669381869,
      "learning_rate": 7.906466555880287e-07,
      "loss": 2.6494,
      "step": 3655
    },
    {
      "epoch": 0.4731850988769136,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5877796411514282,
      "eval_runtime": 14.9692,
      "eval_samples_per_second": 2.939,
      "eval_steps_per_second": 0.2,
      "step": 3655
    },
    {
      "epoch": 0.4733145612842671,
      "grad_norm": 1.9220860961258883,
      "learning_rate": 7.903742636780895e-07,
      "loss": 2.4949,
      "step": 3656
    },
    {
      "epoch": 0.47344402369162053,
      "grad_norm": 2.0942082472127814,
      "learning_rate": 7.901018379871047e-07,
      "loss": 2.3389,
      "step": 3657
    },
    {
      "epoch": 0.473573486098974,
      "grad_norm": 3.3252440437042887,
      "learning_rate": 7.898293785707231e-07,
      "loss": 2.8555,
      "step": 3658
    },
    {
      "epoch": 0.4737029485063275,
      "grad_norm": 1.6969295377177454,
      "learning_rate": 7.895568854846001e-07,
      "loss": 2.656,
      "step": 3659
    },
    {
      "epoch": 0.47383241091368095,
      "grad_norm": 1.1628809286153323,
      "learning_rate": 7.89284358784398e-07,
      "loss": 2.2322,
      "step": 3660
    },
    {
      "epoch": 0.47383241091368095,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5892223119735718,
      "eval_runtime": 14.5752,
      "eval_samples_per_second": 3.019,
      "eval_steps_per_second": 0.206,
      "step": 3660
    },
    {
      "epoch": 0.4739618733210344,
      "grad_norm": 1.4014953449138379,
      "learning_rate": 7.890117985257853e-07,
      "loss": 2.4456,
      "step": 3661
    },
    {
      "epoch": 0.47409133572838785,
      "grad_norm": 1.7908022122096063,
      "learning_rate": 7.887392047644385e-07,
      "loss": 2.7241,
      "step": 3662
    },
    {
      "epoch": 0.47422079813574136,
      "grad_norm": 2.45014592304961,
      "learning_rate": 7.884665775560404e-07,
      "loss": 2.3882,
      "step": 3663
    },
    {
      "epoch": 0.4743502605430948,
      "grad_norm": 1.6672833725737877,
      "learning_rate": 7.881939169562805e-07,
      "loss": 2.4238,
      "step": 3664
    },
    {
      "epoch": 0.47447972295044827,
      "grad_norm": 4.58009051603523,
      "learning_rate": 7.879212230208552e-07,
      "loss": 2.7588,
      "step": 3665
    },
    {
      "epoch": 0.47447972295044827,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5863147974014282,
      "eval_runtime": 14.3211,
      "eval_samples_per_second": 3.072,
      "eval_steps_per_second": 0.209,
      "step": 3665
    },
    {
      "epoch": 0.4746091853578017,
      "grad_norm": 1.2909944526768258,
      "learning_rate": 7.876484958054678e-07,
      "loss": 2.4045,
      "step": 3666
    },
    {
      "epoch": 0.4747386477651552,
      "grad_norm": 1.9253731831295025,
      "learning_rate": 7.873757353658286e-07,
      "loss": 2.437,
      "step": 3667
    },
    {
      "epoch": 0.4748681101725087,
      "grad_norm": 1.493828060078587,
      "learning_rate": 7.871029417576541e-07,
      "loss": 2.4282,
      "step": 3668
    },
    {
      "epoch": 0.47499757257986214,
      "grad_norm": 1.7833120546842636,
      "learning_rate": 7.868301150366682e-07,
      "loss": 2.4453,
      "step": 3669
    },
    {
      "epoch": 0.4751270349872156,
      "grad_norm": 1.5134766080861666,
      "learning_rate": 7.865572552586009e-07,
      "loss": 2.4148,
      "step": 3670
    },
    {
      "epoch": 0.4751270349872156,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5864036083221436,
      "eval_runtime": 14.7312,
      "eval_samples_per_second": 2.987,
      "eval_steps_per_second": 0.204,
      "step": 3670
    },
    {
      "epoch": 0.47525649739456904,
      "grad_norm": 1.6231341720498271,
      "learning_rate": 7.862843624791896e-07,
      "loss": 2.5667,
      "step": 3671
    },
    {
      "epoch": 0.4753859598019225,
      "grad_norm": 2.340347213473683,
      "learning_rate": 7.860114367541785e-07,
      "loss": 2.4146,
      "step": 3672
    },
    {
      "epoch": 0.475515422209276,
      "grad_norm": 1.719479769558792,
      "learning_rate": 7.857384781393177e-07,
      "loss": 2.4636,
      "step": 3673
    },
    {
      "epoch": 0.47564488461662946,
      "grad_norm": 1.2499263146873119,
      "learning_rate": 7.85465486690365e-07,
      "loss": 2.2085,
      "step": 3674
    },
    {
      "epoch": 0.4757743470239829,
      "grad_norm": 2.618130554292385,
      "learning_rate": 7.851924624630836e-07,
      "loss": 2.7778,
      "step": 3675
    },
    {
      "epoch": 0.4757743470239829,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5892223119735718,
      "eval_runtime": 13.7603,
      "eval_samples_per_second": 3.198,
      "eval_steps_per_second": 0.218,
      "step": 3675
    },
    {
      "epoch": 0.47590380943133637,
      "grad_norm": 2.6395850335243884,
      "learning_rate": 7.84919405513245e-07,
      "loss": 2.47,
      "step": 3676
    },
    {
      "epoch": 0.4760332718386898,
      "grad_norm": 3.095993944547029,
      "learning_rate": 7.846463158966267e-07,
      "loss": 2.6088,
      "step": 3677
    },
    {
      "epoch": 0.47616273424604333,
      "grad_norm": 1.3857736384241748,
      "learning_rate": 7.843731936690123e-07,
      "loss": 2.4314,
      "step": 3678
    },
    {
      "epoch": 0.4762921966533968,
      "grad_norm": 1.2359118674120453,
      "learning_rate": 7.841000388861927e-07,
      "loss": 2.3379,
      "step": 3679
    },
    {
      "epoch": 0.47642165906075024,
      "grad_norm": 1.5659393378335085,
      "learning_rate": 7.838268516039653e-07,
      "loss": 2.4626,
      "step": 3680
    },
    {
      "epoch": 0.47642165906075024,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.584916591644287,
      "eval_runtime": 14.2106,
      "eval_samples_per_second": 3.096,
      "eval_steps_per_second": 0.211,
      "step": 3680
    },
    {
      "epoch": 0.4765511214681037,
      "grad_norm": 1.6248855815364758,
      "learning_rate": 7.835536318781342e-07,
      "loss": 2.3647,
      "step": 3681
    },
    {
      "epoch": 0.47668058387545714,
      "grad_norm": 2.3366036662156886,
      "learning_rate": 7.832803797645101e-07,
      "loss": 2.623,
      "step": 3682
    },
    {
      "epoch": 0.47681004628281065,
      "grad_norm": 1.9540910015132569,
      "learning_rate": 7.830070953189105e-07,
      "loss": 2.2856,
      "step": 3683
    },
    {
      "epoch": 0.4769395086901641,
      "grad_norm": 1.7453645913102007,
      "learning_rate": 7.827337785971586e-07,
      "loss": 2.4363,
      "step": 3684
    },
    {
      "epoch": 0.47706897109751756,
      "grad_norm": 3.870806146670109,
      "learning_rate": 7.824604296550854e-07,
      "loss": 2.8291,
      "step": 3685
    },
    {
      "epoch": 0.47706897109751756,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5828746557235718,
      "eval_runtime": 14.9179,
      "eval_samples_per_second": 2.949,
      "eval_steps_per_second": 0.201,
      "step": 3685
    },
    {
      "epoch": 0.477198433504871,
      "grad_norm": 1.824831790233935,
      "learning_rate": 7.821870485485281e-07,
      "loss": 2.499,
      "step": 3686
    },
    {
      "epoch": 0.47732789591222446,
      "grad_norm": 2.0851171491837874,
      "learning_rate": 7.819136353333298e-07,
      "loss": 2.5444,
      "step": 3687
    },
    {
      "epoch": 0.477457358319578,
      "grad_norm": 1.2957353852878712,
      "learning_rate": 7.816401900653414e-07,
      "loss": 2.3779,
      "step": 3688
    },
    {
      "epoch": 0.4775868207269314,
      "grad_norm": 6.37040799261014,
      "learning_rate": 7.813667128004187e-07,
      "loss": 2.7166,
      "step": 3689
    },
    {
      "epoch": 0.4777162831342849,
      "grad_norm": 2.27888901263236,
      "learning_rate": 7.810932035944257e-07,
      "loss": 2.6482,
      "step": 3690
    },
    {
      "epoch": 0.4777162831342849,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5834294557571411,
      "eval_runtime": 14.8773,
      "eval_samples_per_second": 2.958,
      "eval_steps_per_second": 0.202,
      "step": 3690
    },
    {
      "epoch": 0.47784574554163833,
      "grad_norm": 1.7235541416427815,
      "learning_rate": 7.808196625032323e-07,
      "loss": 2.4021,
      "step": 3691
    },
    {
      "epoch": 0.4779752079489918,
      "grad_norm": 1.5463941959221092,
      "learning_rate": 7.805460895827143e-07,
      "loss": 2.6846,
      "step": 3692
    },
    {
      "epoch": 0.4781046703563453,
      "grad_norm": 1.354331992013068,
      "learning_rate": 7.802724848887545e-07,
      "loss": 2.5376,
      "step": 3693
    },
    {
      "epoch": 0.47823413276369875,
      "grad_norm": 2.122552087688215,
      "learning_rate": 7.799988484772426e-07,
      "loss": 2.3826,
      "step": 3694
    },
    {
      "epoch": 0.4783635951710522,
      "grad_norm": 2.316364150505932,
      "learning_rate": 7.797251804040742e-07,
      "loss": 2.6699,
      "step": 3695
    },
    {
      "epoch": 0.4783635951710522,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5855823755264282,
      "eval_runtime": 15.2034,
      "eval_samples_per_second": 2.894,
      "eval_steps_per_second": 0.197,
      "step": 3695
    },
    {
      "epoch": 0.47849305757840566,
      "grad_norm": 2.3547673697725853,
      "learning_rate": 7.794514807251518e-07,
      "loss": 2.4785,
      "step": 3696
    },
    {
      "epoch": 0.4786225199857591,
      "grad_norm": 2.080325531391821,
      "learning_rate": 7.791777494963837e-07,
      "loss": 2.2725,
      "step": 3697
    },
    {
      "epoch": 0.4787519823931126,
      "grad_norm": 1.9976893564121538,
      "learning_rate": 7.78903986773685e-07,
      "loss": 2.3774,
      "step": 3698
    },
    {
      "epoch": 0.47888144480046607,
      "grad_norm": 1.559551048113188,
      "learning_rate": 7.786301926129778e-07,
      "loss": 2.5544,
      "step": 3699
    },
    {
      "epoch": 0.4790109072078195,
      "grad_norm": 2.3145842669312824,
      "learning_rate": 7.783563670701897e-07,
      "loss": 2.6377,
      "step": 3700
    },
    {
      "epoch": 0.4790109072078195,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5814098119735718,
      "eval_runtime": 14.2466,
      "eval_samples_per_second": 3.088,
      "eval_steps_per_second": 0.211,
      "step": 3700
    },
    {
      "epoch": 0.479140369615173,
      "grad_norm": 1.8498585357715334,
      "learning_rate": 7.780825102012554e-07,
      "loss": 2.5061,
      "step": 3701
    },
    {
      "epoch": 0.47926983202252643,
      "grad_norm": 2.648286277629586,
      "learning_rate": 7.778086220621153e-07,
      "loss": 2.3884,
      "step": 3702
    },
    {
      "epoch": 0.47939929442987994,
      "grad_norm": 2.170406607629059,
      "learning_rate": 7.77534702708717e-07,
      "loss": 2.2874,
      "step": 3703
    },
    {
      "epoch": 0.4795287568372334,
      "grad_norm": 1.2229933943224742,
      "learning_rate": 7.772607521970139e-07,
      "loss": 2.4102,
      "step": 3704
    },
    {
      "epoch": 0.47965821924458685,
      "grad_norm": 2.0640529545408723,
      "learning_rate": 7.769867705829662e-07,
      "loss": 2.4568,
      "step": 3705
    },
    {
      "epoch": 0.47965821924458685,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.581498622894287,
      "eval_runtime": 14.3845,
      "eval_samples_per_second": 3.059,
      "eval_steps_per_second": 0.209,
      "step": 3705
    },
    {
      "epoch": 0.4797876816519403,
      "grad_norm": 2.153704351820282,
      "learning_rate": 7.767127579225396e-07,
      "loss": 2.4873,
      "step": 3706
    },
    {
      "epoch": 0.4799171440592938,
      "grad_norm": 2.1574806866970895,
      "learning_rate": 7.764387142717072e-07,
      "loss": 2.5564,
      "step": 3707
    },
    {
      "epoch": 0.48004660646664726,
      "grad_norm": 2.20367580910148,
      "learning_rate": 7.76164639686448e-07,
      "loss": 2.5771,
      "step": 3708
    },
    {
      "epoch": 0.4801760688740007,
      "grad_norm": 2.9764950702641344,
      "learning_rate": 7.758905342227471e-07,
      "loss": 2.4888,
      "step": 3709
    },
    {
      "epoch": 0.48030553128135417,
      "grad_norm": 4.206729882439295,
      "learning_rate": 7.756163979365963e-07,
      "loss": 3.125,
      "step": 3710
    },
    {
      "epoch": 0.48030553128135417,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.58349609375,
      "eval_runtime": 14.6897,
      "eval_samples_per_second": 2.995,
      "eval_steps_per_second": 0.204,
      "step": 3710
    },
    {
      "epoch": 0.4804349936887076,
      "grad_norm": 1.954075134769091,
      "learning_rate": 7.753422308839932e-07,
      "loss": 2.5083,
      "step": 3711
    },
    {
      "epoch": 0.48056445609606113,
      "grad_norm": 2.457147486965655,
      "learning_rate": 7.750680331209423e-07,
      "loss": 2.7341,
      "step": 3712
    },
    {
      "epoch": 0.4806939185034146,
      "grad_norm": 3.3155815697949573,
      "learning_rate": 7.747938047034539e-07,
      "loss": 2.4932,
      "step": 3713
    },
    {
      "epoch": 0.48082338091076804,
      "grad_norm": 1.3728561597210085,
      "learning_rate": 7.745195456875446e-07,
      "loss": 2.2705,
      "step": 3714
    },
    {
      "epoch": 0.4809528433181215,
      "grad_norm": 1.4511225081906338,
      "learning_rate": 7.742452561292375e-07,
      "loss": 2.4126,
      "step": 3715
    },
    {
      "epoch": 0.4809528433181215,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5842063426971436,
      "eval_runtime": 14.4649,
      "eval_samples_per_second": 3.042,
      "eval_steps_per_second": 0.207,
      "step": 3715
    },
    {
      "epoch": 0.48108230572547495,
      "grad_norm": 1.6307034616060005,
      "learning_rate": 7.739709360845618e-07,
      "loss": 2.5022,
      "step": 3716
    },
    {
      "epoch": 0.48121176813282845,
      "grad_norm": 1.9373055476282899,
      "learning_rate": 7.736965856095529e-07,
      "loss": 2.7395,
      "step": 3717
    },
    {
      "epoch": 0.4813412305401819,
      "grad_norm": 1.901467686955576,
      "learning_rate": 7.734222047602523e-07,
      "loss": 2.168,
      "step": 3718
    },
    {
      "epoch": 0.48147069294753536,
      "grad_norm": 2.0944980091406924,
      "learning_rate": 7.731477935927081e-07,
      "loss": 2.418,
      "step": 3719
    },
    {
      "epoch": 0.4816001553548888,
      "grad_norm": 3.293475368402424,
      "learning_rate": 7.728733521629743e-07,
      "loss": 2.5547,
      "step": 3720
    },
    {
      "epoch": 0.4816001553548888,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5814541578292847,
      "eval_runtime": 14.6659,
      "eval_samples_per_second": 3.0,
      "eval_steps_per_second": 0.205,
      "step": 3720
    },
    {
      "epoch": 0.48172961776224227,
      "grad_norm": 2.36494113794181,
      "learning_rate": 7.725988805271109e-07,
      "loss": 2.5623,
      "step": 3721
    },
    {
      "epoch": 0.4818590801695958,
      "grad_norm": 1.3858640979437618,
      "learning_rate": 7.723243787411847e-07,
      "loss": 2.4854,
      "step": 3722
    },
    {
      "epoch": 0.48198854257694923,
      "grad_norm": 1.7821239943084766,
      "learning_rate": 7.720498468612679e-07,
      "loss": 2.4141,
      "step": 3723
    },
    {
      "epoch": 0.4821180049843027,
      "grad_norm": 1.6248640942210661,
      "learning_rate": 7.717752849434395e-07,
      "loss": 2.584,
      "step": 3724
    },
    {
      "epoch": 0.48224746739165614,
      "grad_norm": 1.2430576428639852,
      "learning_rate": 7.715006930437838e-07,
      "loss": 2.3594,
      "step": 3725
    },
    {
      "epoch": 0.48224746739165614,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5827858448028564,
      "eval_runtime": 16.2434,
      "eval_samples_per_second": 2.709,
      "eval_steps_per_second": 0.185,
      "step": 3725
    },
    {
      "epoch": 0.4823769297990096,
      "grad_norm": 3.7523010693157937,
      "learning_rate": 7.712260712183922e-07,
      "loss": 2.4336,
      "step": 3726
    },
    {
      "epoch": 0.4825063922063631,
      "grad_norm": 2.7849366469205035,
      "learning_rate": 7.709514195233618e-07,
      "loss": 2.354,
      "step": 3727
    },
    {
      "epoch": 0.48263585461371655,
      "grad_norm": 1.8079914434319315,
      "learning_rate": 7.706767380147955e-07,
      "loss": 2.3965,
      "step": 3728
    },
    {
      "epoch": 0.48276531702107,
      "grad_norm": 1.5272232049006351,
      "learning_rate": 7.70402026748803e-07,
      "loss": 2.4004,
      "step": 3729
    },
    {
      "epoch": 0.48289477942842346,
      "grad_norm": 1.4507466332466517,
      "learning_rate": 7.701272857814991e-07,
      "loss": 2.4417,
      "step": 3730
    },
    {
      "epoch": 0.48289477942842346,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5869362354278564,
      "eval_runtime": 15.0861,
      "eval_samples_per_second": 2.917,
      "eval_steps_per_second": 0.199,
      "step": 3730
    },
    {
      "epoch": 0.4830242418357769,
      "grad_norm": 1.5855997926424137,
      "learning_rate": 7.698525151690056e-07,
      "loss": 2.3074,
      "step": 3731
    },
    {
      "epoch": 0.4831537042431304,
      "grad_norm": 2.276389780672671,
      "learning_rate": 7.6957771496745e-07,
      "loss": 2.4673,
      "step": 3732
    },
    {
      "epoch": 0.4832831666504839,
      "grad_norm": 2.5975593439310316,
      "learning_rate": 7.693028852329655e-07,
      "loss": 2.4802,
      "step": 3733
    },
    {
      "epoch": 0.48341262905783733,
      "grad_norm": 5.229809334928725,
      "learning_rate": 7.690280260216917e-07,
      "loss": 2.5776,
      "step": 3734
    },
    {
      "epoch": 0.4835420914651908,
      "grad_norm": 2.566943958441565,
      "learning_rate": 7.687531373897743e-07,
      "loss": 2.5281,
      "step": 3735
    },
    {
      "epoch": 0.4835420914651908,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5834072828292847,
      "eval_runtime": 14.059,
      "eval_samples_per_second": 3.13,
      "eval_steps_per_second": 0.213,
      "step": 3735
    },
    {
      "epoch": 0.48367155387254424,
      "grad_norm": 1.2468401957330484,
      "learning_rate": 7.68478219393365e-07,
      "loss": 2.2517,
      "step": 3736
    },
    {
      "epoch": 0.48380101627989774,
      "grad_norm": 1.5363808577193312,
      "learning_rate": 7.682032720886213e-07,
      "loss": 2.3936,
      "step": 3737
    },
    {
      "epoch": 0.4839304786872512,
      "grad_norm": 3.4440986799875177,
      "learning_rate": 7.679282955317063e-07,
      "loss": 2.8218,
      "step": 3738
    },
    {
      "epoch": 0.48405994109460465,
      "grad_norm": 1.5846203120545803,
      "learning_rate": 7.676532897787903e-07,
      "loss": 2.5742,
      "step": 3739
    },
    {
      "epoch": 0.4841894035019581,
      "grad_norm": 2.2258609691683886,
      "learning_rate": 7.673782548860481e-07,
      "loss": 2.3489,
      "step": 3740
    },
    {
      "epoch": 0.4841894035019581,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5833629369735718,
      "eval_runtime": 14.3313,
      "eval_samples_per_second": 3.07,
      "eval_steps_per_second": 0.209,
      "step": 3740
    },
    {
      "epoch": 0.48431886590931156,
      "grad_norm": 1.8681705513875824,
      "learning_rate": 7.671031909096619e-07,
      "loss": 2.4861,
      "step": 3741
    },
    {
      "epoch": 0.48444832831666507,
      "grad_norm": 2.754221053995706,
      "learning_rate": 7.668280979058182e-07,
      "loss": 2.8,
      "step": 3742
    },
    {
      "epoch": 0.4845777907240185,
      "grad_norm": 2.8174720182799904,
      "learning_rate": 7.665529759307108e-07,
      "loss": 2.6074,
      "step": 3743
    },
    {
      "epoch": 0.484707253131372,
      "grad_norm": 1.624618610155561,
      "learning_rate": 7.662778250405392e-07,
      "loss": 2.4326,
      "step": 3744
    },
    {
      "epoch": 0.4848367155387254,
      "grad_norm": 1.7181149937346343,
      "learning_rate": 7.660026452915081e-07,
      "loss": 2.4961,
      "step": 3745
    },
    {
      "epoch": 0.4848367155387254,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5826305150985718,
      "eval_runtime": 13.5985,
      "eval_samples_per_second": 3.236,
      "eval_steps_per_second": 0.221,
      "step": 3745
    },
    {
      "epoch": 0.4849661779460789,
      "grad_norm": 1.4223643090294513,
      "learning_rate": 7.657274367398288e-07,
      "loss": 2.3142,
      "step": 3746
    },
    {
      "epoch": 0.4850956403534324,
      "grad_norm": 1.4142145578127674,
      "learning_rate": 7.65452199441718e-07,
      "loss": 2.374,
      "step": 3747
    },
    {
      "epoch": 0.48522510276078584,
      "grad_norm": 1.387880401134795,
      "learning_rate": 7.651769334533987e-07,
      "loss": 2.5703,
      "step": 3748
    },
    {
      "epoch": 0.4853545651681393,
      "grad_norm": 1.2781165790519644,
      "learning_rate": 7.649016388310994e-07,
      "loss": 2.3257,
      "step": 3749
    },
    {
      "epoch": 0.48548402757549275,
      "grad_norm": 2.682387195459793,
      "learning_rate": 7.646263156310549e-07,
      "loss": 2.9258,
      "step": 3750
    },
    {
      "epoch": 0.48548402757549275,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5848056077957153,
      "eval_runtime": 13.5334,
      "eval_samples_per_second": 3.251,
      "eval_steps_per_second": 0.222,
      "step": 3750
    },
    {
      "epoch": 0.48561348998284626,
      "grad_norm": 2.115571873693488,
      "learning_rate": 7.643509639095053e-07,
      "loss": 2.3838,
      "step": 3751
    },
    {
      "epoch": 0.4857429523901997,
      "grad_norm": 1.645931359818129,
      "learning_rate": 7.640755837226965e-07,
      "loss": 2.5327,
      "step": 3752
    },
    {
      "epoch": 0.48587241479755316,
      "grad_norm": 1.3347804292284098,
      "learning_rate": 7.638001751268812e-07,
      "loss": 2.3938,
      "step": 3753
    },
    {
      "epoch": 0.4860018772049066,
      "grad_norm": 2.9185762888402813,
      "learning_rate": 7.635247381783165e-07,
      "loss": 2.8037,
      "step": 3754
    },
    {
      "epoch": 0.48613133961226007,
      "grad_norm": 1.715877741532721,
      "learning_rate": 7.632492729332664e-07,
      "loss": 2.5388,
      "step": 3755
    },
    {
      "epoch": 0.48613133961226007,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5834739208221436,
      "eval_runtime": 14.5946,
      "eval_samples_per_second": 3.015,
      "eval_steps_per_second": 0.206,
      "step": 3755
    },
    {
      "epoch": 0.4862608020196136,
      "grad_norm": 2.406502812620039,
      "learning_rate": 7.629737794480001e-07,
      "loss": 2.5679,
      "step": 3756
    },
    {
      "epoch": 0.48639026442696703,
      "grad_norm": 2.749982137582843,
      "learning_rate": 7.626982577787925e-07,
      "loss": 2.4883,
      "step": 3757
    },
    {
      "epoch": 0.4865197268343205,
      "grad_norm": 2.599471880256545,
      "learning_rate": 7.624227079819251e-07,
      "loss": 2.4792,
      "step": 3758
    },
    {
      "epoch": 0.48664918924167394,
      "grad_norm": 2.2316005291715233,
      "learning_rate": 7.62147130113684e-07,
      "loss": 2.2654,
      "step": 3759
    },
    {
      "epoch": 0.4867786516490274,
      "grad_norm": 1.61230855683663,
      "learning_rate": 7.618715242303616e-07,
      "loss": 2.479,
      "step": 3760
    },
    {
      "epoch": 0.4867786516490274,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5819424390792847,
      "eval_runtime": 14.186,
      "eval_samples_per_second": 3.102,
      "eval_steps_per_second": 0.211,
      "step": 3760
    },
    {
      "epoch": 0.4869081140563809,
      "grad_norm": 1.5997071523205781,
      "learning_rate": 7.615958903882562e-07,
      "loss": 2.6201,
      "step": 3761
    },
    {
      "epoch": 0.48703757646373436,
      "grad_norm": 1.263696374332714,
      "learning_rate": 7.613202286436716e-07,
      "loss": 2.4778,
      "step": 3762
    },
    {
      "epoch": 0.4871670388710878,
      "grad_norm": 2.0725752991641766,
      "learning_rate": 7.610445390529172e-07,
      "loss": 2.3997,
      "step": 3763
    },
    {
      "epoch": 0.48729650127844126,
      "grad_norm": 2.061199888559606,
      "learning_rate": 7.607688216723083e-07,
      "loss": 2.511,
      "step": 3764
    },
    {
      "epoch": 0.4874259636857947,
      "grad_norm": 2.1263173244821743,
      "learning_rate": 7.604930765581655e-07,
      "loss": 2.6099,
      "step": 3765
    },
    {
      "epoch": 0.4874259636857947,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5812321901321411,
      "eval_runtime": 14.4445,
      "eval_samples_per_second": 3.046,
      "eval_steps_per_second": 0.208,
      "step": 3765
    },
    {
      "epoch": 0.4875554260931482,
      "grad_norm": 2.603026123162556,
      "learning_rate": 7.602173037668155e-07,
      "loss": 2.6643,
      "step": 3766
    },
    {
      "epoch": 0.4876848885005017,
      "grad_norm": 1.821807698417415,
      "learning_rate": 7.599415033545906e-07,
      "loss": 2.2778,
      "step": 3767
    },
    {
      "epoch": 0.48781435090785513,
      "grad_norm": 2.176638610369739,
      "learning_rate": 7.596656753778284e-07,
      "loss": 2.3701,
      "step": 3768
    },
    {
      "epoch": 0.4879438133152086,
      "grad_norm": 2.748369441027363,
      "learning_rate": 7.593898198928725e-07,
      "loss": 2.3882,
      "step": 3769
    },
    {
      "epoch": 0.48807327572256204,
      "grad_norm": 2.673306715070197,
      "learning_rate": 7.591139369560717e-07,
      "loss": 2.5186,
      "step": 3770
    },
    {
      "epoch": 0.48807327572256204,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5784579515457153,
      "eval_runtime": 14.2177,
      "eval_samples_per_second": 3.095,
      "eval_steps_per_second": 0.211,
      "step": 3770
    },
    {
      "epoch": 0.48820273812991555,
      "grad_norm": 1.9444332531518242,
      "learning_rate": 7.588380266237811e-07,
      "loss": 2.4504,
      "step": 3771
    },
    {
      "epoch": 0.488332200537269,
      "grad_norm": 2.10486202104313,
      "learning_rate": 7.585620889523609e-07,
      "loss": 2.5176,
      "step": 3772
    },
    {
      "epoch": 0.48846166294462245,
      "grad_norm": 3.366771178935237,
      "learning_rate": 7.582861239981767e-07,
      "loss": 3.1654,
      "step": 3773
    },
    {
      "epoch": 0.4885911253519759,
      "grad_norm": 2.1460931983728173,
      "learning_rate": 7.580101318176001e-07,
      "loss": 2.4222,
      "step": 3774
    },
    {
      "epoch": 0.48872058775932936,
      "grad_norm": 1.9874755259185424,
      "learning_rate": 7.577341124670082e-07,
      "loss": 2.5044,
      "step": 3775
    },
    {
      "epoch": 0.48872058775932936,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.58056640625,
      "eval_runtime": 14.8579,
      "eval_samples_per_second": 2.961,
      "eval_steps_per_second": 0.202,
      "step": 3775
    },
    {
      "epoch": 0.48885005016668287,
      "grad_norm": 1.417694617421724,
      "learning_rate": 7.574580660027833e-07,
      "loss": 2.5955,
      "step": 3776
    },
    {
      "epoch": 0.4889795125740363,
      "grad_norm": 1.5890110186592572,
      "learning_rate": 7.571819924813138e-07,
      "loss": 2.4797,
      "step": 3777
    },
    {
      "epoch": 0.4891089749813898,
      "grad_norm": 1.7896499862697623,
      "learning_rate": 7.56905891958993e-07,
      "loss": 2.6299,
      "step": 3778
    },
    {
      "epoch": 0.48923843738874323,
      "grad_norm": 1.4127749632758106,
      "learning_rate": 7.566297644922202e-07,
      "loss": 2.3528,
      "step": 3779
    },
    {
      "epoch": 0.4893678997960967,
      "grad_norm": 1.506628794055902,
      "learning_rate": 7.563536101373999e-07,
      "loss": 2.3928,
      "step": 3780
    },
    {
      "epoch": 0.4893678997960967,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5820090770721436,
      "eval_runtime": 15.0042,
      "eval_samples_per_second": 2.933,
      "eval_steps_per_second": 0.2,
      "step": 3780
    },
    {
      "epoch": 0.4894973622034502,
      "grad_norm": 1.835602741807069,
      "learning_rate": 7.560774289509423e-07,
      "loss": 2.4072,
      "step": 3781
    },
    {
      "epoch": 0.48962682461080365,
      "grad_norm": 1.872658240403593,
      "learning_rate": 7.55801220989263e-07,
      "loss": 2.5349,
      "step": 3782
    },
    {
      "epoch": 0.4897562870181571,
      "grad_norm": 2.059157933072759,
      "learning_rate": 7.555249863087833e-07,
      "loss": 2.9045,
      "step": 3783
    },
    {
      "epoch": 0.48988574942551055,
      "grad_norm": 4.654972077338492,
      "learning_rate": 7.552487249659293e-07,
      "loss": 2.7256,
      "step": 3784
    },
    {
      "epoch": 0.490015211832864,
      "grad_norm": 4.282968235896998,
      "learning_rate": 7.549724370171332e-07,
      "loss": 3.1448,
      "step": 3785
    },
    {
      "epoch": 0.490015211832864,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.582075595855713,
      "eval_runtime": 15.2266,
      "eval_samples_per_second": 2.89,
      "eval_steps_per_second": 0.197,
      "step": 3785
    },
    {
      "epoch": 0.4901446742402175,
      "grad_norm": 1.7328692945690685,
      "learning_rate": 7.546961225188326e-07,
      "loss": 2.5132,
      "step": 3786
    },
    {
      "epoch": 0.49027413664757097,
      "grad_norm": 1.9279686658187183,
      "learning_rate": 7.544197815274701e-07,
      "loss": 2.3301,
      "step": 3787
    },
    {
      "epoch": 0.4904035990549244,
      "grad_norm": 1.8035244152791998,
      "learning_rate": 7.541434140994939e-07,
      "loss": 2.4041,
      "step": 3788
    },
    {
      "epoch": 0.4905330614622779,
      "grad_norm": 1.4612428600763918,
      "learning_rate": 7.538670202913578e-07,
      "loss": 2.2925,
      "step": 3789
    },
    {
      "epoch": 0.49066252386963133,
      "grad_norm": 2.40106368432215,
      "learning_rate": 7.535906001595208e-07,
      "loss": 2.3408,
      "step": 3790
    },
    {
      "epoch": 0.49066252386963133,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5819646120071411,
      "eval_runtime": 13.9747,
      "eval_samples_per_second": 3.149,
      "eval_steps_per_second": 0.215,
      "step": 3790
    },
    {
      "epoch": 0.49079198627698484,
      "grad_norm": 1.4759548614521,
      "learning_rate": 7.533141537604474e-07,
      "loss": 2.4529,
      "step": 3791
    },
    {
      "epoch": 0.4909214486843383,
      "grad_norm": 1.806778938780607,
      "learning_rate": 7.530376811506073e-07,
      "loss": 2.3044,
      "step": 3792
    },
    {
      "epoch": 0.49105091109169174,
      "grad_norm": 2.4843736152069993,
      "learning_rate": 7.527611823864753e-07,
      "loss": 2.6257,
      "step": 3793
    },
    {
      "epoch": 0.4911803734990452,
      "grad_norm": 1.3744583028057233,
      "learning_rate": 7.524846575245326e-07,
      "loss": 2.2251,
      "step": 3794
    },
    {
      "epoch": 0.4913098359063987,
      "grad_norm": 2.4159379021213367,
      "learning_rate": 7.522081066212644e-07,
      "loss": 2.8132,
      "step": 3795
    },
    {
      "epoch": 0.4913098359063987,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5798561573028564,
      "eval_runtime": 15.1108,
      "eval_samples_per_second": 2.912,
      "eval_steps_per_second": 0.199,
      "step": 3795
    },
    {
      "epoch": 0.49143929831375216,
      "grad_norm": 1.9351467813129752,
      "learning_rate": 7.519315297331619e-07,
      "loss": 2.4399,
      "step": 3796
    },
    {
      "epoch": 0.4915687607211056,
      "grad_norm": 1.7565517773032409,
      "learning_rate": 7.51654926916722e-07,
      "loss": 2.4822,
      "step": 3797
    },
    {
      "epoch": 0.49169822312845907,
      "grad_norm": 1.8909031824520097,
      "learning_rate": 7.513782982284458e-07,
      "loss": 2.5505,
      "step": 3798
    },
    {
      "epoch": 0.4918276855358125,
      "grad_norm": 2.0266424653585844,
      "learning_rate": 7.511016437248407e-07,
      "loss": 2.2991,
      "step": 3799
    },
    {
      "epoch": 0.49195714794316603,
      "grad_norm": 1.9321306233451563,
      "learning_rate": 7.508249634624191e-07,
      "loss": 2.5488,
      "step": 3800
    },
    {
      "epoch": 0.49195714794316603,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5777698755264282,
      "eval_runtime": 15.4091,
      "eval_samples_per_second": 2.855,
      "eval_steps_per_second": 0.195,
      "step": 3800
    },
    {
      "epoch": 0.4920866103505195,
      "grad_norm": 2.249859434112645,
      "learning_rate": 7.505482574976981e-07,
      "loss": 2.4429,
      "step": 3801
    },
    {
      "epoch": 0.49221607275787294,
      "grad_norm": 1.9578910076494158,
      "learning_rate": 7.502715258872009e-07,
      "loss": 2.5437,
      "step": 3802
    },
    {
      "epoch": 0.4923455351652264,
      "grad_norm": 1.6559026230322365,
      "learning_rate": 7.499947686874553e-07,
      "loss": 2.4465,
      "step": 3803
    },
    {
      "epoch": 0.49247499757257984,
      "grad_norm": 1.971016537092765,
      "learning_rate": 7.497179859549948e-07,
      "loss": 2.4014,
      "step": 3804
    },
    {
      "epoch": 0.49260445997993335,
      "grad_norm": 2.1029887299355785,
      "learning_rate": 7.494411777463578e-07,
      "loss": 2.5227,
      "step": 3805
    },
    {
      "epoch": 0.49260445997993335,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5769264698028564,
      "eval_runtime": 14.0806,
      "eval_samples_per_second": 3.125,
      "eval_steps_per_second": 0.213,
      "step": 3805
    },
    {
      "epoch": 0.4927339223872868,
      "grad_norm": 1.6023708569528559,
      "learning_rate": 7.491643441180876e-07,
      "loss": 2.5186,
      "step": 3806
    },
    {
      "epoch": 0.49286338479464026,
      "grad_norm": 1.8521373085127388,
      "learning_rate": 7.488874851267336e-07,
      "loss": 2.781,
      "step": 3807
    },
    {
      "epoch": 0.4929928472019937,
      "grad_norm": 1.7576352957359789,
      "learning_rate": 7.486106008288499e-07,
      "loss": 2.4624,
      "step": 3808
    },
    {
      "epoch": 0.49312230960934716,
      "grad_norm": 1.819612450542925,
      "learning_rate": 7.483336912809952e-07,
      "loss": 2.3779,
      "step": 3809
    },
    {
      "epoch": 0.4932517720167007,
      "grad_norm": 1.5098069548233064,
      "learning_rate": 7.480567565397344e-07,
      "loss": 2.5537,
      "step": 3810
    },
    {
      "epoch": 0.4932517720167007,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5764381885528564,
      "eval_runtime": 14.0195,
      "eval_samples_per_second": 3.138,
      "eval_steps_per_second": 0.214,
      "step": 3810
    },
    {
      "epoch": 0.4933812344240541,
      "grad_norm": 1.419732364950187,
      "learning_rate": 7.477797966616366e-07,
      "loss": 2.4734,
      "step": 3811
    },
    {
      "epoch": 0.4935106968314076,
      "grad_norm": 1.3622137753538932,
      "learning_rate": 7.475028117032769e-07,
      "loss": 2.4253,
      "step": 3812
    },
    {
      "epoch": 0.49364015923876103,
      "grad_norm": 2.1862426362924943,
      "learning_rate": 7.472258017212348e-07,
      "loss": 2.4448,
      "step": 3813
    },
    {
      "epoch": 0.4937696216461145,
      "grad_norm": 1.903007143071898,
      "learning_rate": 7.469487667720954e-07,
      "loss": 2.6011,
      "step": 3814
    },
    {
      "epoch": 0.493899084053468,
      "grad_norm": 1.379961011257503,
      "learning_rate": 7.466717069124485e-07,
      "loss": 2.4663,
      "step": 3815
    },
    {
      "epoch": 0.493899084053468,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5742853879928589,
      "eval_runtime": 14.1682,
      "eval_samples_per_second": 3.106,
      "eval_steps_per_second": 0.212,
      "step": 3815
    },
    {
      "epoch": 0.49402854646082145,
      "grad_norm": 4.395612935373734,
      "learning_rate": 7.463946221988894e-07,
      "loss": 2.6821,
      "step": 3816
    },
    {
      "epoch": 0.4941580088681749,
      "grad_norm": 1.9107067356600425,
      "learning_rate": 7.46117512688018e-07,
      "loss": 2.3472,
      "step": 3817
    },
    {
      "epoch": 0.49428747127552836,
      "grad_norm": 1.5604783790335544,
      "learning_rate": 7.458403784364397e-07,
      "loss": 2.5479,
      "step": 3818
    },
    {
      "epoch": 0.4944169336828818,
      "grad_norm": 1.9559556506019429,
      "learning_rate": 7.45563219500765e-07,
      "loss": 2.4883,
      "step": 3819
    },
    {
      "epoch": 0.4945463960902353,
      "grad_norm": 3.671482138816925,
      "learning_rate": 7.452860359376087e-07,
      "loss": 2.6936,
      "step": 3820
    },
    {
      "epoch": 0.4945463960902353,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5728648900985718,
      "eval_runtime": 14.8591,
      "eval_samples_per_second": 2.961,
      "eval_steps_per_second": 0.202,
      "step": 3820
    },
    {
      "epoch": 0.49467585849758877,
      "grad_norm": 1.394284610016648,
      "learning_rate": 7.450088278035916e-07,
      "loss": 2.5481,
      "step": 3821
    },
    {
      "epoch": 0.4948053209049422,
      "grad_norm": 1.7266522832586098,
      "learning_rate": 7.447315951553391e-07,
      "loss": 2.3232,
      "step": 3822
    },
    {
      "epoch": 0.4949347833122957,
      "grad_norm": 2.038626792647284,
      "learning_rate": 7.444543380494813e-07,
      "loss": 2.8459,
      "step": 3823
    },
    {
      "epoch": 0.49506424571964913,
      "grad_norm": 2.8266099786324745,
      "learning_rate": 7.441770565426537e-07,
      "loss": 2.9866,
      "step": 3824
    },
    {
      "epoch": 0.49519370812700264,
      "grad_norm": 2.0805122319699954,
      "learning_rate": 7.438997506914968e-07,
      "loss": 2.5251,
      "step": 3825
    },
    {
      "epoch": 0.49519370812700264,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5750178098678589,
      "eval_runtime": 13.5697,
      "eval_samples_per_second": 3.243,
      "eval_steps_per_second": 0.221,
      "step": 3825
    },
    {
      "epoch": 0.4953231705343561,
      "grad_norm": 2.0092854279822165,
      "learning_rate": 7.436224205526557e-07,
      "loss": 2.5103,
      "step": 3826
    },
    {
      "epoch": 0.49545263294170955,
      "grad_norm": 2.6995394306328153,
      "learning_rate": 7.433450661827811e-07,
      "loss": 3.4343,
      "step": 3827
    },
    {
      "epoch": 0.495582095349063,
      "grad_norm": 2.0084445623864977,
      "learning_rate": 7.430676876385278e-07,
      "loss": 2.4236,
      "step": 3828
    },
    {
      "epoch": 0.49571155775641645,
      "grad_norm": 1.8805177996097957,
      "learning_rate": 7.42790284976556e-07,
      "loss": 2.429,
      "step": 3829
    },
    {
      "epoch": 0.49584102016376996,
      "grad_norm": 1.8158045047655909,
      "learning_rate": 7.425128582535313e-07,
      "loss": 2.6643,
      "step": 3830
    },
    {
      "epoch": 0.49584102016376996,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5728648900985718,
      "eval_runtime": 13.3939,
      "eval_samples_per_second": 3.285,
      "eval_steps_per_second": 0.224,
      "step": 3830
    },
    {
      "epoch": 0.4959704825711234,
      "grad_norm": 1.811621330528476,
      "learning_rate": 7.422354075261232e-07,
      "loss": 2.2358,
      "step": 3831
    },
    {
      "epoch": 0.49609994497847687,
      "grad_norm": 2.313425687473408,
      "learning_rate": 7.419579328510069e-07,
      "loss": 2.4932,
      "step": 3832
    },
    {
      "epoch": 0.4962294073858303,
      "grad_norm": 2.6069791484869964,
      "learning_rate": 7.416804342848623e-07,
      "loss": 2.7866,
      "step": 3833
    },
    {
      "epoch": 0.4963588697931838,
      "grad_norm": 1.5991832552492549,
      "learning_rate": 7.414029118843738e-07,
      "loss": 2.4744,
      "step": 3834
    },
    {
      "epoch": 0.4964883322005373,
      "grad_norm": 1.8881290900592334,
      "learning_rate": 7.411253657062313e-07,
      "loss": 2.345,
      "step": 3835
    },
    {
      "epoch": 0.4964883322005373,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5743075609207153,
      "eval_runtime": 14.4064,
      "eval_samples_per_second": 3.054,
      "eval_steps_per_second": 0.208,
      "step": 3835
    },
    {
      "epoch": 0.49661779460789074,
      "grad_norm": 3.204383757763302,
      "learning_rate": 7.408477958071291e-07,
      "loss": 2.4419,
      "step": 3836
    },
    {
      "epoch": 0.4967472570152442,
      "grad_norm": 1.3868120509253685,
      "learning_rate": 7.405702022437666e-07,
      "loss": 2.4214,
      "step": 3837
    },
    {
      "epoch": 0.49687671942259765,
      "grad_norm": 1.3155182370097174,
      "learning_rate": 7.402925850728475e-07,
      "loss": 2.4441,
      "step": 3838
    },
    {
      "epoch": 0.49700618182995115,
      "grad_norm": 1.8934473387512372,
      "learning_rate": 7.400149443510814e-07,
      "loss": 2.4619,
      "step": 3839
    },
    {
      "epoch": 0.4971356442373046,
      "grad_norm": 1.3701282877287062,
      "learning_rate": 7.397372801351814e-07,
      "loss": 2.3213,
      "step": 3840
    },
    {
      "epoch": 0.4971356442373046,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5686701536178589,
      "eval_runtime": 14.2887,
      "eval_samples_per_second": 3.079,
      "eval_steps_per_second": 0.21,
      "step": 3840
    },
    {
      "epoch": 0.49726510664465806,
      "grad_norm": 2.381051717378634,
      "learning_rate": 7.394595924818665e-07,
      "loss": 2.6194,
      "step": 3841
    },
    {
      "epoch": 0.4973945690520115,
      "grad_norm": 3.912663294756281,
      "learning_rate": 7.391818814478595e-07,
      "loss": 2.627,
      "step": 3842
    },
    {
      "epoch": 0.49752403145936497,
      "grad_norm": 1.8801207395755364,
      "learning_rate": 7.38904147089889e-07,
      "loss": 2.5083,
      "step": 3843
    },
    {
      "epoch": 0.4976534938667185,
      "grad_norm": 1.8499822206012733,
      "learning_rate": 7.38626389464688e-07,
      "loss": 2.484,
      "step": 3844
    },
    {
      "epoch": 0.49778295627407193,
      "grad_norm": 1.9300649295209942,
      "learning_rate": 7.383486086289936e-07,
      "loss": 2.6553,
      "step": 3845
    },
    {
      "epoch": 0.49778295627407193,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5678932666778564,
      "eval_runtime": 14.9147,
      "eval_samples_per_second": 2.95,
      "eval_steps_per_second": 0.201,
      "step": 3845
    },
    {
      "epoch": 0.4979124186814254,
      "grad_norm": 1.4945354598537437,
      "learning_rate": 7.380708046395485e-07,
      "loss": 2.3804,
      "step": 3846
    },
    {
      "epoch": 0.49804188108877884,
      "grad_norm": 1.287166983856708,
      "learning_rate": 7.377929775530997e-07,
      "loss": 2.3335,
      "step": 3847
    },
    {
      "epoch": 0.4981713434961323,
      "grad_norm": 2.041089072972178,
      "learning_rate": 7.375151274263989e-07,
      "loss": 2.4495,
      "step": 3848
    },
    {
      "epoch": 0.4983008059034858,
      "grad_norm": 1.734460772955245,
      "learning_rate": 7.37237254316203e-07,
      "loss": 2.5708,
      "step": 3849
    },
    {
      "epoch": 0.49843026831083925,
      "grad_norm": 2.155783042283747,
      "learning_rate": 7.369593582792727e-07,
      "loss": 2.552,
      "step": 3850
    },
    {
      "epoch": 0.49843026831083925,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5707341432571411,
      "eval_runtime": 13.9557,
      "eval_samples_per_second": 3.153,
      "eval_steps_per_second": 0.215,
      "step": 3850
    },
    {
      "epoch": 0.4985597307181927,
      "grad_norm": 1.8270992707659444,
      "learning_rate": 7.366814393723742e-07,
      "loss": 2.5442,
      "step": 3851
    },
    {
      "epoch": 0.49868919312554616,
      "grad_norm": 1.526756141936261,
      "learning_rate": 7.364034976522783e-07,
      "loss": 2.5999,
      "step": 3852
    },
    {
      "epoch": 0.4988186555328996,
      "grad_norm": 2.1889065723581616,
      "learning_rate": 7.361255331757597e-07,
      "loss": 2.689,
      "step": 3853
    },
    {
      "epoch": 0.4989481179402531,
      "grad_norm": 1.4077063254995932,
      "learning_rate": 7.358475459995985e-07,
      "loss": 2.5684,
      "step": 3854
    },
    {
      "epoch": 0.4990775803476066,
      "grad_norm": 2.28805826148618,
      "learning_rate": 7.355695361805794e-07,
      "loss": 2.3989,
      "step": 3855
    },
    {
      "epoch": 0.4990775803476066,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5650302171707153,
      "eval_runtime": 14.6487,
      "eval_samples_per_second": 3.004,
      "eval_steps_per_second": 0.205,
      "step": 3855
    },
    {
      "epoch": 0.49920704275496003,
      "grad_norm": 1.5251554579287894,
      "learning_rate": 7.352915037754912e-07,
      "loss": 2.5068,
      "step": 3856
    },
    {
      "epoch": 0.4993365051623135,
      "grad_norm": 2.1011453941041363,
      "learning_rate": 7.35013448841128e-07,
      "loss": 2.8315,
      "step": 3857
    },
    {
      "epoch": 0.49946596756966694,
      "grad_norm": 1.1925082996351501,
      "learning_rate": 7.347353714342881e-07,
      "loss": 2.272,
      "step": 3858
    },
    {
      "epoch": 0.49959542997702044,
      "grad_norm": 1.7177547031808236,
      "learning_rate": 7.344572716117742e-07,
      "loss": 2.6719,
      "step": 3859
    },
    {
      "epoch": 0.4997248923843739,
      "grad_norm": 2.021125481308124,
      "learning_rate": 7.34179149430394e-07,
      "loss": 2.4043,
      "step": 3860
    },
    {
      "epoch": 0.4997248923843739,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5636540651321411,
      "eval_runtime": 14.5011,
      "eval_samples_per_second": 3.034,
      "eval_steps_per_second": 0.207,
      "step": 3860
    },
    {
      "epoch": 0.49985435479172735,
      "grad_norm": 1.3750962233632043,
      "learning_rate": 7.339010049469596e-07,
      "loss": 2.2122,
      "step": 3861
    },
    {
      "epoch": 0.4999838171990808,
      "grad_norm": 1.5203316871122106,
      "learning_rate": 7.336228382182875e-07,
      "loss": 2.4861,
      "step": 3862
    },
    {
      "epoch": 0.5001132796064343,
      "grad_norm": 2.0263422902254677,
      "learning_rate": 7.333446493011991e-07,
      "loss": 2.418,
      "step": 3863
    },
    {
      "epoch": 0.5002427420137877,
      "grad_norm": 2.407502207446351,
      "learning_rate": 7.330664382525201e-07,
      "loss": 2.5923,
      "step": 3864
    },
    {
      "epoch": 0.5003722044211412,
      "grad_norm": 2.111981918023866,
      "learning_rate": 7.327882051290805e-07,
      "loss": 2.9321,
      "step": 3865
    },
    {
      "epoch": 0.5003722044211412,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5658291578292847,
      "eval_runtime": 13.6826,
      "eval_samples_per_second": 3.216,
      "eval_steps_per_second": 0.219,
      "step": 3865
    },
    {
      "epoch": 0.5005016668284947,
      "grad_norm": 2.0362133554203052,
      "learning_rate": 7.325099499877154e-07,
      "loss": 2.4763,
      "step": 3866
    },
    {
      "epoch": 0.5006311292358482,
      "grad_norm": 1.3532069968107463,
      "learning_rate": 7.32231672885264e-07,
      "loss": 2.4005,
      "step": 3867
    },
    {
      "epoch": 0.5007605916432016,
      "grad_norm": 1.8315827762511798,
      "learning_rate": 7.319533738785696e-07,
      "loss": 2.5701,
      "step": 3868
    },
    {
      "epoch": 0.5008900540505551,
      "grad_norm": 1.9170065064572093,
      "learning_rate": 7.316750530244809e-07,
      "loss": 2.4468,
      "step": 3869
    },
    {
      "epoch": 0.5010195164579085,
      "grad_norm": 2.0260918662704515,
      "learning_rate": 7.313967103798504e-07,
      "loss": 2.6094,
      "step": 3870
    },
    {
      "epoch": 0.5010195164579085,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5665394067764282,
      "eval_runtime": 13.5102,
      "eval_samples_per_second": 3.257,
      "eval_steps_per_second": 0.222,
      "step": 3870
    },
    {
      "epoch": 0.501148978865262,
      "grad_norm": 1.5321029807546402,
      "learning_rate": 7.311183460015353e-07,
      "loss": 2.4431,
      "step": 3871
    },
    {
      "epoch": 0.5012784412726154,
      "grad_norm": 1.7593700539247659,
      "learning_rate": 7.308399599463973e-07,
      "loss": 2.3701,
      "step": 3872
    },
    {
      "epoch": 0.5014079036799689,
      "grad_norm": 1.882865198824058,
      "learning_rate": 7.305615522713018e-07,
      "loss": 2.4648,
      "step": 3873
    },
    {
      "epoch": 0.5015373660873224,
      "grad_norm": 1.4862211139501802,
      "learning_rate": 7.302831230331197e-07,
      "loss": 2.4026,
      "step": 3874
    },
    {
      "epoch": 0.5016668284946758,
      "grad_norm": 2.947035413963958,
      "learning_rate": 7.300046722887259e-07,
      "loss": 2.6794,
      "step": 3875
    },
    {
      "epoch": 0.5016668284946758,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5650745630264282,
      "eval_runtime": 14.482,
      "eval_samples_per_second": 3.038,
      "eval_steps_per_second": 0.207,
      "step": 3875
    },
    {
      "epoch": 0.5017962909020294,
      "grad_norm": 2.41367090513958,
      "learning_rate": 7.297262000949993e-07,
      "loss": 2.4866,
      "step": 3876
    },
    {
      "epoch": 0.5019257533093828,
      "grad_norm": 1.9178427941864862,
      "learning_rate": 7.294477065088237e-07,
      "loss": 2.5181,
      "step": 3877
    },
    {
      "epoch": 0.5020552157167363,
      "grad_norm": 2.326835907076696,
      "learning_rate": 7.291691915870869e-07,
      "loss": 2.3979,
      "step": 3878
    },
    {
      "epoch": 0.5021846781240897,
      "grad_norm": 3.4223439682426218,
      "learning_rate": 7.288906553866815e-07,
      "loss": 2.4858,
      "step": 3879
    },
    {
      "epoch": 0.5023141405314432,
      "grad_norm": 2.908525290796561,
      "learning_rate": 7.286120979645038e-07,
      "loss": 2.4551,
      "step": 3880
    },
    {
      "epoch": 0.5023141405314432,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.56005859375,
      "eval_runtime": 14.2513,
      "eval_samples_per_second": 3.087,
      "eval_steps_per_second": 0.211,
      "step": 3880
    },
    {
      "epoch": 0.5024436029387966,
      "grad_norm": 3.0164933433794934,
      "learning_rate": 7.28333519377455e-07,
      "loss": 2.4829,
      "step": 3881
    },
    {
      "epoch": 0.5025730653461501,
      "grad_norm": 2.490369015888935,
      "learning_rate": 7.280549196824405e-07,
      "loss": 2.5938,
      "step": 3882
    },
    {
      "epoch": 0.5027025277535035,
      "grad_norm": 1.8304420042134857,
      "learning_rate": 7.277762989363696e-07,
      "loss": 2.417,
      "step": 3883
    },
    {
      "epoch": 0.502831990160857,
      "grad_norm": 4.388618020094771,
      "learning_rate": 7.274976571961566e-07,
      "loss": 2.3291,
      "step": 3884
    },
    {
      "epoch": 0.5029614525682105,
      "grad_norm": 4.2393510992134456,
      "learning_rate": 7.272189945187196e-07,
      "loss": 2.6987,
      "step": 3885
    },
    {
      "epoch": 0.5029614525682105,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5587490797042847,
      "eval_runtime": 13.4433,
      "eval_samples_per_second": 3.273,
      "eval_steps_per_second": 0.223,
      "step": 3885
    },
    {
      "epoch": 0.503090914975564,
      "grad_norm": 1.3787967172423385,
      "learning_rate": 7.269403109609811e-07,
      "loss": 2.3911,
      "step": 3886
    },
    {
      "epoch": 0.5032203773829175,
      "grad_norm": 2.333546024328004,
      "learning_rate": 7.266616065798679e-07,
      "loss": 2.7061,
      "step": 3887
    },
    {
      "epoch": 0.5033498397902709,
      "grad_norm": 3.205586913066561,
      "learning_rate": 7.263828814323107e-07,
      "loss": 2.228,
      "step": 3888
    },
    {
      "epoch": 0.5034793021976244,
      "grad_norm": 2.4211409160020594,
      "learning_rate": 7.261041355752453e-07,
      "loss": 2.8258,
      "step": 3889
    },
    {
      "epoch": 0.5036087646049778,
      "grad_norm": 3.4646879677254367,
      "learning_rate": 7.258253690656109e-07,
      "loss": 3.0598,
      "step": 3890
    },
    {
      "epoch": 0.5036087646049778,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5587269067764282,
      "eval_runtime": 13.2947,
      "eval_samples_per_second": 3.31,
      "eval_steps_per_second": 0.226,
      "step": 3890
    },
    {
      "epoch": 0.5037382270123313,
      "grad_norm": 2.6102513763882227,
      "learning_rate": 7.255465819603512e-07,
      "loss": 2.6096,
      "step": 3891
    },
    {
      "epoch": 0.5038676894196847,
      "grad_norm": 1.4244695399444645,
      "learning_rate": 7.25267774316414e-07,
      "loss": 2.51,
      "step": 3892
    },
    {
      "epoch": 0.5039971518270382,
      "grad_norm": 1.7718620456748764,
      "learning_rate": 7.249889461907517e-07,
      "loss": 2.3584,
      "step": 3893
    },
    {
      "epoch": 0.5041266142343916,
      "grad_norm": 1.8930299446891008,
      "learning_rate": 7.247100976403204e-07,
      "loss": 2.6296,
      "step": 3894
    },
    {
      "epoch": 0.5042560766417451,
      "grad_norm": 1.7490787792980549,
      "learning_rate": 7.244312287220806e-07,
      "loss": 2.4651,
      "step": 3895
    },
    {
      "epoch": 0.5042560766417451,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5616122484207153,
      "eval_runtime": 13.649,
      "eval_samples_per_second": 3.224,
      "eval_steps_per_second": 0.22,
      "step": 3895
    },
    {
      "epoch": 0.5043855390490987,
      "grad_norm": 1.9619250480295667,
      "learning_rate": 7.241523394929971e-07,
      "loss": 2.4622,
      "step": 3896
    },
    {
      "epoch": 0.5045150014564521,
      "grad_norm": 1.8031755554424054,
      "learning_rate": 7.238734300100383e-07,
      "loss": 2.4492,
      "step": 3897
    },
    {
      "epoch": 0.5046444638638056,
      "grad_norm": 3.7399693913344443,
      "learning_rate": 7.235945003301776e-07,
      "loss": 2.6377,
      "step": 3898
    },
    {
      "epoch": 0.504773926271159,
      "grad_norm": 3.5956869862531398,
      "learning_rate": 7.233155505103916e-07,
      "loss": 2.8918,
      "step": 3899
    },
    {
      "epoch": 0.5049033886785125,
      "grad_norm": 2.3889617369703333,
      "learning_rate": 7.230365806076618e-07,
      "loss": 2.5273,
      "step": 3900
    },
    {
      "epoch": 0.5049033886785125,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.560990810394287,
      "eval_runtime": 15.2679,
      "eval_samples_per_second": 2.882,
      "eval_steps_per_second": 0.196,
      "step": 3900
    },
    {
      "epoch": 0.5050328510858659,
      "grad_norm": 1.4238818349690794,
      "learning_rate": 7.227575906789731e-07,
      "loss": 2.6147,
      "step": 3901
    },
    {
      "epoch": 0.5051623134932194,
      "grad_norm": 1.6769802198278922,
      "learning_rate": 7.224785807813152e-07,
      "loss": 2.4956,
      "step": 3902
    },
    {
      "epoch": 0.5052917759005728,
      "grad_norm": 1.7444086808256676,
      "learning_rate": 7.221995509716815e-07,
      "loss": 2.5999,
      "step": 3903
    },
    {
      "epoch": 0.5054212383079263,
      "grad_norm": 1.4715754415043543,
      "learning_rate": 7.219205013070692e-07,
      "loss": 2.3098,
      "step": 3904
    },
    {
      "epoch": 0.5055507007152799,
      "grad_norm": 1.7070369466948605,
      "learning_rate": 7.216414318444799e-07,
      "loss": 2.5869,
      "step": 3905
    },
    {
      "epoch": 0.5055507007152799,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.559525966644287,
      "eval_runtime": 14.4639,
      "eval_samples_per_second": 3.042,
      "eval_steps_per_second": 0.207,
      "step": 3905
    },
    {
      "epoch": 0.5056801631226333,
      "grad_norm": 2.655850881190418,
      "learning_rate": 7.213623426409194e-07,
      "loss": 2.6255,
      "step": 3906
    },
    {
      "epoch": 0.5058096255299868,
      "grad_norm": 3.762788410686041,
      "learning_rate": 7.210832337533973e-07,
      "loss": 2.5059,
      "step": 3907
    },
    {
      "epoch": 0.5059390879373402,
      "grad_norm": 1.948873950767315,
      "learning_rate": 7.208041052389273e-07,
      "loss": 2.6294,
      "step": 3908
    },
    {
      "epoch": 0.5060685503446937,
      "grad_norm": 1.5547507570351142,
      "learning_rate": 7.205249571545268e-07,
      "loss": 2.5928,
      "step": 3909
    },
    {
      "epoch": 0.5061980127520471,
      "grad_norm": 3.789349689907264,
      "learning_rate": 7.202457895572176e-07,
      "loss": 2.895,
      "step": 3910
    },
    {
      "epoch": 0.5061980127520471,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5601474046707153,
      "eval_runtime": 14.714,
      "eval_samples_per_second": 2.99,
      "eval_steps_per_second": 0.204,
      "step": 3910
    },
    {
      "epoch": 0.5063274751594006,
      "grad_norm": 2.6536148555537125,
      "learning_rate": 7.199666025040253e-07,
      "loss": 2.9846,
      "step": 3911
    },
    {
      "epoch": 0.506456937566754,
      "grad_norm": 1.193291118307913,
      "learning_rate": 7.196873960519795e-07,
      "loss": 2.2935,
      "step": 3912
    },
    {
      "epoch": 0.5065863999741075,
      "grad_norm": 1.9758267587284766,
      "learning_rate": 7.194081702581141e-07,
      "loss": 2.604,
      "step": 3913
    },
    {
      "epoch": 0.5067158623814609,
      "grad_norm": 1.3407589899149672,
      "learning_rate": 7.191289251794663e-07,
      "loss": 2.3889,
      "step": 3914
    },
    {
      "epoch": 0.5068453247888145,
      "grad_norm": 1.6273721120582303,
      "learning_rate": 7.188496608730773e-07,
      "loss": 2.5203,
      "step": 3915
    },
    {
      "epoch": 0.5068453247888145,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.562943935394287,
      "eval_runtime": 14.343,
      "eval_samples_per_second": 3.068,
      "eval_steps_per_second": 0.209,
      "step": 3915
    },
    {
      "epoch": 0.506974787196168,
      "grad_norm": 2.3531567114958154,
      "learning_rate": 7.185703773959932e-07,
      "loss": 2.4966,
      "step": 3916
    },
    {
      "epoch": 0.5071042496035214,
      "grad_norm": 1.3115626162294656,
      "learning_rate": 7.182910748052627e-07,
      "loss": 2.3882,
      "step": 3917
    },
    {
      "epoch": 0.5072337120108749,
      "grad_norm": 1.4346453004413782,
      "learning_rate": 7.180117531579394e-07,
      "loss": 2.3844,
      "step": 3918
    },
    {
      "epoch": 0.5073631744182283,
      "grad_norm": 2.0330923507423715,
      "learning_rate": 7.177324125110801e-07,
      "loss": 2.386,
      "step": 3919
    },
    {
      "epoch": 0.5074926368255818,
      "grad_norm": 2.489217714225753,
      "learning_rate": 7.174530529217459e-07,
      "loss": 2.5239,
      "step": 3920
    },
    {
      "epoch": 0.5074926368255818,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5594815015792847,
      "eval_runtime": 13.2785,
      "eval_samples_per_second": 3.314,
      "eval_steps_per_second": 0.226,
      "step": 3920
    },
    {
      "epoch": 0.5076220992329352,
      "grad_norm": 2.594991739503318,
      "learning_rate": 7.171736744470019e-07,
      "loss": 2.5317,
      "step": 3921
    },
    {
      "epoch": 0.5077515616402887,
      "grad_norm": 1.5725199053060537,
      "learning_rate": 7.168942771439167e-07,
      "loss": 2.4084,
      "step": 3922
    },
    {
      "epoch": 0.5078810240476421,
      "grad_norm": 1.881095004664746,
      "learning_rate": 7.166148610695625e-07,
      "loss": 2.4331,
      "step": 3923
    },
    {
      "epoch": 0.5080104864549956,
      "grad_norm": 2.4266930175903747,
      "learning_rate": 7.16335426281016e-07,
      "loss": 2.5654,
      "step": 3924
    },
    {
      "epoch": 0.5081399488623491,
      "grad_norm": 1.796259741508993,
      "learning_rate": 7.160559728353576e-07,
      "loss": 2.4663,
      "step": 3925
    },
    {
      "epoch": 0.5081399488623491,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5594815015792847,
      "eval_runtime": 14.1639,
      "eval_samples_per_second": 3.106,
      "eval_steps_per_second": 0.212,
      "step": 3925
    },
    {
      "epoch": 0.5082694112697026,
      "grad_norm": 1.4423450777449243,
      "learning_rate": 7.157765007896709e-07,
      "loss": 2.3232,
      "step": 3926
    },
    {
      "epoch": 0.508398873677056,
      "grad_norm": 2.194804788908348,
      "learning_rate": 7.154970102010441e-07,
      "loss": 2.4922,
      "step": 3927
    },
    {
      "epoch": 0.5085283360844095,
      "grad_norm": 2.2222308648950824,
      "learning_rate": 7.152175011265688e-07,
      "loss": 2.4585,
      "step": 3928
    },
    {
      "epoch": 0.508657798491763,
      "grad_norm": 1.524060422558183,
      "learning_rate": 7.1493797362334e-07,
      "loss": 2.5251,
      "step": 3929
    },
    {
      "epoch": 0.5087872608991164,
      "grad_norm": 2.2265936815451086,
      "learning_rate": 7.146584277484574e-07,
      "loss": 2.3047,
      "step": 3930
    },
    {
      "epoch": 0.5087872608991164,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5573508739471436,
      "eval_runtime": 14.3982,
      "eval_samples_per_second": 3.056,
      "eval_steps_per_second": 0.208,
      "step": 3930
    },
    {
      "epoch": 0.5089167233064699,
      "grad_norm": 1.6994007040797923,
      "learning_rate": 7.143788635590235e-07,
      "loss": 2.5957,
      "step": 3931
    },
    {
      "epoch": 0.5090461857138233,
      "grad_norm": 1.5348308774429156,
      "learning_rate": 7.140992811121449e-07,
      "loss": 2.4495,
      "step": 3932
    },
    {
      "epoch": 0.5091756481211768,
      "grad_norm": 2.0577768073031377,
      "learning_rate": 7.138196804649325e-07,
      "loss": 2.7041,
      "step": 3933
    },
    {
      "epoch": 0.5093051105285302,
      "grad_norm": 2.1932914550121456,
      "learning_rate": 7.135400616744998e-07,
      "loss": 2.5571,
      "step": 3934
    },
    {
      "epoch": 0.5094345729358838,
      "grad_norm": 1.7350578899840934,
      "learning_rate": 7.132604247979651e-07,
      "loss": 2.5276,
      "step": 3935
    },
    {
      "epoch": 0.5094345729358838,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5574396848678589,
      "eval_runtime": 13.5905,
      "eval_samples_per_second": 3.238,
      "eval_steps_per_second": 0.221,
      "step": 3935
    },
    {
      "epoch": 0.5095640353432372,
      "grad_norm": 1.3493533949797523,
      "learning_rate": 7.129807698924496e-07,
      "loss": 2.3362,
      "step": 3936
    },
    {
      "epoch": 0.5096934977505907,
      "grad_norm": 1.497058971274792,
      "learning_rate": 7.127010970150784e-07,
      "loss": 2.262,
      "step": 3937
    },
    {
      "epoch": 0.5098229601579441,
      "grad_norm": 2.2848055261385163,
      "learning_rate": 7.124214062229806e-07,
      "loss": 2.6392,
      "step": 3938
    },
    {
      "epoch": 0.5099524225652976,
      "grad_norm": 1.5324710314121792,
      "learning_rate": 7.121416975732889e-07,
      "loss": 2.5166,
      "step": 3939
    },
    {
      "epoch": 0.5100818849726511,
      "grad_norm": 2.2047869311356485,
      "learning_rate": 7.118619711231389e-07,
      "loss": 2.6123,
      "step": 3940
    },
    {
      "epoch": 0.5100818849726511,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5545543432235718,
      "eval_runtime": 14.0838,
      "eval_samples_per_second": 3.124,
      "eval_steps_per_second": 0.213,
      "step": 3940
    },
    {
      "epoch": 0.5102113473800045,
      "grad_norm": 1.6734007618058986,
      "learning_rate": 7.115822269296708e-07,
      "loss": 2.4174,
      "step": 3941
    },
    {
      "epoch": 0.510340809787358,
      "grad_norm": 1.9927410656670073,
      "learning_rate": 7.113024650500278e-07,
      "loss": 2.4111,
      "step": 3942
    },
    {
      "epoch": 0.5104702721947114,
      "grad_norm": 3.482649138463418,
      "learning_rate": 7.110226855413571e-07,
      "loss": 2.5056,
      "step": 3943
    },
    {
      "epoch": 0.510599734602065,
      "grad_norm": 1.7149359690220838,
      "learning_rate": 7.107428884608095e-07,
      "loss": 2.4897,
      "step": 3944
    },
    {
      "epoch": 0.5107291970094184,
      "grad_norm": 1.2134008456724723,
      "learning_rate": 7.104630738655387e-07,
      "loss": 2.3611,
      "step": 3945
    },
    {
      "epoch": 0.5107291970094184,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.556596279144287,
      "eval_runtime": 14.6474,
      "eval_samples_per_second": 3.004,
      "eval_steps_per_second": 0.205,
      "step": 3945
    },
    {
      "epoch": 0.5108586594167719,
      "grad_norm": 2.0132074726290883,
      "learning_rate": 7.101832418127029e-07,
      "loss": 2.4846,
      "step": 3946
    },
    {
      "epoch": 0.5109881218241253,
      "grad_norm": 1.5541086630709071,
      "learning_rate": 7.099033923594634e-07,
      "loss": 2.3997,
      "step": 3947
    },
    {
      "epoch": 0.5111175842314788,
      "grad_norm": 1.3299252453677401,
      "learning_rate": 7.09623525562985e-07,
      "loss": 2.3718,
      "step": 3948
    },
    {
      "epoch": 0.5112470466388322,
      "grad_norm": 2.5221613480053224,
      "learning_rate": 7.093436414804363e-07,
      "loss": 2.4031,
      "step": 3949
    },
    {
      "epoch": 0.5113765090461857,
      "grad_norm": 1.4096505860276634,
      "learning_rate": 7.090637401689894e-07,
      "loss": 2.3093,
      "step": 3950
    },
    {
      "epoch": 0.5113765090461857,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5552645921707153,
      "eval_runtime": 13.9075,
      "eval_samples_per_second": 3.164,
      "eval_steps_per_second": 0.216,
      "step": 3950
    },
    {
      "epoch": 0.5115059714535392,
      "grad_norm": 2.066049643975334,
      "learning_rate": 7.087838216858194e-07,
      "loss": 2.4053,
      "step": 3951
    },
    {
      "epoch": 0.5116354338608926,
      "grad_norm": 1.7286364982543136,
      "learning_rate": 7.085038860881058e-07,
      "loss": 2.4419,
      "step": 3952
    },
    {
      "epoch": 0.5117648962682461,
      "grad_norm": 1.4051318477919965,
      "learning_rate": 7.082239334330309e-07,
      "loss": 2.396,
      "step": 3953
    },
    {
      "epoch": 0.5118943586755996,
      "grad_norm": 1.6146455394077133,
      "learning_rate": 7.079439637777808e-07,
      "loss": 2.5259,
      "step": 3954
    },
    {
      "epoch": 0.5120238210829531,
      "grad_norm": 3.449450569319739,
      "learning_rate": 7.076639771795447e-07,
      "loss": 2.5874,
      "step": 3955
    },
    {
      "epoch": 0.5120238210829531,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5573508739471436,
      "eval_runtime": 15.0765,
      "eval_samples_per_second": 2.918,
      "eval_steps_per_second": 0.199,
      "step": 3955
    },
    {
      "epoch": 0.5121532834903065,
      "grad_norm": 2.060301125091945,
      "learning_rate": 7.073839736955159e-07,
      "loss": 2.7292,
      "step": 3956
    },
    {
      "epoch": 0.51228274589766,
      "grad_norm": 2.9594007014426653,
      "learning_rate": 7.071039533828905e-07,
      "loss": 2.5938,
      "step": 3957
    },
    {
      "epoch": 0.5124122083050134,
      "grad_norm": 1.8220953475582509,
      "learning_rate": 7.068239162988687e-07,
      "loss": 2.478,
      "step": 3958
    },
    {
      "epoch": 0.5125416707123669,
      "grad_norm": 2.5455585823633564,
      "learning_rate": 7.065438625006533e-07,
      "loss": 2.5444,
      "step": 3959
    },
    {
      "epoch": 0.5126711331197203,
      "grad_norm": 2.134688878493036,
      "learning_rate": 7.06263792045451e-07,
      "loss": 2.4739,
      "step": 3960
    },
    {
      "epoch": 0.5126711331197203,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5510475635528564,
      "eval_runtime": 14.6108,
      "eval_samples_per_second": 3.011,
      "eval_steps_per_second": 0.205,
      "step": 3960
    },
    {
      "epoch": 0.5128005955270738,
      "grad_norm": 4.072923458133599,
      "learning_rate": 7.059837049904723e-07,
      "loss": 2.8989,
      "step": 3961
    },
    {
      "epoch": 0.5129300579344273,
      "grad_norm": 1.5879376346767393,
      "learning_rate": 7.057036013929304e-07,
      "loss": 2.5625,
      "step": 3962
    },
    {
      "epoch": 0.5130595203417807,
      "grad_norm": 1.853374831338135,
      "learning_rate": 7.05423481310042e-07,
      "loss": 2.3704,
      "step": 3963
    },
    {
      "epoch": 0.5131889827491343,
      "grad_norm": 3.11869981616625,
      "learning_rate": 7.051433447990275e-07,
      "loss": 2.665,
      "step": 3964
    },
    {
      "epoch": 0.5133184451564877,
      "grad_norm": 1.9902384840853993,
      "learning_rate": 7.048631919171103e-07,
      "loss": 2.4248,
      "step": 3965
    },
    {
      "epoch": 0.5133184451564877,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.55517578125,
      "eval_runtime": 13.9622,
      "eval_samples_per_second": 3.151,
      "eval_steps_per_second": 0.215,
      "step": 3965
    },
    {
      "epoch": 0.5134479075638412,
      "grad_norm": 1.4680747921815727,
      "learning_rate": 7.045830227215175e-07,
      "loss": 2.2361,
      "step": 3966
    },
    {
      "epoch": 0.5135773699711946,
      "grad_norm": 1.8960457945657165,
      "learning_rate": 7.043028372694794e-07,
      "loss": 2.5029,
      "step": 3967
    },
    {
      "epoch": 0.5137068323785481,
      "grad_norm": 1.7457386696546948,
      "learning_rate": 7.04022635618229e-07,
      "loss": 2.4343,
      "step": 3968
    },
    {
      "epoch": 0.5138362947859015,
      "grad_norm": 1.7655426299646468,
      "learning_rate": 7.037424178250038e-07,
      "loss": 2.4026,
      "step": 3969
    },
    {
      "epoch": 0.513965757193255,
      "grad_norm": 1.485243791585457,
      "learning_rate": 7.034621839470436e-07,
      "loss": 2.4216,
      "step": 3970
    },
    {
      "epoch": 0.513965757193255,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.552290439605713,
      "eval_runtime": 13.9204,
      "eval_samples_per_second": 3.161,
      "eval_steps_per_second": 0.216,
      "step": 3970
    },
    {
      "epoch": 0.5140952196006084,
      "grad_norm": 3.732439611569006,
      "learning_rate": 7.031819340415922e-07,
      "loss": 2.4043,
      "step": 3971
    },
    {
      "epoch": 0.5142246820079619,
      "grad_norm": 3.1296279316006124,
      "learning_rate": 7.029016681658959e-07,
      "loss": 2.5725,
      "step": 3972
    },
    {
      "epoch": 0.5143541444153154,
      "grad_norm": 1.833534723667105,
      "learning_rate": 7.026213863772048e-07,
      "loss": 2.7883,
      "step": 3973
    },
    {
      "epoch": 0.5144836068226689,
      "grad_norm": 1.7938191067710845,
      "learning_rate": 7.023410887327723e-07,
      "loss": 2.6987,
      "step": 3974
    },
    {
      "epoch": 0.5146130692300224,
      "grad_norm": 1.924824007484302,
      "learning_rate": 7.02060775289855e-07,
      "loss": 2.5598,
      "step": 3975
    },
    {
      "epoch": 0.5146130692300224,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5531116724014282,
      "eval_runtime": 14.4951,
      "eval_samples_per_second": 3.036,
      "eval_steps_per_second": 0.207,
      "step": 3975
    },
    {
      "epoch": 0.5147425316373758,
      "grad_norm": 1.832494485582077,
      "learning_rate": 7.017804461057122e-07,
      "loss": 2.7869,
      "step": 3976
    },
    {
      "epoch": 0.5148719940447293,
      "grad_norm": 3.2238741576907795,
      "learning_rate": 7.01500101237607e-07,
      "loss": 2.5723,
      "step": 3977
    },
    {
      "epoch": 0.5150014564520827,
      "grad_norm": 2.179860322047722,
      "learning_rate": 7.012197407428056e-07,
      "loss": 2.5591,
      "step": 3978
    },
    {
      "epoch": 0.5151309188594362,
      "grad_norm": 2.1896499209017257,
      "learning_rate": 7.009393646785773e-07,
      "loss": 2.4724,
      "step": 3979
    },
    {
      "epoch": 0.5152603812667896,
      "grad_norm": 1.3999990607693729,
      "learning_rate": 7.006589731021946e-07,
      "loss": 2.3999,
      "step": 3980
    },
    {
      "epoch": 0.5152603812667896,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5510032176971436,
      "eval_runtime": 14.3346,
      "eval_samples_per_second": 3.069,
      "eval_steps_per_second": 0.209,
      "step": 3980
    },
    {
      "epoch": 0.5153898436741431,
      "grad_norm": 2.923499299527811,
      "learning_rate": 7.003785660709329e-07,
      "loss": 2.4344,
      "step": 3981
    },
    {
      "epoch": 0.5155193060814965,
      "grad_norm": 2.7561421328259437,
      "learning_rate": 7.000981436420713e-07,
      "loss": 2.582,
      "step": 3982
    },
    {
      "epoch": 0.51564876848885,
      "grad_norm": 1.590944618509092,
      "learning_rate": 6.998177058728917e-07,
      "loss": 2.3236,
      "step": 3983
    },
    {
      "epoch": 0.5157782308962036,
      "grad_norm": 1.4916695803307185,
      "learning_rate": 6.995372528206793e-07,
      "loss": 2.3311,
      "step": 3984
    },
    {
      "epoch": 0.515907693303557,
      "grad_norm": 3.491227262486324,
      "learning_rate": 6.99256784542722e-07,
      "loss": 2.7056,
      "step": 3985
    },
    {
      "epoch": 0.515907693303557,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.550337314605713,
      "eval_runtime": 14.7473,
      "eval_samples_per_second": 2.984,
      "eval_steps_per_second": 0.203,
      "step": 3985
    },
    {
      "epoch": 0.5160371557109105,
      "grad_norm": 1.812358430861342,
      "learning_rate": 6.989763010963116e-07,
      "loss": 2.6284,
      "step": 3986
    },
    {
      "epoch": 0.5161666181182639,
      "grad_norm": 3.0750105845363764,
      "learning_rate": 6.98695802538742e-07,
      "loss": 2.5698,
      "step": 3987
    },
    {
      "epoch": 0.5162960805256174,
      "grad_norm": 1.6985289106885613,
      "learning_rate": 6.984152889273111e-07,
      "loss": 2.4294,
      "step": 3988
    },
    {
      "epoch": 0.5164255429329708,
      "grad_norm": 1.7012894884849699,
      "learning_rate": 6.981347603193194e-07,
      "loss": 2.5469,
      "step": 3989
    },
    {
      "epoch": 0.5165550053403243,
      "grad_norm": 2.6308187205415035,
      "learning_rate": 6.978542167720705e-07,
      "loss": 2.7463,
      "step": 3990
    },
    {
      "epoch": 0.5165550053403243,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5511142015457153,
      "eval_runtime": 14.3667,
      "eval_samples_per_second": 3.063,
      "eval_steps_per_second": 0.209,
      "step": 3990
    },
    {
      "epoch": 0.5166844677476777,
      "grad_norm": 1.2373409490242695,
      "learning_rate": 6.97573658342871e-07,
      "loss": 2.2629,
      "step": 3991
    },
    {
      "epoch": 0.5168139301550312,
      "grad_norm": 2.1358560904142605,
      "learning_rate": 6.97293085089031e-07,
      "loss": 2.4246,
      "step": 3992
    },
    {
      "epoch": 0.5169433925623848,
      "grad_norm": 1.7000094902465779,
      "learning_rate": 6.97012497067863e-07,
      "loss": 2.4238,
      "step": 3993
    },
    {
      "epoch": 0.5170728549697382,
      "grad_norm": 2.7223489131476413,
      "learning_rate": 6.967318943366828e-07,
      "loss": 2.5874,
      "step": 3994
    },
    {
      "epoch": 0.5172023173770917,
      "grad_norm": 5.732676217185138,
      "learning_rate": 6.964512769528094e-07,
      "loss": 3.3306,
      "step": 3995
    },
    {
      "epoch": 0.5172023173770917,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.548828125,
      "eval_runtime": 14.9967,
      "eval_samples_per_second": 2.934,
      "eval_steps_per_second": 0.2,
      "step": 3995
    },
    {
      "epoch": 0.5173317797844451,
      "grad_norm": 1.7775936567587456,
      "learning_rate": 6.96170644973564e-07,
      "loss": 2.6416,
      "step": 3996
    },
    {
      "epoch": 0.5174612421917986,
      "grad_norm": 1.7078130269268268,
      "learning_rate": 6.958899984562724e-07,
      "loss": 2.4714,
      "step": 3997
    },
    {
      "epoch": 0.517590704599152,
      "grad_norm": 1.9191300921989958,
      "learning_rate": 6.956093374582614e-07,
      "loss": 2.3521,
      "step": 3998
    },
    {
      "epoch": 0.5177201670065055,
      "grad_norm": 1.5271119778326883,
      "learning_rate": 6.95328662036862e-07,
      "loss": 2.3293,
      "step": 3999
    },
    {
      "epoch": 0.5178496294138589,
      "grad_norm": 1.501818407490641,
      "learning_rate": 6.95047972249408e-07,
      "loss": 2.5022,
      "step": 4000
    },
    {
      "epoch": 0.5178496294138589,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5495827198028564,
      "eval_runtime": 14.299,
      "eval_samples_per_second": 3.077,
      "eval_steps_per_second": 0.21,
      "step": 4000
    },
    {
      "epoch": 0.5179790918212124,
      "grad_norm": 2.3046162616095063,
      "learning_rate": 6.947672681532358e-07,
      "loss": 2.5112,
      "step": 4001
    },
    {
      "epoch": 0.5181085542285658,
      "grad_norm": 1.274892818735378,
      "learning_rate": 6.944865498056849e-07,
      "loss": 2.2561,
      "step": 4002
    },
    {
      "epoch": 0.5182380166359194,
      "grad_norm": 1.3678346798583385,
      "learning_rate": 6.942058172640981e-07,
      "loss": 2.2942,
      "step": 4003
    },
    {
      "epoch": 0.5183674790432728,
      "grad_norm": 1.6670188504381802,
      "learning_rate": 6.9392507058582e-07,
      "loss": 2.4956,
      "step": 4004
    },
    {
      "epoch": 0.5184969414506263,
      "grad_norm": 1.8297421986920928,
      "learning_rate": 6.936443098281993e-07,
      "loss": 2.6304,
      "step": 4005
    },
    {
      "epoch": 0.5184969414506263,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5488947629928589,
      "eval_runtime": 13.9259,
      "eval_samples_per_second": 3.16,
      "eval_steps_per_second": 0.215,
      "step": 4005
    },
    {
      "epoch": 0.5186264038579798,
      "grad_norm": 2.281361998524018,
      "learning_rate": 6.93363535048587e-07,
      "loss": 2.3857,
      "step": 4006
    },
    {
      "epoch": 0.5187558662653332,
      "grad_norm": 1.8496701172040177,
      "learning_rate": 6.93082746304337e-07,
      "loss": 2.5203,
      "step": 4007
    },
    {
      "epoch": 0.5188853286726867,
      "grad_norm": 2.9370386479711414,
      "learning_rate": 6.928019436528062e-07,
      "loss": 2.5742,
      "step": 4008
    },
    {
      "epoch": 0.5190147910800401,
      "grad_norm": 1.5975234790838664,
      "learning_rate": 6.92521127151354e-07,
      "loss": 2.3896,
      "step": 4009
    },
    {
      "epoch": 0.5191442534873936,
      "grad_norm": 1.5829937977805089,
      "learning_rate": 6.922402968573429e-07,
      "loss": 2.3433,
      "step": 4010
    },
    {
      "epoch": 0.5191442534873936,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.545365810394287,
      "eval_runtime": 13.7942,
      "eval_samples_per_second": 3.19,
      "eval_steps_per_second": 0.217,
      "step": 4010
    },
    {
      "epoch": 0.519273715894747,
      "grad_norm": 2.259205106524019,
      "learning_rate": 6.919594528281384e-07,
      "loss": 2.7288,
      "step": 4011
    },
    {
      "epoch": 0.5194031783021005,
      "grad_norm": 1.8968202224773139,
      "learning_rate": 6.916785951211084e-07,
      "loss": 2.1353,
      "step": 4012
    },
    {
      "epoch": 0.519532640709454,
      "grad_norm": 2.9708380498163365,
      "learning_rate": 6.913977237936239e-07,
      "loss": 3.1262,
      "step": 4013
    },
    {
      "epoch": 0.5196621031168075,
      "grad_norm": 1.4358213582854178,
      "learning_rate": 6.911168389030584e-07,
      "loss": 2.5811,
      "step": 4014
    },
    {
      "epoch": 0.519791565524161,
      "grad_norm": 1.5399669742781321,
      "learning_rate": 6.908359405067885e-07,
      "loss": 2.3269,
      "step": 4015
    },
    {
      "epoch": 0.519791565524161,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5490056276321411,
      "eval_runtime": 14.1061,
      "eval_samples_per_second": 3.119,
      "eval_steps_per_second": 0.213,
      "step": 4015
    },
    {
      "epoch": 0.5199210279315144,
      "grad_norm": 1.3322671984492085,
      "learning_rate": 6.905550286621933e-07,
      "loss": 2.4346,
      "step": 4016
    },
    {
      "epoch": 0.5200504903388679,
      "grad_norm": 1.5827936533204134,
      "learning_rate": 6.902741034266548e-07,
      "loss": 2.3481,
      "step": 4017
    },
    {
      "epoch": 0.5201799527462213,
      "grad_norm": 1.2818233116674882,
      "learning_rate": 6.899931648575575e-07,
      "loss": 2.4534,
      "step": 4018
    },
    {
      "epoch": 0.5203094151535748,
      "grad_norm": 1.8177028507779016,
      "learning_rate": 6.89712213012289e-07,
      "loss": 2.6448,
      "step": 4019
    },
    {
      "epoch": 0.5204388775609282,
      "grad_norm": 2.977795867828807,
      "learning_rate": 6.894312479482394e-07,
      "loss": 2.3749,
      "step": 4020
    },
    {
      "epoch": 0.5204388775609282,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5468083620071411,
      "eval_runtime": 13.2784,
      "eval_samples_per_second": 3.314,
      "eval_steps_per_second": 0.226,
      "step": 4020
    },
    {
      "epoch": 0.5205683399682817,
      "grad_norm": 3.8761697200983587,
      "learning_rate": 6.891502697228014e-07,
      "loss": 2.6353,
      "step": 4021
    },
    {
      "epoch": 0.5206978023756351,
      "grad_norm": 2.0277540872485567,
      "learning_rate": 6.888692783933706e-07,
      "loss": 2.3862,
      "step": 4022
    },
    {
      "epoch": 0.5208272647829887,
      "grad_norm": 2.5307349281963223,
      "learning_rate": 6.88588274017345e-07,
      "loss": 2.5737,
      "step": 4023
    },
    {
      "epoch": 0.5209567271903421,
      "grad_norm": 2.069943371719611,
      "learning_rate": 6.883072566521255e-07,
      "loss": 2.4773,
      "step": 4024
    },
    {
      "epoch": 0.5210861895976956,
      "grad_norm": 2.153232460800363,
      "learning_rate": 6.880262263551159e-07,
      "loss": 2.4248,
      "step": 4025
    },
    {
      "epoch": 0.5210861895976956,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5490056276321411,
      "eval_runtime": 13.8471,
      "eval_samples_per_second": 3.178,
      "eval_steps_per_second": 0.217,
      "step": 4025
    },
    {
      "epoch": 0.521215652005049,
      "grad_norm": 1.6353442719300793,
      "learning_rate": 6.877451831837219e-07,
      "loss": 2.3752,
      "step": 4026
    },
    {
      "epoch": 0.5213451144124025,
      "grad_norm": 2.2208258264771548,
      "learning_rate": 6.874641271953525e-07,
      "loss": 2.4895,
      "step": 4027
    },
    {
      "epoch": 0.521474576819756,
      "grad_norm": 2.213323882831792,
      "learning_rate": 6.871830584474189e-07,
      "loss": 2.5103,
      "step": 4028
    },
    {
      "epoch": 0.5216040392271094,
      "grad_norm": 1.9378846944831578,
      "learning_rate": 6.869019769973354e-07,
      "loss": 2.4146,
      "step": 4029
    },
    {
      "epoch": 0.5217335016344629,
      "grad_norm": 1.8529593255096792,
      "learning_rate": 6.866208829025182e-07,
      "loss": 2.3682,
      "step": 4030
    },
    {
      "epoch": 0.5217335016344629,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.544633388519287,
      "eval_runtime": 13.697,
      "eval_samples_per_second": 3.212,
      "eval_steps_per_second": 0.219,
      "step": 4030
    },
    {
      "epoch": 0.5218629640418163,
      "grad_norm": 1.6765713081694857,
      "learning_rate": 6.86339776220387e-07,
      "loss": 2.269,
      "step": 4031
    },
    {
      "epoch": 0.5219924264491699,
      "grad_norm": 2.158720777122161,
      "learning_rate": 6.860586570083629e-07,
      "loss": 2.4163,
      "step": 4032
    },
    {
      "epoch": 0.5221218888565233,
      "grad_norm": 1.8396302914952403,
      "learning_rate": 6.857775253238706e-07,
      "loss": 2.6018,
      "step": 4033
    },
    {
      "epoch": 0.5222513512638768,
      "grad_norm": 2.89006099770384,
      "learning_rate": 6.854963812243369e-07,
      "loss": 2.7969,
      "step": 4034
    },
    {
      "epoch": 0.5223808136712302,
      "grad_norm": 1.2566850857976133,
      "learning_rate": 6.852152247671911e-07,
      "loss": 2.2957,
      "step": 4035
    },
    {
      "epoch": 0.5223808136712302,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5445889234542847,
      "eval_runtime": 13.8081,
      "eval_samples_per_second": 3.187,
      "eval_steps_per_second": 0.217,
      "step": 4035
    },
    {
      "epoch": 0.5225102760785837,
      "grad_norm": 1.5335537435467135,
      "learning_rate": 6.84934056009865e-07,
      "loss": 2.46,
      "step": 4036
    },
    {
      "epoch": 0.5226397384859371,
      "grad_norm": 3.795728024773409,
      "learning_rate": 6.846528750097933e-07,
      "loss": 2.5276,
      "step": 4037
    },
    {
      "epoch": 0.5227692008932906,
      "grad_norm": 1.3880231422149454,
      "learning_rate": 6.843716818244129e-07,
      "loss": 2.48,
      "step": 4038
    },
    {
      "epoch": 0.522898663300644,
      "grad_norm": 1.4284678586421233,
      "learning_rate": 6.840904765111629e-07,
      "loss": 2.4761,
      "step": 4039
    },
    {
      "epoch": 0.5230281257079975,
      "grad_norm": 2.2523568981713273,
      "learning_rate": 6.838092591274855e-07,
      "loss": 2.5649,
      "step": 4040
    },
    {
      "epoch": 0.5230281257079975,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5475407838821411,
      "eval_runtime": 15.0819,
      "eval_samples_per_second": 2.917,
      "eval_steps_per_second": 0.199,
      "step": 4040
    },
    {
      "epoch": 0.523157588115351,
      "grad_norm": 1.2029077852565317,
      "learning_rate": 6.835280297308247e-07,
      "loss": 2.3252,
      "step": 4041
    },
    {
      "epoch": 0.5232870505227045,
      "grad_norm": 1.3756174207563119,
      "learning_rate": 6.832467883786276e-07,
      "loss": 2.3965,
      "step": 4042
    },
    {
      "epoch": 0.523416512930058,
      "grad_norm": 1.4355098504264527,
      "learning_rate": 6.829655351283434e-07,
      "loss": 2.4001,
      "step": 4043
    },
    {
      "epoch": 0.5235459753374114,
      "grad_norm": 2.45334991143882,
      "learning_rate": 6.826842700374238e-07,
      "loss": 2.6587,
      "step": 4044
    },
    {
      "epoch": 0.5236754377447649,
      "grad_norm": 1.531338741425381,
      "learning_rate": 6.824029931633227e-07,
      "loss": 2.3848,
      "step": 4045
    },
    {
      "epoch": 0.5236754377447649,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5446555614471436,
      "eval_runtime": 14.4656,
      "eval_samples_per_second": 3.042,
      "eval_steps_per_second": 0.207,
      "step": 4045
    },
    {
      "epoch": 0.5238049001521183,
      "grad_norm": 1.8417709072515376,
      "learning_rate": 6.821217045634968e-07,
      "loss": 2.2681,
      "step": 4046
    },
    {
      "epoch": 0.5239343625594718,
      "grad_norm": 2.226281162429195,
      "learning_rate": 6.81840404295405e-07,
      "loss": 2.2002,
      "step": 4047
    },
    {
      "epoch": 0.5240638249668252,
      "grad_norm": 2.2354565806073716,
      "learning_rate": 6.815590924165085e-07,
      "loss": 2.6931,
      "step": 4048
    },
    {
      "epoch": 0.5241932873741787,
      "grad_norm": 1.9605992545028588,
      "learning_rate": 6.812777689842708e-07,
      "loss": 2.3494,
      "step": 4049
    },
    {
      "epoch": 0.5243227497815321,
      "grad_norm": 1.6412602427597318,
      "learning_rate": 6.809964340561583e-07,
      "loss": 2.4158,
      "step": 4050
    },
    {
      "epoch": 0.5243227497815321,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.543257236480713,
      "eval_runtime": 14.0658,
      "eval_samples_per_second": 3.128,
      "eval_steps_per_second": 0.213,
      "step": 4050
    },
    {
      "epoch": 0.5244522121888856,
      "grad_norm": 1.649246079162323,
      "learning_rate": 6.80715087689639e-07,
      "loss": 2.3794,
      "step": 4051
    },
    {
      "epoch": 0.5245816745962392,
      "grad_norm": 2.5638443834562086,
      "learning_rate": 6.804337299421837e-07,
      "loss": 2.5198,
      "step": 4052
    },
    {
      "epoch": 0.5247111370035926,
      "grad_norm": 2.1241400341109533,
      "learning_rate": 6.801523608712659e-07,
      "loss": 2.5381,
      "step": 4053
    },
    {
      "epoch": 0.5248405994109461,
      "grad_norm": 4.498602466975845,
      "learning_rate": 6.798709805343601e-07,
      "loss": 2.5005,
      "step": 4054
    },
    {
      "epoch": 0.5249700618182995,
      "grad_norm": 2.781289622954872,
      "learning_rate": 6.795895889889443e-07,
      "loss": 2.6262,
      "step": 4055
    },
    {
      "epoch": 0.5249700618182995,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5418590307235718,
      "eval_runtime": 15.221,
      "eval_samples_per_second": 2.891,
      "eval_steps_per_second": 0.197,
      "step": 4055
    },
    {
      "epoch": 0.525099524225653,
      "grad_norm": 3.0928074247533215,
      "learning_rate": 6.793081862924987e-07,
      "loss": 2.7434,
      "step": 4056
    },
    {
      "epoch": 0.5252289866330064,
      "grad_norm": 2.0552540618606683,
      "learning_rate": 6.790267725025052e-07,
      "loss": 2.5503,
      "step": 4057
    },
    {
      "epoch": 0.5253584490403599,
      "grad_norm": 1.8352263219916283,
      "learning_rate": 6.787453476764483e-07,
      "loss": 2.3716,
      "step": 4058
    },
    {
      "epoch": 0.5254879114477133,
      "grad_norm": 1.8612845516455199,
      "learning_rate": 6.784639118718146e-07,
      "loss": 2.5055,
      "step": 4059
    },
    {
      "epoch": 0.5256173738550668,
      "grad_norm": 2.5825970119011976,
      "learning_rate": 6.781824651460934e-07,
      "loss": 2.7236,
      "step": 4060
    },
    {
      "epoch": 0.5256173738550668,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5419255495071411,
      "eval_runtime": 14.1009,
      "eval_samples_per_second": 3.12,
      "eval_steps_per_second": 0.213,
      "step": 4060
    },
    {
      "epoch": 0.5257468362624202,
      "grad_norm": 3.364303685325622,
      "learning_rate": 6.77901007556776e-07,
      "loss": 2.4731,
      "step": 4061
    },
    {
      "epoch": 0.5258762986697738,
      "grad_norm": 1.4290990338587204,
      "learning_rate": 6.776195391613552e-07,
      "loss": 2.2952,
      "step": 4062
    },
    {
      "epoch": 0.5260057610771273,
      "grad_norm": 2.0174493980427184,
      "learning_rate": 6.773380600173271e-07,
      "loss": 2.6475,
      "step": 4063
    },
    {
      "epoch": 0.5261352234844807,
      "grad_norm": 1.8965380072530311,
      "learning_rate": 6.770565701821893e-07,
      "loss": 2.6648,
      "step": 4064
    },
    {
      "epoch": 0.5262646858918342,
      "grad_norm": 2.940568522815605,
      "learning_rate": 6.76775069713442e-07,
      "loss": 2.2603,
      "step": 4065
    },
    {
      "epoch": 0.5262646858918342,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5426357984542847,
      "eval_runtime": 14.3713,
      "eval_samples_per_second": 3.062,
      "eval_steps_per_second": 0.209,
      "step": 4065
    },
    {
      "epoch": 0.5263941482991876,
      "grad_norm": 1.276968373334518,
      "learning_rate": 6.764935586685873e-07,
      "loss": 2.2014,
      "step": 4066
    },
    {
      "epoch": 0.5265236107065411,
      "grad_norm": 2.467185405032055,
      "learning_rate": 6.762120371051296e-07,
      "loss": 2.8103,
      "step": 4067
    },
    {
      "epoch": 0.5266530731138945,
      "grad_norm": 1.6662267916379692,
      "learning_rate": 6.75930505080575e-07,
      "loss": 2.4114,
      "step": 4068
    },
    {
      "epoch": 0.526782535521248,
      "grad_norm": 1.369969145436834,
      "learning_rate": 6.756489626524326e-07,
      "loss": 2.4111,
      "step": 4069
    },
    {
      "epoch": 0.5269119979286014,
      "grad_norm": 1.3334883998849394,
      "learning_rate": 6.753674098782131e-07,
      "loss": 2.3872,
      "step": 4070
    },
    {
      "epoch": 0.5269119979286014,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5459872484207153,
      "eval_runtime": 14.3671,
      "eval_samples_per_second": 3.063,
      "eval_steps_per_second": 0.209,
      "step": 4070
    },
    {
      "epoch": 0.5270414603359549,
      "grad_norm": 1.7450215341448398,
      "learning_rate": 6.750858468154291e-07,
      "loss": 2.6187,
      "step": 4071
    },
    {
      "epoch": 0.5271709227433085,
      "grad_norm": 2.390558197174125,
      "learning_rate": 6.748042735215957e-07,
      "loss": 2.5771,
      "step": 4072
    },
    {
      "epoch": 0.5273003851506619,
      "grad_norm": 1.2022305620236275,
      "learning_rate": 6.745226900542298e-07,
      "loss": 2.2388,
      "step": 4073
    },
    {
      "epoch": 0.5274298475580154,
      "grad_norm": 1.8862366979778469,
      "learning_rate": 6.742410964708509e-07,
      "loss": 2.6208,
      "step": 4074
    },
    {
      "epoch": 0.5275593099653688,
      "grad_norm": 1.8438641976208832,
      "learning_rate": 6.7395949282898e-07,
      "loss": 2.5164,
      "step": 4075
    },
    {
      "epoch": 0.5275593099653688,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5446999073028564,
      "eval_runtime": 15.033,
      "eval_samples_per_second": 2.927,
      "eval_steps_per_second": 0.2,
      "step": 4075
    },
    {
      "epoch": 0.5276887723727223,
      "grad_norm": 2.251961850422397,
      "learning_rate": 6.736778791861403e-07,
      "loss": 2.428,
      "step": 4076
    },
    {
      "epoch": 0.5278182347800757,
      "grad_norm": 1.820521444568679,
      "learning_rate": 6.733962555998569e-07,
      "loss": 2.3325,
      "step": 4077
    },
    {
      "epoch": 0.5279476971874292,
      "grad_norm": 1.9899972253345364,
      "learning_rate": 6.731146221276575e-07,
      "loss": 2.5508,
      "step": 4078
    },
    {
      "epoch": 0.5280771595947826,
      "grad_norm": 2.0062209838055205,
      "learning_rate": 6.728329788270713e-07,
      "loss": 2.7676,
      "step": 4079
    },
    {
      "epoch": 0.5282066220021361,
      "grad_norm": 2.3639440229428352,
      "learning_rate": 6.725513257556296e-07,
      "loss": 2.4801,
      "step": 4080
    },
    {
      "epoch": 0.5282066220021361,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5431462526321411,
      "eval_runtime": 14.5526,
      "eval_samples_per_second": 3.024,
      "eval_steps_per_second": 0.206,
      "step": 4080
    },
    {
      "epoch": 0.5283360844094896,
      "grad_norm": 1.8032323195283315,
      "learning_rate": 6.72269662970866e-07,
      "loss": 2.3364,
      "step": 4081
    },
    {
      "epoch": 0.5284655468168431,
      "grad_norm": 2.0547809072510845,
      "learning_rate": 6.719879905303152e-07,
      "loss": 2.533,
      "step": 4082
    },
    {
      "epoch": 0.5285950092241966,
      "grad_norm": 2.206939725330552,
      "learning_rate": 6.717063084915153e-07,
      "loss": 2.5942,
      "step": 4083
    },
    {
      "epoch": 0.52872447163155,
      "grad_norm": 2.092499638160591,
      "learning_rate": 6.71424616912005e-07,
      "loss": 2.3096,
      "step": 4084
    },
    {
      "epoch": 0.5288539340389035,
      "grad_norm": 1.9624236100112382,
      "learning_rate": 6.711429158493259e-07,
      "loss": 2.2903,
      "step": 4085
    },
    {
      "epoch": 0.5288539340389035,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.543212890625,
      "eval_runtime": 14.9978,
      "eval_samples_per_second": 2.934,
      "eval_steps_per_second": 0.2,
      "step": 4085
    },
    {
      "epoch": 0.5289833964462569,
      "grad_norm": 2.652543670973225,
      "learning_rate": 6.708612053610207e-07,
      "loss": 2.6018,
      "step": 4086
    },
    {
      "epoch": 0.5291128588536104,
      "grad_norm": 1.377147046044787,
      "learning_rate": 6.705794855046349e-07,
      "loss": 2.344,
      "step": 4087
    },
    {
      "epoch": 0.5292423212609638,
      "grad_norm": 1.6605268731313942,
      "learning_rate": 6.702977563377153e-07,
      "loss": 2.2849,
      "step": 4088
    },
    {
      "epoch": 0.5293717836683173,
      "grad_norm": 2.027139648045615,
      "learning_rate": 6.700160179178109e-07,
      "loss": 2.5793,
      "step": 4089
    },
    {
      "epoch": 0.5295012460756707,
      "grad_norm": 1.3878072962026224,
      "learning_rate": 6.697342703024724e-07,
      "loss": 2.302,
      "step": 4090
    },
    {
      "epoch": 0.5295012460756707,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5409489870071411,
      "eval_runtime": 13.6891,
      "eval_samples_per_second": 3.214,
      "eval_steps_per_second": 0.219,
      "step": 4090
    },
    {
      "epoch": 0.5296307084830243,
      "grad_norm": 1.3877954933325167,
      "learning_rate": 6.694525135492523e-07,
      "loss": 2.4109,
      "step": 4091
    },
    {
      "epoch": 0.5297601708903777,
      "grad_norm": 1.3435178934375451,
      "learning_rate": 6.691707477157055e-07,
      "loss": 2.2239,
      "step": 4092
    },
    {
      "epoch": 0.5298896332977312,
      "grad_norm": 3.237947114871341,
      "learning_rate": 6.688889728593881e-07,
      "loss": 2.8003,
      "step": 4093
    },
    {
      "epoch": 0.5300190957050847,
      "grad_norm": 1.5673185934234255,
      "learning_rate": 6.686071890378584e-07,
      "loss": 2.395,
      "step": 4094
    },
    {
      "epoch": 0.5301485581124381,
      "grad_norm": 2.051905149309898,
      "learning_rate": 6.683253963086764e-07,
      "loss": 2.5671,
      "step": 4095
    },
    {
      "epoch": 0.5301485581124381,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5425692796707153,
      "eval_runtime": 13.9015,
      "eval_samples_per_second": 3.165,
      "eval_steps_per_second": 0.216,
      "step": 4095
    },
    {
      "epoch": 0.5302780205197916,
      "grad_norm": 1.849445812197663,
      "learning_rate": 6.680435947294041e-07,
      "loss": 2.6562,
      "step": 4096
    },
    {
      "epoch": 0.530407482927145,
      "grad_norm": 2.6091837391899984,
      "learning_rate": 6.677617843576052e-07,
      "loss": 2.7395,
      "step": 4097
    },
    {
      "epoch": 0.5305369453344985,
      "grad_norm": 1.795511185540601,
      "learning_rate": 6.67479965250845e-07,
      "loss": 2.4978,
      "step": 4098
    },
    {
      "epoch": 0.5306664077418519,
      "grad_norm": 2.3018198330380897,
      "learning_rate": 6.67198137466691e-07,
      "loss": 2.5415,
      "step": 4099
    },
    {
      "epoch": 0.5307958701492054,
      "grad_norm": 1.3505060470833354,
      "learning_rate": 6.66916301062712e-07,
      "loss": 2.3435,
      "step": 4100
    },
    {
      "epoch": 0.5307958701492054,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5418146848678589,
      "eval_runtime": 14.2788,
      "eval_samples_per_second": 3.081,
      "eval_steps_per_second": 0.21,
      "step": 4100
    },
    {
      "epoch": 0.5309253325565589,
      "grad_norm": 1.3975667251220192,
      "learning_rate": 6.666344560964788e-07,
      "loss": 2.2227,
      "step": 4101
    },
    {
      "epoch": 0.5310547949639124,
      "grad_norm": 1.5539573154826345,
      "learning_rate": 6.663526026255641e-07,
      "loss": 2.4458,
      "step": 4102
    },
    {
      "epoch": 0.5311842573712658,
      "grad_norm": 2.344456631655416,
      "learning_rate": 6.660707407075423e-07,
      "loss": 2.5728,
      "step": 4103
    },
    {
      "epoch": 0.5313137197786193,
      "grad_norm": 2.124932888646604,
      "learning_rate": 6.657888703999892e-07,
      "loss": 2.606,
      "step": 4104
    },
    {
      "epoch": 0.5314431821859728,
      "grad_norm": 1.748060373519471,
      "learning_rate": 6.655069917604822e-07,
      "loss": 2.4187,
      "step": 4105
    },
    {
      "epoch": 0.5314431821859728,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5445224046707153,
      "eval_runtime": 14.5717,
      "eval_samples_per_second": 3.02,
      "eval_steps_per_second": 0.206,
      "step": 4105
    },
    {
      "epoch": 0.5315726445933262,
      "grad_norm": 1.543397018032056,
      "learning_rate": 6.652251048466016e-07,
      "loss": 2.3669,
      "step": 4106
    },
    {
      "epoch": 0.5317021070006797,
      "grad_norm": 2.7426462524261597,
      "learning_rate": 6.649432097159278e-07,
      "loss": 2.3608,
      "step": 4107
    },
    {
      "epoch": 0.5318315694080331,
      "grad_norm": 2.2296753841059096,
      "learning_rate": 6.646613064260439e-07,
      "loss": 2.6189,
      "step": 4108
    },
    {
      "epoch": 0.5319610318153866,
      "grad_norm": 1.5627983300053554,
      "learning_rate": 6.64379395034534e-07,
      "loss": 2.3882,
      "step": 4109
    },
    {
      "epoch": 0.53209049422274,
      "grad_norm": 1.994126962465963,
      "learning_rate": 6.640974755989846e-07,
      "loss": 2.4468,
      "step": 4110
    },
    {
      "epoch": 0.53209049422274,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.545942783355713,
      "eval_runtime": 13.7882,
      "eval_samples_per_second": 3.191,
      "eval_steps_per_second": 0.218,
      "step": 4110
    },
    {
      "epoch": 0.5322199566300936,
      "grad_norm": 1.3010825712566525,
      "learning_rate": 6.638155481769836e-07,
      "loss": 2.4158,
      "step": 4111
    },
    {
      "epoch": 0.532349419037447,
      "grad_norm": 2.0143028073878986,
      "learning_rate": 6.635336128261198e-07,
      "loss": 2.3218,
      "step": 4112
    },
    {
      "epoch": 0.5324788814448005,
      "grad_norm": 2.252579202225458,
      "learning_rate": 6.632516696039847e-07,
      "loss": 2.5225,
      "step": 4113
    },
    {
      "epoch": 0.5326083438521539,
      "grad_norm": 3.0033230100128896,
      "learning_rate": 6.629697185681706e-07,
      "loss": 2.813,
      "step": 4114
    },
    {
      "epoch": 0.5327378062595074,
      "grad_norm": 1.572045835248253,
      "learning_rate": 6.626877597762721e-07,
      "loss": 2.6624,
      "step": 4115
    },
    {
      "epoch": 0.5327378062595074,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5383079051971436,
      "eval_runtime": 13.7374,
      "eval_samples_per_second": 3.203,
      "eval_steps_per_second": 0.218,
      "step": 4115
    },
    {
      "epoch": 0.5328672686668608,
      "grad_norm": 2.0625186151816677,
      "learning_rate": 6.624057932858846e-07,
      "loss": 2.5728,
      "step": 4116
    },
    {
      "epoch": 0.5329967310742143,
      "grad_norm": 2.4584282769901447,
      "learning_rate": 6.621238191546057e-07,
      "loss": 2.5815,
      "step": 4117
    },
    {
      "epoch": 0.5331261934815678,
      "grad_norm": 3.739672032630529,
      "learning_rate": 6.618418374400341e-07,
      "loss": 2.9109,
      "step": 4118
    },
    {
      "epoch": 0.5332556558889212,
      "grad_norm": 2.22265460600069,
      "learning_rate": 6.615598481997707e-07,
      "loss": 2.6689,
      "step": 4119
    },
    {
      "epoch": 0.5333851182962748,
      "grad_norm": 1.269188844666255,
      "learning_rate": 6.612778514914172e-07,
      "loss": 2.48,
      "step": 4120
    },
    {
      "epoch": 0.5333851182962748,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5423473119735718,
      "eval_runtime": 14.4042,
      "eval_samples_per_second": 3.055,
      "eval_steps_per_second": 0.208,
      "step": 4120
    },
    {
      "epoch": 0.5335145807036282,
      "grad_norm": 1.7695175092101905,
      "learning_rate": 6.609958473725774e-07,
      "loss": 2.5378,
      "step": 4121
    },
    {
      "epoch": 0.5336440431109817,
      "grad_norm": 2.928718448413823,
      "learning_rate": 6.607138359008558e-07,
      "loss": 2.4871,
      "step": 4122
    },
    {
      "epoch": 0.5337735055183351,
      "grad_norm": 3.269442266337227,
      "learning_rate": 6.604318171338595e-07,
      "loss": 3.1372,
      "step": 4123
    },
    {
      "epoch": 0.5339029679256886,
      "grad_norm": 1.1861803220090243,
      "learning_rate": 6.601497911291964e-07,
      "loss": 2.2971,
      "step": 4124
    },
    {
      "epoch": 0.534032430333042,
      "grad_norm": 1.578764117420061,
      "learning_rate": 6.598677579444761e-07,
      "loss": 2.5684,
      "step": 4125
    },
    {
      "epoch": 0.534032430333042,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5388849973678589,
      "eval_runtime": 14.2809,
      "eval_samples_per_second": 3.081,
      "eval_steps_per_second": 0.21,
      "step": 4125
    },
    {
      "epoch": 0.5341618927403955,
      "grad_norm": 2.086286422744801,
      "learning_rate": 6.595857176373094e-07,
      "loss": 2.4326,
      "step": 4126
    },
    {
      "epoch": 0.534291355147749,
      "grad_norm": 2.2371130585898817,
      "learning_rate": 6.593036702653087e-07,
      "loss": 2.2461,
      "step": 4127
    },
    {
      "epoch": 0.5344208175551024,
      "grad_norm": 1.3533887585013036,
      "learning_rate": 6.590216158860884e-07,
      "loss": 2.5505,
      "step": 4128
    },
    {
      "epoch": 0.5345502799624559,
      "grad_norm": 3.522146648313255,
      "learning_rate": 6.587395545572632e-07,
      "loss": 2.916,
      "step": 4129
    },
    {
      "epoch": 0.5346797423698094,
      "grad_norm": 1.7016474924008749,
      "learning_rate": 6.584574863364501e-07,
      "loss": 2.2529,
      "step": 4130
    },
    {
      "epoch": 0.5346797423698094,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5431462526321411,
      "eval_runtime": 14.761,
      "eval_samples_per_second": 2.981,
      "eval_steps_per_second": 0.203,
      "step": 4130
    },
    {
      "epoch": 0.5348092047771629,
      "grad_norm": 1.5306119576768233,
      "learning_rate": 6.581754112812673e-07,
      "loss": 2.4597,
      "step": 4131
    },
    {
      "epoch": 0.5349386671845163,
      "grad_norm": 1.262342395383607,
      "learning_rate": 6.578933294493344e-07,
      "loss": 2.2175,
      "step": 4132
    },
    {
      "epoch": 0.5350681295918698,
      "grad_norm": 2.7684296116377642,
      "learning_rate": 6.576112408982721e-07,
      "loss": 2.4918,
      "step": 4133
    },
    {
      "epoch": 0.5351975919992232,
      "grad_norm": 4.262698770451707,
      "learning_rate": 6.573291456857031e-07,
      "loss": 2.2244,
      "step": 4134
    },
    {
      "epoch": 0.5353270544065767,
      "grad_norm": 1.5831741269025994,
      "learning_rate": 6.570470438692507e-07,
      "loss": 2.5874,
      "step": 4135
    },
    {
      "epoch": 0.5353270544065767,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.540971279144287,
      "eval_runtime": 14.3312,
      "eval_samples_per_second": 3.07,
      "eval_steps_per_second": 0.209,
      "step": 4135
    },
    {
      "epoch": 0.5354565168139301,
      "grad_norm": 2.356478319566696,
      "learning_rate": 6.567649355065399e-07,
      "loss": 2.522,
      "step": 4136
    },
    {
      "epoch": 0.5355859792212836,
      "grad_norm": 2.2150551599205524,
      "learning_rate": 6.564828206551973e-07,
      "loss": 2.3005,
      "step": 4137
    },
    {
      "epoch": 0.535715441628637,
      "grad_norm": 1.4103500087613847,
      "learning_rate": 6.562006993728504e-07,
      "loss": 2.22,
      "step": 4138
    },
    {
      "epoch": 0.5358449040359905,
      "grad_norm": 1.6810955401953784,
      "learning_rate": 6.559185717171283e-07,
      "loss": 2.436,
      "step": 4139
    },
    {
      "epoch": 0.5359743664433441,
      "grad_norm": 1.6689741274499452,
      "learning_rate": 6.556364377456612e-07,
      "loss": 2.321,
      "step": 4140
    },
    {
      "epoch": 0.5359743664433441,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5423916578292847,
      "eval_runtime": 14.1164,
      "eval_samples_per_second": 3.117,
      "eval_steps_per_second": 0.213,
      "step": 4140
    },
    {
      "epoch": 0.5361038288506975,
      "grad_norm": 2.5132137293247543,
      "learning_rate": 6.553542975160805e-07,
      "loss": 2.4812,
      "step": 4141
    },
    {
      "epoch": 0.536233291258051,
      "grad_norm": 1.6537520547321056,
      "learning_rate": 6.550721510860196e-07,
      "loss": 2.3928,
      "step": 4142
    },
    {
      "epoch": 0.5363627536654044,
      "grad_norm": 4.443183664121167,
      "learning_rate": 6.547899985131122e-07,
      "loss": 2.6509,
      "step": 4143
    },
    {
      "epoch": 0.5364922160727579,
      "grad_norm": 2.591581579111484,
      "learning_rate": 6.545078398549938e-07,
      "loss": 2.4866,
      "step": 4144
    },
    {
      "epoch": 0.5366216784801113,
      "grad_norm": 1.972224143849903,
      "learning_rate": 6.542256751693009e-07,
      "loss": 2.5684,
      "step": 4145
    },
    {
      "epoch": 0.5366216784801113,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.540238857269287,
      "eval_runtime": 14.4653,
      "eval_samples_per_second": 3.042,
      "eval_steps_per_second": 0.207,
      "step": 4145
    },
    {
      "epoch": 0.5367511408874648,
      "grad_norm": 2.434647424660376,
      "learning_rate": 6.539435045136716e-07,
      "loss": 2.4805,
      "step": 4146
    },
    {
      "epoch": 0.5368806032948182,
      "grad_norm": 2.7099396153766784,
      "learning_rate": 6.536613279457449e-07,
      "loss": 2.2415,
      "step": 4147
    },
    {
      "epoch": 0.5370100657021717,
      "grad_norm": 2.5720628154340246,
      "learning_rate": 6.533791455231608e-07,
      "loss": 2.1792,
      "step": 4148
    },
    {
      "epoch": 0.5371395281095251,
      "grad_norm": 1.5318086603967571,
      "learning_rate": 6.530969573035613e-07,
      "loss": 2.3389,
      "step": 4149
    },
    {
      "epoch": 0.5372689905168787,
      "grad_norm": 1.6718272624233645,
      "learning_rate": 6.528147633445885e-07,
      "loss": 2.4729,
      "step": 4150
    },
    {
      "epoch": 0.5372689905168787,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5431240797042847,
      "eval_runtime": 13.9217,
      "eval_samples_per_second": 3.161,
      "eval_steps_per_second": 0.215,
      "step": 4150
    },
    {
      "epoch": 0.5373984529242322,
      "grad_norm": 1.6339644167646443,
      "learning_rate": 6.525325637038867e-07,
      "loss": 2.5388,
      "step": 4151
    },
    {
      "epoch": 0.5375279153315856,
      "grad_norm": 1.699395503065127,
      "learning_rate": 6.522503584391006e-07,
      "loss": 2.5259,
      "step": 4152
    },
    {
      "epoch": 0.5376573777389391,
      "grad_norm": 1.8218315704132828,
      "learning_rate": 6.519681476078765e-07,
      "loss": 2.686,
      "step": 4153
    },
    {
      "epoch": 0.5377868401462925,
      "grad_norm": 2.4566200313116506,
      "learning_rate": 6.516859312678615e-07,
      "loss": 2.3599,
      "step": 4154
    },
    {
      "epoch": 0.537916302553646,
      "grad_norm": 1.6803469253255623,
      "learning_rate": 6.514037094767042e-07,
      "loss": 2.3262,
      "step": 4155
    },
    {
      "epoch": 0.537916302553646,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.539595127105713,
      "eval_runtime": 14.257,
      "eval_samples_per_second": 3.086,
      "eval_steps_per_second": 0.21,
      "step": 4155
    },
    {
      "epoch": 0.5380457649609994,
      "grad_norm": 1.5664799659933828,
      "learning_rate": 6.511214822920542e-07,
      "loss": 2.4006,
      "step": 4156
    },
    {
      "epoch": 0.5381752273683529,
      "grad_norm": 2.8993082389735854,
      "learning_rate": 6.508392497715619e-07,
      "loss": 2.6189,
      "step": 4157
    },
    {
      "epoch": 0.5383046897757063,
      "grad_norm": 1.356183825550859,
      "learning_rate": 6.50557011972879e-07,
      "loss": 2.4287,
      "step": 4158
    },
    {
      "epoch": 0.5384341521830598,
      "grad_norm": 1.3362634663670692,
      "learning_rate": 6.502747689536586e-07,
      "loss": 2.3394,
      "step": 4159
    },
    {
      "epoch": 0.5385636145904134,
      "grad_norm": 1.6283922815925258,
      "learning_rate": 6.49992520771554e-07,
      "loss": 2.4009,
      "step": 4160
    },
    {
      "epoch": 0.5385636145904134,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.542436122894287,
      "eval_runtime": 14.245,
      "eval_samples_per_second": 3.089,
      "eval_steps_per_second": 0.211,
      "step": 4160
    },
    {
      "epoch": 0.5386930769977668,
      "grad_norm": 1.5901340799808281,
      "learning_rate": 6.49710267484221e-07,
      "loss": 2.426,
      "step": 4161
    },
    {
      "epoch": 0.5388225394051203,
      "grad_norm": 1.4668920979501585,
      "learning_rate": 6.494280091493147e-07,
      "loss": 2.4539,
      "step": 4162
    },
    {
      "epoch": 0.5389520018124737,
      "grad_norm": 1.3498471310082636,
      "learning_rate": 6.491457458244921e-07,
      "loss": 2.3096,
      "step": 4163
    },
    {
      "epoch": 0.5390814642198272,
      "grad_norm": 1.761705768284713,
      "learning_rate": 6.488634775674118e-07,
      "loss": 2.3091,
      "step": 4164
    },
    {
      "epoch": 0.5392109266271806,
      "grad_norm": 2.8964785889206657,
      "learning_rate": 6.485812044357326e-07,
      "loss": 3.2476,
      "step": 4165
    },
    {
      "epoch": 0.5392109266271806,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.541748046875,
      "eval_runtime": 13.8564,
      "eval_samples_per_second": 3.175,
      "eval_steps_per_second": 0.217,
      "step": 4165
    },
    {
      "epoch": 0.5393403890345341,
      "grad_norm": 1.522326332860526,
      "learning_rate": 6.482989264871144e-07,
      "loss": 2.4175,
      "step": 4166
    },
    {
      "epoch": 0.5394698514418875,
      "grad_norm": 2.5806067591169173,
      "learning_rate": 6.48016643779218e-07,
      "loss": 2.2136,
      "step": 4167
    },
    {
      "epoch": 0.539599313849241,
      "grad_norm": 1.6801112313729012,
      "learning_rate": 6.477343563697057e-07,
      "loss": 2.5496,
      "step": 4168
    },
    {
      "epoch": 0.5397287762565945,
      "grad_norm": 1.2116867136132727,
      "learning_rate": 6.474520643162403e-07,
      "loss": 2.2166,
      "step": 4169
    },
    {
      "epoch": 0.539858238663948,
      "grad_norm": 3.5642090884438304,
      "learning_rate": 6.471697676764859e-07,
      "loss": 2.9839,
      "step": 4170
    },
    {
      "epoch": 0.539858238663948,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.540771484375,
      "eval_runtime": 13.4534,
      "eval_samples_per_second": 3.271,
      "eval_steps_per_second": 0.223,
      "step": 4170
    },
    {
      "epoch": 0.5399877010713015,
      "grad_norm": 1.9239674828001385,
      "learning_rate": 6.46887466508107e-07,
      "loss": 2.4031,
      "step": 4171
    },
    {
      "epoch": 0.5401171634786549,
      "grad_norm": 3.0916557154110262,
      "learning_rate": 6.466051608687691e-07,
      "loss": 2.3848,
      "step": 4172
    },
    {
      "epoch": 0.5402466258860084,
      "grad_norm": 2.224857846619929,
      "learning_rate": 6.463228508161396e-07,
      "loss": 2.8987,
      "step": 4173
    },
    {
      "epoch": 0.5403760882933618,
      "grad_norm": 1.4484856016602865,
      "learning_rate": 6.460405364078853e-07,
      "loss": 2.3452,
      "step": 4174
    },
    {
      "epoch": 0.5405055507007153,
      "grad_norm": 3.5544299907319408,
      "learning_rate": 6.457582177016753e-07,
      "loss": 2.6343,
      "step": 4175
    },
    {
      "epoch": 0.5405055507007153,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5425026416778564,
      "eval_runtime": 14.1343,
      "eval_samples_per_second": 3.113,
      "eval_steps_per_second": 0.212,
      "step": 4175
    },
    {
      "epoch": 0.5406350131080687,
      "grad_norm": 2.4112992276312366,
      "learning_rate": 6.454758947551783e-07,
      "loss": 2.7356,
      "step": 4176
    },
    {
      "epoch": 0.5407644755154222,
      "grad_norm": 1.174249291916097,
      "learning_rate": 6.451935676260646e-07,
      "loss": 2.3081,
      "step": 4177
    },
    {
      "epoch": 0.5408939379227756,
      "grad_norm": 3.502429903794039,
      "learning_rate": 6.449112363720059e-07,
      "loss": 2.5579,
      "step": 4178
    },
    {
      "epoch": 0.5410234003301292,
      "grad_norm": 1.996221574938458,
      "learning_rate": 6.446289010506733e-07,
      "loss": 2.5537,
      "step": 4179
    },
    {
      "epoch": 0.5411528627374826,
      "grad_norm": 1.6937322360129383,
      "learning_rate": 6.443465617197399e-07,
      "loss": 2.2275,
      "step": 4180
    },
    {
      "epoch": 0.5411528627374826,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5429909229278564,
      "eval_runtime": 14.6676,
      "eval_samples_per_second": 3.0,
      "eval_steps_per_second": 0.205,
      "step": 4180
    },
    {
      "epoch": 0.5412823251448361,
      "grad_norm": 1.7482707874539745,
      "learning_rate": 6.440642184368791e-07,
      "loss": 2.4971,
      "step": 4181
    },
    {
      "epoch": 0.5414117875521895,
      "grad_norm": 2.371147087751884,
      "learning_rate": 6.437818712597652e-07,
      "loss": 2.2202,
      "step": 4182
    },
    {
      "epoch": 0.541541249959543,
      "grad_norm": 1.2245064008136306,
      "learning_rate": 6.434995202460734e-07,
      "loss": 2.3198,
      "step": 4183
    },
    {
      "epoch": 0.5416707123668965,
      "grad_norm": 1.4041473304260155,
      "learning_rate": 6.432171654534798e-07,
      "loss": 2.4062,
      "step": 4184
    },
    {
      "epoch": 0.5418001747742499,
      "grad_norm": 1.953007238042008,
      "learning_rate": 6.429348069396608e-07,
      "loss": 2.3882,
      "step": 4185
    },
    {
      "epoch": 0.5418001747742499,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5472744703292847,
      "eval_runtime": 15.1198,
      "eval_samples_per_second": 2.91,
      "eval_steps_per_second": 0.198,
      "step": 4185
    },
    {
      "epoch": 0.5419296371816034,
      "grad_norm": 1.6887981284720208,
      "learning_rate": 6.426524447622939e-07,
      "loss": 2.2783,
      "step": 4186
    },
    {
      "epoch": 0.5420590995889568,
      "grad_norm": 1.4744105133337015,
      "learning_rate": 6.423700789790574e-07,
      "loss": 2.4482,
      "step": 4187
    },
    {
      "epoch": 0.5421885619963103,
      "grad_norm": 1.0822522146984528,
      "learning_rate": 6.420877096476302e-07,
      "loss": 2.1797,
      "step": 4188
    },
    {
      "epoch": 0.5423180244036638,
      "grad_norm": 2.441575103806913,
      "learning_rate": 6.418053368256921e-07,
      "loss": 2.8159,
      "step": 4189
    },
    {
      "epoch": 0.5424474868110173,
      "grad_norm": 1.4886098520316235,
      "learning_rate": 6.41522960570923e-07,
      "loss": 2.3335,
      "step": 4190
    },
    {
      "epoch": 0.5424474868110173,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5444557666778564,
      "eval_runtime": 13.9678,
      "eval_samples_per_second": 3.15,
      "eval_steps_per_second": 0.215,
      "step": 4190
    },
    {
      "epoch": 0.5425769492183707,
      "grad_norm": 1.9406476922778426,
      "learning_rate": 6.412405809410044e-07,
      "loss": 2.626,
      "step": 4191
    },
    {
      "epoch": 0.5427064116257242,
      "grad_norm": 2.3393791085247666,
      "learning_rate": 6.409581979936182e-07,
      "loss": 2.5432,
      "step": 4192
    },
    {
      "epoch": 0.5428358740330776,
      "grad_norm": 2.2694496386550984,
      "learning_rate": 6.406758117864465e-07,
      "loss": 2.2891,
      "step": 4193
    },
    {
      "epoch": 0.5429653364404311,
      "grad_norm": 1.6616169648274188,
      "learning_rate": 6.403934223771723e-07,
      "loss": 2.1733,
      "step": 4194
    },
    {
      "epoch": 0.5430947988477846,
      "grad_norm": 1.7415433677808254,
      "learning_rate": 6.401110298234797e-07,
      "loss": 2.4353,
      "step": 4195
    },
    {
      "epoch": 0.5430947988477846,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.546586513519287,
      "eval_runtime": 14.4774,
      "eval_samples_per_second": 3.039,
      "eval_steps_per_second": 0.207,
      "step": 4195
    },
    {
      "epoch": 0.543224261255138,
      "grad_norm": 2.1996264029862456,
      "learning_rate": 6.39828634183053e-07,
      "loss": 2.5642,
      "step": 4196
    },
    {
      "epoch": 0.5433537236624915,
      "grad_norm": 2.8892846477758334,
      "learning_rate": 6.395462355135771e-07,
      "loss": 2.3782,
      "step": 4197
    },
    {
      "epoch": 0.5434831860698449,
      "grad_norm": 2.087094379000283,
      "learning_rate": 6.39263833872738e-07,
      "loss": 2.4329,
      "step": 4198
    },
    {
      "epoch": 0.5436126484771985,
      "grad_norm": 1.3318643034196034,
      "learning_rate": 6.389814293182217e-07,
      "loss": 2.4294,
      "step": 4199
    },
    {
      "epoch": 0.5437421108845519,
      "grad_norm": 1.1322866633949251,
      "learning_rate": 6.386990219077149e-07,
      "loss": 2.1982,
      "step": 4200
    },
    {
      "epoch": 0.5437421108845519,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5423694849014282,
      "eval_runtime": 14.232,
      "eval_samples_per_second": 3.092,
      "eval_steps_per_second": 0.211,
      "step": 4200
    },
    {
      "epoch": 0.5438715732919054,
      "grad_norm": 2.5152851648724677,
      "learning_rate": 6.384166116989052e-07,
      "loss": 2.447,
      "step": 4201
    },
    {
      "epoch": 0.5440010356992588,
      "grad_norm": 2.620407039986505,
      "learning_rate": 6.381341987494806e-07,
      "loss": 2.3965,
      "step": 4202
    },
    {
      "epoch": 0.5441304981066123,
      "grad_norm": 2.748626575249578,
      "learning_rate": 6.378517831171298e-07,
      "loss": 2.5415,
      "step": 4203
    },
    {
      "epoch": 0.5442599605139657,
      "grad_norm": 4.251181066911142,
      "learning_rate": 6.375693648595417e-07,
      "loss": 2.6841,
      "step": 4204
    },
    {
      "epoch": 0.5443894229213192,
      "grad_norm": 1.189665140244015,
      "learning_rate": 6.37286944034406e-07,
      "loss": 2.3726,
      "step": 4205
    },
    {
      "epoch": 0.5443894229213192,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5437678098678589,
      "eval_runtime": 14.3562,
      "eval_samples_per_second": 3.065,
      "eval_steps_per_second": 0.209,
      "step": 4205
    },
    {
      "epoch": 0.5445188853286727,
      "grad_norm": 2.076593322638028,
      "learning_rate": 6.370045206994133e-07,
      "loss": 2.4871,
      "step": 4206
    },
    {
      "epoch": 0.5446483477360261,
      "grad_norm": 1.4388034624808688,
      "learning_rate": 6.367220949122536e-07,
      "loss": 2.3867,
      "step": 4207
    },
    {
      "epoch": 0.5447778101433797,
      "grad_norm": 2.493220957586294,
      "learning_rate": 6.364396667306186e-07,
      "loss": 2.6929,
      "step": 4208
    },
    {
      "epoch": 0.5449072725507331,
      "grad_norm": 3.426691707669274,
      "learning_rate": 6.361572362121997e-07,
      "loss": 2.447,
      "step": 4209
    },
    {
      "epoch": 0.5450367349580866,
      "grad_norm": 1.4371328325093933,
      "learning_rate": 6.35874803414689e-07,
      "loss": 2.3025,
      "step": 4210
    },
    {
      "epoch": 0.5450367349580866,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5452326536178589,
      "eval_runtime": 14.5668,
      "eval_samples_per_second": 3.021,
      "eval_steps_per_second": 0.206,
      "step": 4210
    },
    {
      "epoch": 0.54516619736544,
      "grad_norm": 1.6803092972037372,
      "learning_rate": 6.355923683957796e-07,
      "loss": 2.4741,
      "step": 4211
    },
    {
      "epoch": 0.5452956597727935,
      "grad_norm": 1.5433213481873256,
      "learning_rate": 6.353099312131641e-07,
      "loss": 2.47,
      "step": 4212
    },
    {
      "epoch": 0.5454251221801469,
      "grad_norm": 1.6174433267701191,
      "learning_rate": 6.35027491924536e-07,
      "loss": 2.5542,
      "step": 4213
    },
    {
      "epoch": 0.5455545845875004,
      "grad_norm": 2.8481688644180476,
      "learning_rate": 6.347450505875898e-07,
      "loss": 2.5293,
      "step": 4214
    },
    {
      "epoch": 0.5456840469948538,
      "grad_norm": 1.408684476527369,
      "learning_rate": 6.344626072600192e-07,
      "loss": 2.2739,
      "step": 4215
    },
    {
      "epoch": 0.5456840469948538,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5416592359542847,
      "eval_runtime": 14.4188,
      "eval_samples_per_second": 3.052,
      "eval_steps_per_second": 0.208,
      "step": 4215
    },
    {
      "epoch": 0.5458135094022073,
      "grad_norm": 1.4355391005547333,
      "learning_rate": 6.341801619995192e-07,
      "loss": 2.1799,
      "step": 4216
    },
    {
      "epoch": 0.5459429718095608,
      "grad_norm": 1.6207047158663324,
      "learning_rate": 6.338977148637854e-07,
      "loss": 2.459,
      "step": 4217
    },
    {
      "epoch": 0.5460724342169143,
      "grad_norm": 1.9303170020412486,
      "learning_rate": 6.336152659105127e-07,
      "loss": 2.5396,
      "step": 4218
    },
    {
      "epoch": 0.5462018966242678,
      "grad_norm": 2.3186850222738093,
      "learning_rate": 6.333328151973974e-07,
      "loss": 2.3923,
      "step": 4219
    },
    {
      "epoch": 0.5463313590316212,
      "grad_norm": 2.0628762619456857,
      "learning_rate": 6.330503627821358e-07,
      "loss": 2.7161,
      "step": 4220
    },
    {
      "epoch": 0.5463313590316212,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5359108448028564,
      "eval_runtime": 13.5612,
      "eval_samples_per_second": 3.245,
      "eval_steps_per_second": 0.221,
      "step": 4220
    },
    {
      "epoch": 0.5464608214389747,
      "grad_norm": 1.7857025300073248,
      "learning_rate": 6.327679087224244e-07,
      "loss": 2.5642,
      "step": 4221
    },
    {
      "epoch": 0.5465902838463281,
      "grad_norm": 1.8159878219151797,
      "learning_rate": 6.324854530759598e-07,
      "loss": 2.5762,
      "step": 4222
    },
    {
      "epoch": 0.5467197462536816,
      "grad_norm": 1.0916356142439672,
      "learning_rate": 6.322029959004403e-07,
      "loss": 2.1165,
      "step": 4223
    },
    {
      "epoch": 0.546849208661035,
      "grad_norm": 1.6807476770558634,
      "learning_rate": 6.319205372535625e-07,
      "loss": 2.4902,
      "step": 4224
    },
    {
      "epoch": 0.5469786710683885,
      "grad_norm": 1.296506550576479,
      "learning_rate": 6.316380771930249e-07,
      "loss": 2.2305,
      "step": 4225
    },
    {
      "epoch": 0.5469786710683885,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5417258739471436,
      "eval_runtime": 14.1958,
      "eval_samples_per_second": 3.1,
      "eval_steps_per_second": 0.211,
      "step": 4225
    },
    {
      "epoch": 0.5471081334757419,
      "grad_norm": 4.201644742027217,
      "learning_rate": 6.313556157765252e-07,
      "loss": 2.5933,
      "step": 4226
    },
    {
      "epoch": 0.5472375958830954,
      "grad_norm": 2.0314146327915976,
      "learning_rate": 6.310731530617624e-07,
      "loss": 2.5894,
      "step": 4227
    },
    {
      "epoch": 0.547367058290449,
      "grad_norm": 1.8474975378034588,
      "learning_rate": 6.307906891064349e-07,
      "loss": 2.3545,
      "step": 4228
    },
    {
      "epoch": 0.5474965206978024,
      "grad_norm": 1.6345361699498737,
      "learning_rate": 6.305082239682418e-07,
      "loss": 2.5654,
      "step": 4229
    },
    {
      "epoch": 0.5476259831051559,
      "grad_norm": 2.105173565976692,
      "learning_rate": 6.302257577048824e-07,
      "loss": 2.5134,
      "step": 4230
    },
    {
      "epoch": 0.5476259831051559,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5423251390457153,
      "eval_runtime": 14.2269,
      "eval_samples_per_second": 3.093,
      "eval_steps_per_second": 0.211,
      "step": 4230
    },
    {
      "epoch": 0.5477554455125093,
      "grad_norm": 1.48930641554959,
      "learning_rate": 6.299432903740559e-07,
      "loss": 2.4819,
      "step": 4231
    },
    {
      "epoch": 0.5478849079198628,
      "grad_norm": 1.2684982608570394,
      "learning_rate": 6.296608220334623e-07,
      "loss": 2.3135,
      "step": 4232
    },
    {
      "epoch": 0.5480143703272162,
      "grad_norm": 1.4004164751858577,
      "learning_rate": 6.293783527408011e-07,
      "loss": 2.3877,
      "step": 4233
    },
    {
      "epoch": 0.5481438327345697,
      "grad_norm": 2.467489378204922,
      "learning_rate": 6.290958825537729e-07,
      "loss": 2.5466,
      "step": 4234
    },
    {
      "epoch": 0.5482732951419231,
      "grad_norm": 2.37339525455699,
      "learning_rate": 6.288134115300776e-07,
      "loss": 2.5435,
      "step": 4235
    },
    {
      "epoch": 0.5482732951419231,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.541703701019287,
      "eval_runtime": 14.5334,
      "eval_samples_per_second": 3.028,
      "eval_steps_per_second": 0.206,
      "step": 4235
    },
    {
      "epoch": 0.5484027575492766,
      "grad_norm": 1.5220531485817708,
      "learning_rate": 6.285309397274156e-07,
      "loss": 2.3633,
      "step": 4236
    },
    {
      "epoch": 0.54853221995663,
      "grad_norm": 1.8425814639014737,
      "learning_rate": 6.282484672034875e-07,
      "loss": 2.4744,
      "step": 4237
    },
    {
      "epoch": 0.5486616823639836,
      "grad_norm": 1.8795602805475373,
      "learning_rate": 6.279659940159944e-07,
      "loss": 2.3457,
      "step": 4238
    },
    {
      "epoch": 0.5487911447713371,
      "grad_norm": 3.2850836472195053,
      "learning_rate": 6.276835202226368e-07,
      "loss": 2.6023,
      "step": 4239
    },
    {
      "epoch": 0.5489206071786905,
      "grad_norm": 1.9113612809426985,
      "learning_rate": 6.274010458811155e-07,
      "loss": 2.3223,
      "step": 4240
    },
    {
      "epoch": 0.5489206071786905,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.543257236480713,
      "eval_runtime": 14.1931,
      "eval_samples_per_second": 3.1,
      "eval_steps_per_second": 0.211,
      "step": 4240
    },
    {
      "epoch": 0.549050069586044,
      "grad_norm": 1.3494014291103476,
      "learning_rate": 6.271185710491322e-07,
      "loss": 2.405,
      "step": 4241
    },
    {
      "epoch": 0.5491795319933974,
      "grad_norm": 1.619297404680498,
      "learning_rate": 6.268360957843879e-07,
      "loss": 2.3254,
      "step": 4242
    },
    {
      "epoch": 0.5493089944007509,
      "grad_norm": 2.5865667349566017,
      "learning_rate": 6.265536201445837e-07,
      "loss": 2.3953,
      "step": 4243
    },
    {
      "epoch": 0.5494384568081043,
      "grad_norm": 1.891944822875691,
      "learning_rate": 6.262711441874212e-07,
      "loss": 2.4565,
      "step": 4244
    },
    {
      "epoch": 0.5495679192154578,
      "grad_norm": 1.3004415238380485,
      "learning_rate": 6.259886679706015e-07,
      "loss": 2.3877,
      "step": 4245
    },
    {
      "epoch": 0.5495679192154578,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.545210361480713,
      "eval_runtime": 15.0134,
      "eval_samples_per_second": 2.931,
      "eval_steps_per_second": 0.2,
      "step": 4245
    },
    {
      "epoch": 0.5496973816228112,
      "grad_norm": 1.776819784881242,
      "learning_rate": 6.257061915518264e-07,
      "loss": 2.302,
      "step": 4246
    },
    {
      "epoch": 0.5498268440301648,
      "grad_norm": 2.280689985617081,
      "learning_rate": 6.254237149887975e-07,
      "loss": 2.4583,
      "step": 4247
    },
    {
      "epoch": 0.5499563064375182,
      "grad_norm": 2.7203691094558065,
      "learning_rate": 6.251412383392162e-07,
      "loss": 2.5894,
      "step": 4248
    },
    {
      "epoch": 0.5500857688448717,
      "grad_norm": 2.969062585147217,
      "learning_rate": 6.248587616607841e-07,
      "loss": 2.4214,
      "step": 4249
    },
    {
      "epoch": 0.5502152312522252,
      "grad_norm": 1.296427052441223,
      "learning_rate": 6.245762850112027e-07,
      "loss": 2.1907,
      "step": 4250
    },
    {
      "epoch": 0.5502152312522252,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5416370630264282,
      "eval_runtime": 13.8872,
      "eval_samples_per_second": 3.168,
      "eval_steps_per_second": 0.216,
      "step": 4250
    },
    {
      "epoch": 0.5503446936595786,
      "grad_norm": 1.402864582983678,
      "learning_rate": 6.242938084481737e-07,
      "loss": 2.3792,
      "step": 4251
    },
    {
      "epoch": 0.5504741560669321,
      "grad_norm": 1.3751387286957826,
      "learning_rate": 6.240113320293987e-07,
      "loss": 2.3184,
      "step": 4252
    },
    {
      "epoch": 0.5506036184742855,
      "grad_norm": 2.1869111357667874,
      "learning_rate": 6.237288558125792e-07,
      "loss": 2.3779,
      "step": 4253
    },
    {
      "epoch": 0.550733080881639,
      "grad_norm": 1.5283618607460008,
      "learning_rate": 6.234463798554164e-07,
      "loss": 2.1719,
      "step": 4254
    },
    {
      "epoch": 0.5508625432889924,
      "grad_norm": 1.4464525499124974,
      "learning_rate": 6.231639042156124e-07,
      "loss": 2.2969,
      "step": 4255
    },
    {
      "epoch": 0.5508625432889924,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5423916578292847,
      "eval_runtime": 14.0399,
      "eval_samples_per_second": 3.134,
      "eval_steps_per_second": 0.214,
      "step": 4255
    },
    {
      "epoch": 0.5509920056963459,
      "grad_norm": 1.5550531398759564,
      "learning_rate": 6.22881428950868e-07,
      "loss": 2.3955,
      "step": 4256
    },
    {
      "epoch": 0.5511214681036994,
      "grad_norm": 2.0309621929565633,
      "learning_rate": 6.225989541188847e-07,
      "loss": 2.4414,
      "step": 4257
    },
    {
      "epoch": 0.5512509305110529,
      "grad_norm": 1.5501826796899372,
      "learning_rate": 6.223164797773634e-07,
      "loss": 2.3711,
      "step": 4258
    },
    {
      "epoch": 0.5513803929184063,
      "grad_norm": 2.692452107080833,
      "learning_rate": 6.220340059840057e-07,
      "loss": 2.4541,
      "step": 4259
    },
    {
      "epoch": 0.5515098553257598,
      "grad_norm": 1.963313779702523,
      "learning_rate": 6.217515327965128e-07,
      "loss": 2.6089,
      "step": 4260
    },
    {
      "epoch": 0.5515098553257598,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5409489870071411,
      "eval_runtime": 14.1201,
      "eval_samples_per_second": 3.116,
      "eval_steps_per_second": 0.212,
      "step": 4260
    },
    {
      "epoch": 0.5516393177331133,
      "grad_norm": 1.050193208919441,
      "learning_rate": 6.214690602725848e-07,
      "loss": 2.2222,
      "step": 4261
    },
    {
      "epoch": 0.5517687801404667,
      "grad_norm": 1.3499979398460218,
      "learning_rate": 6.211865884699227e-07,
      "loss": 2.3208,
      "step": 4262
    },
    {
      "epoch": 0.5518982425478202,
      "grad_norm": 1.5321107954241715,
      "learning_rate": 6.209041174462273e-07,
      "loss": 2.3252,
      "step": 4263
    },
    {
      "epoch": 0.5520277049551736,
      "grad_norm": 2.990441835117077,
      "learning_rate": 6.206216472591989e-07,
      "loss": 2.4578,
      "step": 4264
    },
    {
      "epoch": 0.5521571673625271,
      "grad_norm": 1.559652379821359,
      "learning_rate": 6.203391779665378e-07,
      "loss": 2.3167,
      "step": 4265
    },
    {
      "epoch": 0.5521571673625271,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5409934520721436,
      "eval_runtime": 13.6904,
      "eval_samples_per_second": 3.214,
      "eval_steps_per_second": 0.219,
      "step": 4265
    },
    {
      "epoch": 0.5522866297698805,
      "grad_norm": 1.8119123071096452,
      "learning_rate": 6.200567096259442e-07,
      "loss": 2.5085,
      "step": 4266
    },
    {
      "epoch": 0.5524160921772341,
      "grad_norm": 1.5331208358071444,
      "learning_rate": 6.197742422951179e-07,
      "loss": 2.3789,
      "step": 4267
    },
    {
      "epoch": 0.5525455545845875,
      "grad_norm": 3.4507193190543246,
      "learning_rate": 6.194917760317583e-07,
      "loss": 2.6619,
      "step": 4268
    },
    {
      "epoch": 0.552675016991941,
      "grad_norm": 1.7432181431325713,
      "learning_rate": 6.192093108935652e-07,
      "loss": 2.5391,
      "step": 4269
    },
    {
      "epoch": 0.5528044793992944,
      "grad_norm": 1.0047514710300003,
      "learning_rate": 6.189268469382377e-07,
      "loss": 2.1633,
      "step": 4270
    },
    {
      "epoch": 0.5528044793992944,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5408824682235718,
      "eval_runtime": 14.2773,
      "eval_samples_per_second": 3.082,
      "eval_steps_per_second": 0.21,
      "step": 4270
    },
    {
      "epoch": 0.5529339418066479,
      "grad_norm": 2.847447302301485,
      "learning_rate": 6.18644384223475e-07,
      "loss": 2.4878,
      "step": 4271
    },
    {
      "epoch": 0.5530634042140014,
      "grad_norm": 3.564824162049398,
      "learning_rate": 6.183619228069754e-07,
      "loss": 2.5781,
      "step": 4272
    },
    {
      "epoch": 0.5531928666213548,
      "grad_norm": 2.190422739362663,
      "learning_rate": 6.180794627464376e-07,
      "loss": 2.5781,
      "step": 4273
    },
    {
      "epoch": 0.5533223290287083,
      "grad_norm": 3.335455848514011,
      "learning_rate": 6.1779700409956e-07,
      "loss": 2.6113,
      "step": 4274
    },
    {
      "epoch": 0.5534517914360617,
      "grad_norm": 3.240592517370985,
      "learning_rate": 6.175145469240402e-07,
      "loss": 2.437,
      "step": 4275
    },
    {
      "epoch": 0.5534517914360617,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5424582958221436,
      "eval_runtime": 14.1941,
      "eval_samples_per_second": 3.1,
      "eval_steps_per_second": 0.211,
      "step": 4275
    },
    {
      "epoch": 0.5535812538434152,
      "grad_norm": 1.1368239219851894,
      "learning_rate": 6.172320912775759e-07,
      "loss": 2.3291,
      "step": 4276
    },
    {
      "epoch": 0.5537107162507687,
      "grad_norm": 1.6462575268764186,
      "learning_rate": 6.169496372178645e-07,
      "loss": 2.5496,
      "step": 4277
    },
    {
      "epoch": 0.5538401786581222,
      "grad_norm": 2.7383191619532736,
      "learning_rate": 6.166671848026027e-07,
      "loss": 2.5112,
      "step": 4278
    },
    {
      "epoch": 0.5539696410654756,
      "grad_norm": 1.3810044211872197,
      "learning_rate": 6.163847340894875e-07,
      "loss": 2.3259,
      "step": 4279
    },
    {
      "epoch": 0.5540991034728291,
      "grad_norm": 1.3782291577102648,
      "learning_rate": 6.161022851362149e-07,
      "loss": 2.3877,
      "step": 4280
    },
    {
      "epoch": 0.5540991034728291,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5381081104278564,
      "eval_runtime": 13.7889,
      "eval_samples_per_second": 3.191,
      "eval_steps_per_second": 0.218,
      "step": 4280
    },
    {
      "epoch": 0.5542285658801825,
      "grad_norm": 1.6345559226615984,
      "learning_rate": 6.158198380004807e-07,
      "loss": 2.3025,
      "step": 4281
    },
    {
      "epoch": 0.554358028287536,
      "grad_norm": 2.134404514212594,
      "learning_rate": 6.155373927399811e-07,
      "loss": 2.6165,
      "step": 4282
    },
    {
      "epoch": 0.5544874906948895,
      "grad_norm": 1.8754157751704479,
      "learning_rate": 6.152549494124105e-07,
      "loss": 2.4297,
      "step": 4283
    },
    {
      "epoch": 0.5546169531022429,
      "grad_norm": 1.4589409114717964,
      "learning_rate": 6.149725080754641e-07,
      "loss": 2.293,
      "step": 4284
    },
    {
      "epoch": 0.5547464155095964,
      "grad_norm": 2.3647478201961722,
      "learning_rate": 6.146900687868361e-07,
      "loss": 2.3967,
      "step": 4285
    },
    {
      "epoch": 0.5547464155095964,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5380192995071411,
      "eval_runtime": 14.1434,
      "eval_samples_per_second": 3.111,
      "eval_steps_per_second": 0.212,
      "step": 4285
    },
    {
      "epoch": 0.5548758779169498,
      "grad_norm": 1.8037358080536374,
      "learning_rate": 6.144076316042206e-07,
      "loss": 2.5027,
      "step": 4286
    },
    {
      "epoch": 0.5550053403243034,
      "grad_norm": 2.010237073853137,
      "learning_rate": 6.14125196585311e-07,
      "loss": 2.3926,
      "step": 4287
    },
    {
      "epoch": 0.5551348027316568,
      "grad_norm": 2.028880778931612,
      "learning_rate": 6.138427637878007e-07,
      "loss": 2.2804,
      "step": 4288
    },
    {
      "epoch": 0.5552642651390103,
      "grad_norm": 1.6381121289506992,
      "learning_rate": 6.135603332693816e-07,
      "loss": 2.3596,
      "step": 4289
    },
    {
      "epoch": 0.5553937275463637,
      "grad_norm": 1.155760356910515,
      "learning_rate": 6.132779050877465e-07,
      "loss": 2.2751,
      "step": 4290
    },
    {
      "epoch": 0.5553937275463637,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5388849973678589,
      "eval_runtime": 14.3499,
      "eval_samples_per_second": 3.066,
      "eval_steps_per_second": 0.209,
      "step": 4290
    },
    {
      "epoch": 0.5555231899537172,
      "grad_norm": 1.4974920409708508,
      "learning_rate": 6.129954793005868e-07,
      "loss": 2.4294,
      "step": 4291
    },
    {
      "epoch": 0.5556526523610706,
      "grad_norm": 1.6028331554743795,
      "learning_rate": 6.127130559655938e-07,
      "loss": 2.47,
      "step": 4292
    },
    {
      "epoch": 0.5557821147684241,
      "grad_norm": 2.846220988905065,
      "learning_rate": 6.124306351404583e-07,
      "loss": 2.5627,
      "step": 4293
    },
    {
      "epoch": 0.5559115771757775,
      "grad_norm": 1.5526617947905548,
      "learning_rate": 6.121482168828704e-07,
      "loss": 2.2983,
      "step": 4294
    },
    {
      "epoch": 0.556041039583131,
      "grad_norm": 2.3601307970385452,
      "learning_rate": 6.118658012505195e-07,
      "loss": 2.4534,
      "step": 4295
    },
    {
      "epoch": 0.556041039583131,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5347567796707153,
      "eval_runtime": 14.4575,
      "eval_samples_per_second": 3.043,
      "eval_steps_per_second": 0.208,
      "step": 4295
    },
    {
      "epoch": 0.5561705019904846,
      "grad_norm": 1.0591689115265042,
      "learning_rate": 6.11583388301095e-07,
      "loss": 2.1467,
      "step": 4296
    },
    {
      "epoch": 0.556299964397838,
      "grad_norm": 3.093537640550959,
      "learning_rate": 6.113009780922853e-07,
      "loss": 2.8789,
      "step": 4297
    },
    {
      "epoch": 0.5564294268051915,
      "grad_norm": 1.4199783428734807,
      "learning_rate": 6.110185706817785e-07,
      "loss": 2.5833,
      "step": 4298
    },
    {
      "epoch": 0.5565588892125449,
      "grad_norm": 1.7854911614667743,
      "learning_rate": 6.107361661272622e-07,
      "loss": 2.6311,
      "step": 4299
    },
    {
      "epoch": 0.5566883516198984,
      "grad_norm": 1.692967792531868,
      "learning_rate": 6.10453764486423e-07,
      "loss": 2.4094,
      "step": 4300
    },
    {
      "epoch": 0.5566883516198984,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.535888671875,
      "eval_runtime": 13.742,
      "eval_samples_per_second": 3.202,
      "eval_steps_per_second": 0.218,
      "step": 4300
    },
    {
      "epoch": 0.5568178140272518,
      "grad_norm": 1.621198340454262,
      "learning_rate": 6.101713658169471e-07,
      "loss": 2.5679,
      "step": 4301
    },
    {
      "epoch": 0.5569472764346053,
      "grad_norm": 1.7649924775661157,
      "learning_rate": 6.098889701765204e-07,
      "loss": 2.3979,
      "step": 4302
    },
    {
      "epoch": 0.5570767388419587,
      "grad_norm": 1.9243137756792954,
      "learning_rate": 6.096065776228279e-07,
      "loss": 2.2397,
      "step": 4303
    },
    {
      "epoch": 0.5572062012493122,
      "grad_norm": 1.1067917529171238,
      "learning_rate": 6.093241882135536e-07,
      "loss": 2.2673,
      "step": 4304
    },
    {
      "epoch": 0.5573356636566656,
      "grad_norm": 1.8037687154397128,
      "learning_rate": 6.090418020063821e-07,
      "loss": 2.3108,
      "step": 4305
    },
    {
      "epoch": 0.5573356636566656,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5361106395721436,
      "eval_runtime": 13.5547,
      "eval_samples_per_second": 3.246,
      "eval_steps_per_second": 0.221,
      "step": 4305
    },
    {
      "epoch": 0.5574651260640192,
      "grad_norm": 1.6941644609536064,
      "learning_rate": 6.087594190589957e-07,
      "loss": 2.4482,
      "step": 4306
    },
    {
      "epoch": 0.5575945884713727,
      "grad_norm": 1.6380269526534845,
      "learning_rate": 6.084770394290772e-07,
      "loss": 2.3022,
      "step": 4307
    },
    {
      "epoch": 0.5577240508787261,
      "grad_norm": 2.0084433010521,
      "learning_rate": 6.081946631743081e-07,
      "loss": 2.6982,
      "step": 4308
    },
    {
      "epoch": 0.5578535132860796,
      "grad_norm": 2.01592731315132,
      "learning_rate": 6.079122903523699e-07,
      "loss": 2.4358,
      "step": 4309
    },
    {
      "epoch": 0.557982975693433,
      "grad_norm": 1.0106757896882663,
      "learning_rate": 6.076299210209429e-07,
      "loss": 2.167,
      "step": 4310
    },
    {
      "epoch": 0.557982975693433,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5338467359542847,
      "eval_runtime": 13.6083,
      "eval_samples_per_second": 3.233,
      "eval_steps_per_second": 0.22,
      "step": 4310
    },
    {
      "epoch": 0.5581124381007865,
      "grad_norm": 1.280734552829846,
      "learning_rate": 6.073475552377065e-07,
      "loss": 2.4146,
      "step": 4311
    },
    {
      "epoch": 0.5582419005081399,
      "grad_norm": 1.670267584144987,
      "learning_rate": 6.070651930603394e-07,
      "loss": 2.2913,
      "step": 4312
    },
    {
      "epoch": 0.5583713629154934,
      "grad_norm": 1.7265189822582097,
      "learning_rate": 6.067828345465203e-07,
      "loss": 2.3843,
      "step": 4313
    },
    {
      "epoch": 0.5585008253228468,
      "grad_norm": 2.002063738470711,
      "learning_rate": 6.065004797539266e-07,
      "loss": 2.3066,
      "step": 4314
    },
    {
      "epoch": 0.5586302877302003,
      "grad_norm": 1.9168009317211723,
      "learning_rate": 6.062181287402349e-07,
      "loss": 2.5767,
      "step": 4315
    },
    {
      "epoch": 0.5586302877302003,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5359553098678589,
      "eval_runtime": 14.0475,
      "eval_samples_per_second": 3.132,
      "eval_steps_per_second": 0.214,
      "step": 4315
    },
    {
      "epoch": 0.5587597501375539,
      "grad_norm": 1.403384796059386,
      "learning_rate": 6.059357815631211e-07,
      "loss": 2.207,
      "step": 4316
    },
    {
      "epoch": 0.5588892125449073,
      "grad_norm": 4.448100370827739,
      "learning_rate": 6.056534382802604e-07,
      "loss": 2.75,
      "step": 4317
    },
    {
      "epoch": 0.5590186749522608,
      "grad_norm": 2.121133381597139,
      "learning_rate": 6.053710989493268e-07,
      "loss": 2.4905,
      "step": 4318
    },
    {
      "epoch": 0.5591481373596142,
      "grad_norm": 6.075048779560015,
      "learning_rate": 6.050887636279943e-07,
      "loss": 3.207,
      "step": 4319
    },
    {
      "epoch": 0.5592775997669677,
      "grad_norm": 1.8484663206337315,
      "learning_rate": 6.048064323739353e-07,
      "loss": 2.4114,
      "step": 4320
    },
    {
      "epoch": 0.5592775997669677,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.533891201019287,
      "eval_runtime": 14.5707,
      "eval_samples_per_second": 3.02,
      "eval_steps_per_second": 0.206,
      "step": 4320
    },
    {
      "epoch": 0.5594070621743211,
      "grad_norm": 4.268453327901289,
      "learning_rate": 6.04524105244822e-07,
      "loss": 2.7529,
      "step": 4321
    },
    {
      "epoch": 0.5595365245816746,
      "grad_norm": 2.1370555224338714,
      "learning_rate": 6.04241782298325e-07,
      "loss": 2.3604,
      "step": 4322
    },
    {
      "epoch": 0.559665986989028,
      "grad_norm": 2.2960287423991486,
      "learning_rate": 6.039594635921149e-07,
      "loss": 2.3677,
      "step": 4323
    },
    {
      "epoch": 0.5597954493963815,
      "grad_norm": 2.8704269197703756,
      "learning_rate": 6.036771491838607e-07,
      "loss": 2.4443,
      "step": 4324
    },
    {
      "epoch": 0.5599249118037349,
      "grad_norm": 1.4825200247912658,
      "learning_rate": 6.03394839131231e-07,
      "loss": 2.5254,
      "step": 4325
    },
    {
      "epoch": 0.5599249118037349,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5359996557235718,
      "eval_runtime": 14.266,
      "eval_samples_per_second": 3.084,
      "eval_steps_per_second": 0.21,
      "step": 4325
    },
    {
      "epoch": 0.5600543742110885,
      "grad_norm": 1.442340105331669,
      "learning_rate": 6.031125334918933e-07,
      "loss": 2.2185,
      "step": 4326
    },
    {
      "epoch": 0.560183836618442,
      "grad_norm": 1.2449860375339152,
      "learning_rate": 6.028302323235144e-07,
      "loss": 2.3076,
      "step": 4327
    },
    {
      "epoch": 0.5603132990257954,
      "grad_norm": 1.7380514771513373,
      "learning_rate": 6.025479356837598e-07,
      "loss": 2.1694,
      "step": 4328
    },
    {
      "epoch": 0.5604427614331489,
      "grad_norm": 2.657681475032726,
      "learning_rate": 6.022656436302944e-07,
      "loss": 2.8411,
      "step": 4329
    },
    {
      "epoch": 0.5605722238405023,
      "grad_norm": 2.0844999473675343,
      "learning_rate": 6.019833562207821e-07,
      "loss": 2.3096,
      "step": 4330
    },
    {
      "epoch": 0.5605722238405023,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5352894067764282,
      "eval_runtime": 14.2294,
      "eval_samples_per_second": 3.092,
      "eval_steps_per_second": 0.211,
      "step": 4330
    },
    {
      "epoch": 0.5607016862478558,
      "grad_norm": 1.5067865702213148,
      "learning_rate": 6.017010735128857e-07,
      "loss": 2.4738,
      "step": 4331
    },
    {
      "epoch": 0.5608311486552092,
      "grad_norm": 2.1401722891417294,
      "learning_rate": 6.014187955642677e-07,
      "loss": 2.3574,
      "step": 4332
    },
    {
      "epoch": 0.5609606110625627,
      "grad_norm": 1.7115771031567406,
      "learning_rate": 6.011365224325884e-07,
      "loss": 2.4009,
      "step": 4333
    },
    {
      "epoch": 0.5610900734699161,
      "grad_norm": 1.6361342118722764,
      "learning_rate": 6.008542541755081e-07,
      "loss": 2.3208,
      "step": 4334
    },
    {
      "epoch": 0.5612195358772697,
      "grad_norm": 1.356496895199791,
      "learning_rate": 6.005719908506856e-07,
      "loss": 2.4309,
      "step": 4335
    },
    {
      "epoch": 0.5612195358772697,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5339133739471436,
      "eval_runtime": 15.2556,
      "eval_samples_per_second": 2.884,
      "eval_steps_per_second": 0.197,
      "step": 4335
    },
    {
      "epoch": 0.5613489982846231,
      "grad_norm": 1.221308370546581,
      "learning_rate": 6.002897325157793e-07,
      "loss": 2.1519,
      "step": 4336
    },
    {
      "epoch": 0.5614784606919766,
      "grad_norm": 2.675358818356185,
      "learning_rate": 6.00007479228446e-07,
      "loss": 2.7185,
      "step": 4337
    },
    {
      "epoch": 0.56160792309933,
      "grad_norm": 1.8337408764339245,
      "learning_rate": 5.997252310463418e-07,
      "loss": 2.47,
      "step": 4338
    },
    {
      "epoch": 0.5617373855066835,
      "grad_norm": 1.7740760470384846,
      "learning_rate": 5.994429880271211e-07,
      "loss": 2.559,
      "step": 4339
    },
    {
      "epoch": 0.561866847914037,
      "grad_norm": 1.927484913934167,
      "learning_rate": 5.991607502284383e-07,
      "loss": 2.5396,
      "step": 4340
    },
    {
      "epoch": 0.561866847914037,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.532470703125,
      "eval_runtime": 13.7042,
      "eval_samples_per_second": 3.211,
      "eval_steps_per_second": 0.219,
      "step": 4340
    },
    {
      "epoch": 0.5619963103213904,
      "grad_norm": 1.7099284959469516,
      "learning_rate": 5.988785177079459e-07,
      "loss": 2.457,
      "step": 4341
    },
    {
      "epoch": 0.5621257727287439,
      "grad_norm": 4.647874236406694,
      "learning_rate": 5.985962905232959e-07,
      "loss": 2.416,
      "step": 4342
    },
    {
      "epoch": 0.5622552351360973,
      "grad_norm": 1.9542066969976484,
      "learning_rate": 5.983140687321386e-07,
      "loss": 2.5364,
      "step": 4343
    },
    {
      "epoch": 0.5623846975434508,
      "grad_norm": 1.3377028975600906,
      "learning_rate": 5.980318523921238e-07,
      "loss": 2.3984,
      "step": 4344
    },
    {
      "epoch": 0.5625141599508043,
      "grad_norm": 2.450650803461354,
      "learning_rate": 5.977496415608997e-07,
      "loss": 2.5984,
      "step": 4345
    },
    {
      "epoch": 0.5625141599508043,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5274325609207153,
      "eval_runtime": 14.2484,
      "eval_samples_per_second": 3.088,
      "eval_steps_per_second": 0.211,
      "step": 4345
    },
    {
      "epoch": 0.5626436223581578,
      "grad_norm": 2.247019952296008,
      "learning_rate": 5.974674362961135e-07,
      "loss": 2.6516,
      "step": 4346
    },
    {
      "epoch": 0.5627730847655112,
      "grad_norm": 2.3293819167551955,
      "learning_rate": 5.971852366554116e-07,
      "loss": 2.8132,
      "step": 4347
    },
    {
      "epoch": 0.5629025471728647,
      "grad_norm": 2.2004301921375125,
      "learning_rate": 5.969030426964389e-07,
      "loss": 2.4553,
      "step": 4348
    },
    {
      "epoch": 0.5630320095802182,
      "grad_norm": 1.5640593573696209,
      "learning_rate": 5.966208544768394e-07,
      "loss": 2.5265,
      "step": 4349
    },
    {
      "epoch": 0.5631614719875716,
      "grad_norm": 2.3508723234105955,
      "learning_rate": 5.963386720542554e-07,
      "loss": 2.3914,
      "step": 4350
    },
    {
      "epoch": 0.5631614719875716,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5288752317428589,
      "eval_runtime": 15.089,
      "eval_samples_per_second": 2.916,
      "eval_steps_per_second": 0.199,
      "step": 4350
    },
    {
      "epoch": 0.5632909343949251,
      "grad_norm": 1.719696224374733,
      "learning_rate": 5.960564954863287e-07,
      "loss": 2.4253,
      "step": 4351
    },
    {
      "epoch": 0.5634203968022785,
      "grad_norm": 1.6928187089090314,
      "learning_rate": 5.957743248306992e-07,
      "loss": 2.3608,
      "step": 4352
    },
    {
      "epoch": 0.563549859209632,
      "grad_norm": 1.2109778019038435,
      "learning_rate": 5.954921601450063e-07,
      "loss": 2.3892,
      "step": 4353
    },
    {
      "epoch": 0.5636793216169854,
      "grad_norm": 1.5762652315588639,
      "learning_rate": 5.952100014868879e-07,
      "loss": 2.5181,
      "step": 4354
    },
    {
      "epoch": 0.563808784024339,
      "grad_norm": 1.3981577179119586,
      "learning_rate": 5.949278489139808e-07,
      "loss": 2.3237,
      "step": 4355
    },
    {
      "epoch": 0.563808784024339,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.531782627105713,
      "eval_runtime": 15.3357,
      "eval_samples_per_second": 2.869,
      "eval_steps_per_second": 0.196,
      "step": 4355
    },
    {
      "epoch": 0.5639382464316924,
      "grad_norm": 1.4754293718042735,
      "learning_rate": 5.946457024839196e-07,
      "loss": 2.3621,
      "step": 4356
    },
    {
      "epoch": 0.5640677088390459,
      "grad_norm": 1.712249368352622,
      "learning_rate": 5.94363562254339e-07,
      "loss": 2.3892,
      "step": 4357
    },
    {
      "epoch": 0.5641971712463993,
      "grad_norm": 2.6254923200702622,
      "learning_rate": 5.940814282828719e-07,
      "loss": 2.3833,
      "step": 4358
    },
    {
      "epoch": 0.5643266336537528,
      "grad_norm": 3.885289785657173,
      "learning_rate": 5.937993006271497e-07,
      "loss": 2.6719,
      "step": 4359
    },
    {
      "epoch": 0.5644560960611062,
      "grad_norm": 2.363808303837714,
      "learning_rate": 5.93517179344803e-07,
      "loss": 2.5852,
      "step": 4360
    },
    {
      "epoch": 0.5644560960611062,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.531538486480713,
      "eval_runtime": 14.0524,
      "eval_samples_per_second": 3.131,
      "eval_steps_per_second": 0.213,
      "step": 4360
    },
    {
      "epoch": 0.5645855584684597,
      "grad_norm": 1.602069822782967,
      "learning_rate": 5.932350644934602e-07,
      "loss": 2.2119,
      "step": 4361
    },
    {
      "epoch": 0.5647150208758132,
      "grad_norm": 1.8281623782924643,
      "learning_rate": 5.929529561307496e-07,
      "loss": 2.6133,
      "step": 4362
    },
    {
      "epoch": 0.5648444832831666,
      "grad_norm": 2.0200877176521717,
      "learning_rate": 5.926708543142971e-07,
      "loss": 2.5155,
      "step": 4363
    },
    {
      "epoch": 0.5649739456905201,
      "grad_norm": 1.8284045627951315,
      "learning_rate": 5.923887591017278e-07,
      "loss": 2.6443,
      "step": 4364
    },
    {
      "epoch": 0.5651034080978736,
      "grad_norm": 2.687058918958252,
      "learning_rate": 5.921066705506657e-07,
      "loss": 2.2771,
      "step": 4365
    },
    {
      "epoch": 0.5651034080978736,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5296519994735718,
      "eval_runtime": 15.087,
      "eval_samples_per_second": 2.916,
      "eval_steps_per_second": 0.199,
      "step": 4365
    },
    {
      "epoch": 0.5652328705052271,
      "grad_norm": 1.8236095391076637,
      "learning_rate": 5.918245887187328e-07,
      "loss": 2.2871,
      "step": 4366
    },
    {
      "epoch": 0.5653623329125805,
      "grad_norm": 1.5896062479065716,
      "learning_rate": 5.915425136635501e-07,
      "loss": 2.192,
      "step": 4367
    },
    {
      "epoch": 0.565491795319934,
      "grad_norm": 1.6472960425831566,
      "learning_rate": 5.912604454427371e-07,
      "loss": 2.5454,
      "step": 4368
    },
    {
      "epoch": 0.5656212577272874,
      "grad_norm": 2.9543109508071046,
      "learning_rate": 5.909783841139119e-07,
      "loss": 2.6353,
      "step": 4369
    },
    {
      "epoch": 0.5657507201346409,
      "grad_norm": 1.4553244028788497,
      "learning_rate": 5.906963297346913e-07,
      "loss": 2.3591,
      "step": 4370
    },
    {
      "epoch": 0.5657507201346409,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5302734375,
      "eval_runtime": 14.4201,
      "eval_samples_per_second": 3.051,
      "eval_steps_per_second": 0.208,
      "step": 4370
    },
    {
      "epoch": 0.5658801825419943,
      "grad_norm": 2.1224827481378656,
      "learning_rate": 5.90414282362691e-07,
      "loss": 2.2778,
      "step": 4371
    },
    {
      "epoch": 0.5660096449493478,
      "grad_norm": 1.182766858602216,
      "learning_rate": 5.901322420555242e-07,
      "loss": 2.2878,
      "step": 4372
    },
    {
      "epoch": 0.5661391073567013,
      "grad_norm": 2.4107673011574464,
      "learning_rate": 5.898502088708038e-07,
      "loss": 2.7124,
      "step": 4373
    },
    {
      "epoch": 0.5662685697640547,
      "grad_norm": 1.819894293678706,
      "learning_rate": 5.895681828661407e-07,
      "loss": 2.4404,
      "step": 4374
    },
    {
      "epoch": 0.5663980321714083,
      "grad_norm": 2.159685988426527,
      "learning_rate": 5.892861640991443e-07,
      "loss": 2.4075,
      "step": 4375
    },
    {
      "epoch": 0.5663980321714083,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.527388095855713,
      "eval_runtime": 14.5617,
      "eval_samples_per_second": 3.022,
      "eval_steps_per_second": 0.206,
      "step": 4375
    },
    {
      "epoch": 0.5665274945787617,
      "grad_norm": 2.043764536141545,
      "learning_rate": 5.890041526274227e-07,
      "loss": 2.4575,
      "step": 4376
    },
    {
      "epoch": 0.5666569569861152,
      "grad_norm": 1.5799764033172574,
      "learning_rate": 5.887221485085831e-07,
      "loss": 2.3267,
      "step": 4377
    },
    {
      "epoch": 0.5667864193934686,
      "grad_norm": 1.3463578493932928,
      "learning_rate": 5.884401518002294e-07,
      "loss": 2.4819,
      "step": 4378
    },
    {
      "epoch": 0.5669158818008221,
      "grad_norm": 2.0349278503485153,
      "learning_rate": 5.88158162559966e-07,
      "loss": 2.4309,
      "step": 4379
    },
    {
      "epoch": 0.5670453442081755,
      "grad_norm": 2.770127127905956,
      "learning_rate": 5.878761808453945e-07,
      "loss": 2.248,
      "step": 4380
    },
    {
      "epoch": 0.5670453442081755,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5281871557235718,
      "eval_runtime": 14.9465,
      "eval_samples_per_second": 2.944,
      "eval_steps_per_second": 0.201,
      "step": 4380
    },
    {
      "epoch": 0.567174806615529,
      "grad_norm": 2.134112231434973,
      "learning_rate": 5.875942067141155e-07,
      "loss": 2.6089,
      "step": 4381
    },
    {
      "epoch": 0.5673042690228824,
      "grad_norm": 2.994367714922088,
      "learning_rate": 5.87312240223728e-07,
      "loss": 2.4468,
      "step": 4382
    },
    {
      "epoch": 0.5674337314302359,
      "grad_norm": 1.7663851518555604,
      "learning_rate": 5.870302814318297e-07,
      "loss": 2.394,
      "step": 4383
    },
    {
      "epoch": 0.5675631938375895,
      "grad_norm": 1.4659378815165225,
      "learning_rate": 5.867483303960156e-07,
      "loss": 2.4839,
      "step": 4384
    },
    {
      "epoch": 0.5676926562449429,
      "grad_norm": 1.1933853110709893,
      "learning_rate": 5.864663871738804e-07,
      "loss": 2.3306,
      "step": 4385
    },
    {
      "epoch": 0.5676926562449429,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5296963453292847,
      "eval_runtime": 14.9068,
      "eval_samples_per_second": 2.952,
      "eval_steps_per_second": 0.201,
      "step": 4385
    },
    {
      "epoch": 0.5678221186522964,
      "grad_norm": 2.4091167840293055,
      "learning_rate": 5.861844518230166e-07,
      "loss": 2.7019,
      "step": 4386
    },
    {
      "epoch": 0.5679515810596498,
      "grad_norm": 1.988042526869207,
      "learning_rate": 5.859025244010154e-07,
      "loss": 2.3662,
      "step": 4387
    },
    {
      "epoch": 0.5680810434670033,
      "grad_norm": 1.8566397518006972,
      "learning_rate": 5.856206049654662e-07,
      "loss": 2.4619,
      "step": 4388
    },
    {
      "epoch": 0.5682105058743567,
      "grad_norm": 1.4740486551990213,
      "learning_rate": 5.853386935739565e-07,
      "loss": 2.446,
      "step": 4389
    },
    {
      "epoch": 0.5683399682817102,
      "grad_norm": 2.8905062659500493,
      "learning_rate": 5.850567902840725e-07,
      "loss": 2.3464,
      "step": 4390
    },
    {
      "epoch": 0.5683399682817102,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5274103879928589,
      "eval_runtime": 13.3875,
      "eval_samples_per_second": 3.287,
      "eval_steps_per_second": 0.224,
      "step": 4390
    },
    {
      "epoch": 0.5684694306890636,
      "grad_norm": 1.4397807669253861,
      "learning_rate": 5.847748951533986e-07,
      "loss": 2.4487,
      "step": 4391
    },
    {
      "epoch": 0.5685988930964171,
      "grad_norm": 1.5826540808394478,
      "learning_rate": 5.844930082395179e-07,
      "loss": 2.4158,
      "step": 4392
    },
    {
      "epoch": 0.5687283555037705,
      "grad_norm": 1.4787467223629225,
      "learning_rate": 5.842111296000111e-07,
      "loss": 2.5039,
      "step": 4393
    },
    {
      "epoch": 0.5688578179111241,
      "grad_norm": 2.338872559020097,
      "learning_rate": 5.83929259292458e-07,
      "loss": 2.543,
      "step": 4394
    },
    {
      "epoch": 0.5689872803184776,
      "grad_norm": 2.0668034917106537,
      "learning_rate": 5.836473973744361e-07,
      "loss": 2.387,
      "step": 4395
    },
    {
      "epoch": 0.5689872803184776,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5296076536178589,
      "eval_runtime": 14.3069,
      "eval_samples_per_second": 3.075,
      "eval_steps_per_second": 0.21,
      "step": 4395
    },
    {
      "epoch": 0.569116742725831,
      "grad_norm": 1.400165456144604,
      "learning_rate": 5.833655439035214e-07,
      "loss": 2.3931,
      "step": 4396
    },
    {
      "epoch": 0.5692462051331845,
      "grad_norm": 1.4042769598962068,
      "learning_rate": 5.830836989372884e-07,
      "loss": 2.2837,
      "step": 4397
    },
    {
      "epoch": 0.5693756675405379,
      "grad_norm": 1.9484711577465819,
      "learning_rate": 5.828018625333092e-07,
      "loss": 2.3738,
      "step": 4398
    },
    {
      "epoch": 0.5695051299478914,
      "grad_norm": 1.9466189831270702,
      "learning_rate": 5.825200347491553e-07,
      "loss": 2.4841,
      "step": 4399
    },
    {
      "epoch": 0.5696345923552448,
      "grad_norm": 1.791127433568799,
      "learning_rate": 5.822382156423951e-07,
      "loss": 2.4585,
      "step": 4400
    },
    {
      "epoch": 0.5696345923552448,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.525346279144287,
      "eval_runtime": 14.0008,
      "eval_samples_per_second": 3.143,
      "eval_steps_per_second": 0.214,
      "step": 4400
    },
    {
      "epoch": 0.5697640547625983,
      "grad_norm": 2.8891716286120426,
      "learning_rate": 5.81956405270596e-07,
      "loss": 2.4109,
      "step": 4401
    },
    {
      "epoch": 0.5698935171699517,
      "grad_norm": 2.7550617339644146,
      "learning_rate": 5.816746036913238e-07,
      "loss": 2.4573,
      "step": 4402
    },
    {
      "epoch": 0.5700229795773052,
      "grad_norm": 2.59713592337112,
      "learning_rate": 5.813928109621417e-07,
      "loss": 2.5371,
      "step": 4403
    },
    {
      "epoch": 0.5701524419846588,
      "grad_norm": 2.5926111759252564,
      "learning_rate": 5.811110271406119e-07,
      "loss": 2.3833,
      "step": 4404
    },
    {
      "epoch": 0.5702819043920122,
      "grad_norm": 2.022572978459419,
      "learning_rate": 5.808292522842948e-07,
      "loss": 2.658,
      "step": 4405
    },
    {
      "epoch": 0.5702819043920122,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.528120517730713,
      "eval_runtime": 13.6321,
      "eval_samples_per_second": 3.228,
      "eval_steps_per_second": 0.22,
      "step": 4405
    },
    {
      "epoch": 0.5704113667993657,
      "grad_norm": 3.3056888083335267,
      "learning_rate": 5.805474864507479e-07,
      "loss": 2.6514,
      "step": 4406
    },
    {
      "epoch": 0.5705408292067191,
      "grad_norm": 1.7753400137407758,
      "learning_rate": 5.802657296975278e-07,
      "loss": 2.4954,
      "step": 4407
    },
    {
      "epoch": 0.5706702916140726,
      "grad_norm": 1.8646872930363578,
      "learning_rate": 5.799839820821892e-07,
      "loss": 2.4954,
      "step": 4408
    },
    {
      "epoch": 0.570799754021426,
      "grad_norm": 2.3099731334676683,
      "learning_rate": 5.797022436622847e-07,
      "loss": 2.5325,
      "step": 4409
    },
    {
      "epoch": 0.5709292164287795,
      "grad_norm": 1.3895810608022325,
      "learning_rate": 5.794205144953654e-07,
      "loss": 2.4773,
      "step": 4410
    },
    {
      "epoch": 0.5709292164287795,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5260564088821411,
      "eval_runtime": 13.7311,
      "eval_samples_per_second": 3.204,
      "eval_steps_per_second": 0.218,
      "step": 4410
    },
    {
      "epoch": 0.5710586788361329,
      "grad_norm": 1.3705900301873855,
      "learning_rate": 5.791387946389794e-07,
      "loss": 2.2119,
      "step": 4411
    },
    {
      "epoch": 0.5711881412434864,
      "grad_norm": 1.9473085089700983,
      "learning_rate": 5.788570841506744e-07,
      "loss": 2.7031,
      "step": 4412
    },
    {
      "epoch": 0.5713176036508398,
      "grad_norm": 1.354616648337542,
      "learning_rate": 5.785753830879951e-07,
      "loss": 2.2866,
      "step": 4413
    },
    {
      "epoch": 0.5714470660581934,
      "grad_norm": 1.9862347815882009,
      "learning_rate": 5.782936915084849e-07,
      "loss": 2.4004,
      "step": 4414
    },
    {
      "epoch": 0.5715765284655469,
      "grad_norm": 1.129499225406658,
      "learning_rate": 5.780120094696848e-07,
      "loss": 2.3333,
      "step": 4415
    },
    {
      "epoch": 0.5715765284655469,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5262340307235718,
      "eval_runtime": 14.5752,
      "eval_samples_per_second": 3.019,
      "eval_steps_per_second": 0.206,
      "step": 4415
    },
    {
      "epoch": 0.5717059908729003,
      "grad_norm": 1.6202474907493667,
      "learning_rate": 5.777303370291345e-07,
      "loss": 2.7019,
      "step": 4416
    },
    {
      "epoch": 0.5718354532802538,
      "grad_norm": 1.9191105811353815,
      "learning_rate": 5.774486742443706e-07,
      "loss": 2.4966,
      "step": 4417
    },
    {
      "epoch": 0.5719649156876072,
      "grad_norm": 2.037320423421432,
      "learning_rate": 5.771670211729289e-07,
      "loss": 2.4553,
      "step": 4418
    },
    {
      "epoch": 0.5720943780949607,
      "grad_norm": 2.510736120238449,
      "learning_rate": 5.768853778723426e-07,
      "loss": 2.5156,
      "step": 4419
    },
    {
      "epoch": 0.5722238405023141,
      "grad_norm": 1.4703227582133935,
      "learning_rate": 5.766037444001432e-07,
      "loss": 2.4041,
      "step": 4420
    },
    {
      "epoch": 0.5722238405023141,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5289639234542847,
      "eval_runtime": 14.9362,
      "eval_samples_per_second": 2.946,
      "eval_steps_per_second": 0.201,
      "step": 4420
    },
    {
      "epoch": 0.5723533029096676,
      "grad_norm": 1.8304570511644205,
      "learning_rate": 5.763221208138599e-07,
      "loss": 2.3875,
      "step": 4421
    },
    {
      "epoch": 0.572482765317021,
      "grad_norm": 2.473405558929656,
      "learning_rate": 5.760405071710203e-07,
      "loss": 2.4817,
      "step": 4422
    },
    {
      "epoch": 0.5726122277243746,
      "grad_norm": 2.281016263263221,
      "learning_rate": 5.757589035291493e-07,
      "loss": 2.4146,
      "step": 4423
    },
    {
      "epoch": 0.572741690131728,
      "grad_norm": 1.8188954735236613,
      "learning_rate": 5.754773099457703e-07,
      "loss": 2.4629,
      "step": 4424
    },
    {
      "epoch": 0.5728711525390815,
      "grad_norm": 1.5946402814098786,
      "learning_rate": 5.751957264784045e-07,
      "loss": 2.4077,
      "step": 4425
    },
    {
      "epoch": 0.5728711525390815,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5260342359542847,
      "eval_runtime": 14.7625,
      "eval_samples_per_second": 2.981,
      "eval_steps_per_second": 0.203,
      "step": 4425
    },
    {
      "epoch": 0.573000614946435,
      "grad_norm": 1.4585366481901132,
      "learning_rate": 5.74914153184571e-07,
      "loss": 2.2961,
      "step": 4426
    },
    {
      "epoch": 0.5731300773537884,
      "grad_norm": 2.3032195190792484,
      "learning_rate": 5.746325901217873e-07,
      "loss": 2.4172,
      "step": 4427
    },
    {
      "epoch": 0.5732595397611419,
      "grad_norm": 1.2597632216703234,
      "learning_rate": 5.743510373475676e-07,
      "loss": 2.4263,
      "step": 4428
    },
    {
      "epoch": 0.5733890021684953,
      "grad_norm": 1.5173838874780994,
      "learning_rate": 5.740694949194252e-07,
      "loss": 2.4333,
      "step": 4429
    },
    {
      "epoch": 0.5735184645758488,
      "grad_norm": 2.401684593596505,
      "learning_rate": 5.737879628948706e-07,
      "loss": 2.3127,
      "step": 4430
    },
    {
      "epoch": 0.5735184645758488,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5231267213821411,
      "eval_runtime": 13.8635,
      "eval_samples_per_second": 3.174,
      "eval_steps_per_second": 0.216,
      "step": 4430
    },
    {
      "epoch": 0.5736479269832022,
      "grad_norm": 1.4510733736592982,
      "learning_rate": 5.735064413314127e-07,
      "loss": 2.311,
      "step": 4431
    },
    {
      "epoch": 0.5737773893905557,
      "grad_norm": 1.6857797273584392,
      "learning_rate": 5.73224930286558e-07,
      "loss": 2.6533,
      "step": 4432
    },
    {
      "epoch": 0.5739068517979092,
      "grad_norm": 1.3458299598883972,
      "learning_rate": 5.729434298178109e-07,
      "loss": 2.2466,
      "step": 4433
    },
    {
      "epoch": 0.5740363142052627,
      "grad_norm": 2.2297329226935325,
      "learning_rate": 5.72661939982673e-07,
      "loss": 2.4878,
      "step": 4434
    },
    {
      "epoch": 0.5741657766126161,
      "grad_norm": 2.213536585372666,
      "learning_rate": 5.723804608386449e-07,
      "loss": 2.4495,
      "step": 4435
    },
    {
      "epoch": 0.5741657766126161,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.525346279144287,
      "eval_runtime": 15.4433,
      "eval_samples_per_second": 2.849,
      "eval_steps_per_second": 0.194,
      "step": 4435
    },
    {
      "epoch": 0.5742952390199696,
      "grad_norm": 1.7665518674641991,
      "learning_rate": 5.720989924432242e-07,
      "loss": 2.4937,
      "step": 4436
    },
    {
      "epoch": 0.574424701427323,
      "grad_norm": 1.9313632127953595,
      "learning_rate": 5.718175348539066e-07,
      "loss": 2.6377,
      "step": 4437
    },
    {
      "epoch": 0.5745541638346765,
      "grad_norm": 1.4411096246688953,
      "learning_rate": 5.715360881281855e-07,
      "loss": 2.4055,
      "step": 4438
    },
    {
      "epoch": 0.57468362624203,
      "grad_norm": 2.4802929117215347,
      "learning_rate": 5.712546523235519e-07,
      "loss": 2.2656,
      "step": 4439
    },
    {
      "epoch": 0.5748130886493834,
      "grad_norm": 2.0575709791767127,
      "learning_rate": 5.70973227497495e-07,
      "loss": 2.4155,
      "step": 4440
    },
    {
      "epoch": 0.5748130886493834,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5282315015792847,
      "eval_runtime": 15.0436,
      "eval_samples_per_second": 2.925,
      "eval_steps_per_second": 0.199,
      "step": 4440
    },
    {
      "epoch": 0.5749425510567369,
      "grad_norm": 1.6075160428351194,
      "learning_rate": 5.706918137075016e-07,
      "loss": 2.3403,
      "step": 4441
    },
    {
      "epoch": 0.5750720134640903,
      "grad_norm": 2.0210791928496072,
      "learning_rate": 5.704104110110558e-07,
      "loss": 2.5894,
      "step": 4442
    },
    {
      "epoch": 0.5752014758714439,
      "grad_norm": 3.581780719538034,
      "learning_rate": 5.701290194656401e-07,
      "loss": 2.5481,
      "step": 4443
    },
    {
      "epoch": 0.5753309382787973,
      "grad_norm": 2.3132010101035094,
      "learning_rate": 5.698476391287346e-07,
      "loss": 2.5286,
      "step": 4444
    },
    {
      "epoch": 0.5754604006861508,
      "grad_norm": 2.1634873054041095,
      "learning_rate": 5.695662700578164e-07,
      "loss": 2.7505,
      "step": 4445
    },
    {
      "epoch": 0.5754604006861508,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5262340307235718,
      "eval_runtime": 15.2581,
      "eval_samples_per_second": 2.884,
      "eval_steps_per_second": 0.197,
      "step": 4445
    },
    {
      "epoch": 0.5755898630935042,
      "grad_norm": 4.710934020276544,
      "learning_rate": 5.692849123103612e-07,
      "loss": 2.8469,
      "step": 4446
    },
    {
      "epoch": 0.5757193255008577,
      "grad_norm": 2.5049038293057797,
      "learning_rate": 5.69003565943842e-07,
      "loss": 2.1226,
      "step": 4447
    },
    {
      "epoch": 0.5758487879082111,
      "grad_norm": 1.85147163157558,
      "learning_rate": 5.687222310157292e-07,
      "loss": 2.5762,
      "step": 4448
    },
    {
      "epoch": 0.5759782503155646,
      "grad_norm": 2.6613966355536616,
      "learning_rate": 5.68440907583492e-07,
      "loss": 2.2188,
      "step": 4449
    },
    {
      "epoch": 0.576107712722918,
      "grad_norm": 3.426319030938563,
      "learning_rate": 5.681595957045953e-07,
      "loss": 2.5869,
      "step": 4450
    },
    {
      "epoch": 0.576107712722918,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5259898900985718,
      "eval_runtime": 14.7706,
      "eval_samples_per_second": 2.979,
      "eval_steps_per_second": 0.203,
      "step": 4450
    },
    {
      "epoch": 0.5762371751302715,
      "grad_norm": 1.383261421992041,
      "learning_rate": 5.678782954365035e-07,
      "loss": 2.3679,
      "step": 4451
    },
    {
      "epoch": 0.576366637537625,
      "grad_norm": 2.1947644694985136,
      "learning_rate": 5.675970068366775e-07,
      "loss": 2.2419,
      "step": 4452
    },
    {
      "epoch": 0.5764960999449785,
      "grad_norm": 3.7721305425759724,
      "learning_rate": 5.673157299625763e-07,
      "loss": 2.3655,
      "step": 4453
    },
    {
      "epoch": 0.576625562352332,
      "grad_norm": 1.2867292570156783,
      "learning_rate": 5.670344648716567e-07,
      "loss": 2.2977,
      "step": 4454
    },
    {
      "epoch": 0.5767550247596854,
      "grad_norm": 1.5538208293258153,
      "learning_rate": 5.667532116213727e-07,
      "loss": 2.4429,
      "step": 4455
    },
    {
      "epoch": 0.5767550247596854,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.52880859375,
      "eval_runtime": 14.1241,
      "eval_samples_per_second": 3.115,
      "eval_steps_per_second": 0.212,
      "step": 4455
    },
    {
      "epoch": 0.5768844871670389,
      "grad_norm": 1.965334129077527,
      "learning_rate": 5.664719702691755e-07,
      "loss": 2.604,
      "step": 4456
    },
    {
      "epoch": 0.5770139495743923,
      "grad_norm": 4.026632870232991,
      "learning_rate": 5.661907408725148e-07,
      "loss": 2.5515,
      "step": 4457
    },
    {
      "epoch": 0.5771434119817458,
      "grad_norm": 4.676562484974604,
      "learning_rate": 5.659095234888372e-07,
      "loss": 2.624,
      "step": 4458
    },
    {
      "epoch": 0.5772728743890992,
      "grad_norm": 1.730909330499002,
      "learning_rate": 5.656283181755872e-07,
      "loss": 2.3916,
      "step": 4459
    },
    {
      "epoch": 0.5774023367964527,
      "grad_norm": 1.5817781744530113,
      "learning_rate": 5.653471249902069e-07,
      "loss": 2.4617,
      "step": 4460
    },
    {
      "epoch": 0.5774023367964527,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5274769067764282,
      "eval_runtime": 14.4784,
      "eval_samples_per_second": 3.039,
      "eval_steps_per_second": 0.207,
      "step": 4460
    },
    {
      "epoch": 0.5775317992038062,
      "grad_norm": 2.818841732056183,
      "learning_rate": 5.65065943990135e-07,
      "loss": 2.3599,
      "step": 4461
    },
    {
      "epoch": 0.5776612616111596,
      "grad_norm": 1.0506262480272102,
      "learning_rate": 5.64784775232809e-07,
      "loss": 2.2012,
      "step": 4462
    },
    {
      "epoch": 0.5777907240185132,
      "grad_norm": 1.5922862410622813,
      "learning_rate": 5.645036187756633e-07,
      "loss": 2.4409,
      "step": 4463
    },
    {
      "epoch": 0.5779201864258666,
      "grad_norm": 1.4695994767055949,
      "learning_rate": 5.642224746761296e-07,
      "loss": 2.3076,
      "step": 4464
    },
    {
      "epoch": 0.5780496488332201,
      "grad_norm": 1.6604665725323235,
      "learning_rate": 5.639413429916372e-07,
      "loss": 2.4216,
      "step": 4465
    },
    {
      "epoch": 0.5780496488332201,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.527388095855713,
      "eval_runtime": 14.5962,
      "eval_samples_per_second": 3.014,
      "eval_steps_per_second": 0.206,
      "step": 4465
    },
    {
      "epoch": 0.5781791112405735,
      "grad_norm": 1.9544969857678574,
      "learning_rate": 5.636602237796134e-07,
      "loss": 2.3052,
      "step": 4466
    },
    {
      "epoch": 0.578308573647927,
      "grad_norm": 1.3974319686009802,
      "learning_rate": 5.633791170974819e-07,
      "loss": 2.457,
      "step": 4467
    },
    {
      "epoch": 0.5784380360552804,
      "grad_norm": 1.7721137061460703,
      "learning_rate": 5.630980230026649e-07,
      "loss": 2.1443,
      "step": 4468
    },
    {
      "epoch": 0.5785674984626339,
      "grad_norm": 1.5041703755625444,
      "learning_rate": 5.628169415525812e-07,
      "loss": 2.4646,
      "step": 4469
    },
    {
      "epoch": 0.5786969608699873,
      "grad_norm": 1.8893228776421431,
      "learning_rate": 5.625358728046478e-07,
      "loss": 2.531,
      "step": 4470
    },
    {
      "epoch": 0.5786969608699873,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.529541015625,
      "eval_runtime": 14.0562,
      "eval_samples_per_second": 3.13,
      "eval_steps_per_second": 0.213,
      "step": 4470
    },
    {
      "epoch": 0.5788264232773408,
      "grad_norm": 2.5719945207698816,
      "learning_rate": 5.622548168162781e-07,
      "loss": 2.6311,
      "step": 4471
    },
    {
      "epoch": 0.5789558856846944,
      "grad_norm": 1.9246790601651034,
      "learning_rate": 5.619737736448844e-07,
      "loss": 2.6304,
      "step": 4472
    },
    {
      "epoch": 0.5790853480920478,
      "grad_norm": 2.0237890836361982,
      "learning_rate": 5.616927433478747e-07,
      "loss": 2.2788,
      "step": 4473
    },
    {
      "epoch": 0.5792148104994013,
      "grad_norm": 2.9359188910380736,
      "learning_rate": 5.614117259826553e-07,
      "loss": 2.4771,
      "step": 4474
    },
    {
      "epoch": 0.5793442729067547,
      "grad_norm": 1.1558513778999286,
      "learning_rate": 5.611307216066296e-07,
      "loss": 2.3115,
      "step": 4475
    },
    {
      "epoch": 0.5793442729067547,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5287864208221436,
      "eval_runtime": 14.0072,
      "eval_samples_per_second": 3.141,
      "eval_steps_per_second": 0.214,
      "step": 4475
    },
    {
      "epoch": 0.5794737353141082,
      "grad_norm": 1.8909683015214567,
      "learning_rate": 5.608497302771987e-07,
      "loss": 2.1995,
      "step": 4476
    },
    {
      "epoch": 0.5796031977214616,
      "grad_norm": 1.4889476465015383,
      "learning_rate": 5.605687520517609e-07,
      "loss": 2.3591,
      "step": 4477
    },
    {
      "epoch": 0.5797326601288151,
      "grad_norm": 1.4010784570748078,
      "learning_rate": 5.602877869877113e-07,
      "loss": 2.4226,
      "step": 4478
    },
    {
      "epoch": 0.5798621225361685,
      "grad_norm": 3.120190715671266,
      "learning_rate": 5.600068351424426e-07,
      "loss": 2.4722,
      "step": 4479
    },
    {
      "epoch": 0.579991584943522,
      "grad_norm": 1.350775285953974,
      "learning_rate": 5.597258965733453e-07,
      "loss": 2.4858,
      "step": 4480
    },
    {
      "epoch": 0.579991584943522,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5295631885528564,
      "eval_runtime": 14.354,
      "eval_samples_per_second": 3.065,
      "eval_steps_per_second": 0.209,
      "step": 4480
    },
    {
      "epoch": 0.5801210473508754,
      "grad_norm": 1.0767364308174234,
      "learning_rate": 5.594449713378068e-07,
      "loss": 2.3174,
      "step": 4481
    },
    {
      "epoch": 0.580250509758229,
      "grad_norm": 1.4328183638098897,
      "learning_rate": 5.591640594932116e-07,
      "loss": 2.394,
      "step": 4482
    },
    {
      "epoch": 0.5803799721655825,
      "grad_norm": 1.6070932440065626,
      "learning_rate": 5.588831610969417e-07,
      "loss": 2.5144,
      "step": 4483
    },
    {
      "epoch": 0.5805094345729359,
      "grad_norm": 4.211060137167303,
      "learning_rate": 5.586022762063762e-07,
      "loss": 2.9375,
      "step": 4484
    },
    {
      "epoch": 0.5806388969802894,
      "grad_norm": 2.2107982240612287,
      "learning_rate": 5.583214048788917e-07,
      "loss": 2.262,
      "step": 4485
    },
    {
      "epoch": 0.5806388969802894,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.528076171875,
      "eval_runtime": 14.5999,
      "eval_samples_per_second": 3.014,
      "eval_steps_per_second": 0.205,
      "step": 4485
    },
    {
      "epoch": 0.5807683593876428,
      "grad_norm": 1.4940383019804222,
      "learning_rate": 5.580405471718617e-07,
      "loss": 2.3945,
      "step": 4486
    },
    {
      "epoch": 0.5808978217949963,
      "grad_norm": 3.9055127690181153,
      "learning_rate": 5.577597031426572e-07,
      "loss": 2.5847,
      "step": 4487
    },
    {
      "epoch": 0.5810272842023497,
      "grad_norm": 1.5190296165711294,
      "learning_rate": 5.574788728486463e-07,
      "loss": 2.4238,
      "step": 4488
    },
    {
      "epoch": 0.5811567466097032,
      "grad_norm": 2.553099489896722,
      "learning_rate": 5.571980563471941e-07,
      "loss": 2.5259,
      "step": 4489
    },
    {
      "epoch": 0.5812862090170566,
      "grad_norm": 1.2215964268693216,
      "learning_rate": 5.569172536956632e-07,
      "loss": 2.3818,
      "step": 4490
    },
    {
      "epoch": 0.5812862090170566,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5287864208221436,
      "eval_runtime": 15.8192,
      "eval_samples_per_second": 2.781,
      "eval_steps_per_second": 0.19,
      "step": 4490
    },
    {
      "epoch": 0.5814156714244101,
      "grad_norm": 1.3329980872126876,
      "learning_rate": 5.566364649514132e-07,
      "loss": 2.4604,
      "step": 4491
    },
    {
      "epoch": 0.5815451338317636,
      "grad_norm": 4.380748109242493,
      "learning_rate": 5.563556901718008e-07,
      "loss": 2.7351,
      "step": 4492
    },
    {
      "epoch": 0.5816745962391171,
      "grad_norm": 1.4112095905544813,
      "learning_rate": 5.560749294141801e-07,
      "loss": 2.3372,
      "step": 4493
    },
    {
      "epoch": 0.5818040586464706,
      "grad_norm": 1.8576856419271832,
      "learning_rate": 5.557941827359023e-07,
      "loss": 2.5603,
      "step": 4494
    },
    {
      "epoch": 0.581933521053824,
      "grad_norm": 1.1999207118768667,
      "learning_rate": 5.555134501943153e-07,
      "loss": 2.4482,
      "step": 4495
    },
    {
      "epoch": 0.581933521053824,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5308948755264282,
      "eval_runtime": 14.637,
      "eval_samples_per_second": 3.006,
      "eval_steps_per_second": 0.205,
      "step": 4495
    },
    {
      "epoch": 0.5820629834611775,
      "grad_norm": 1.9533276199200365,
      "learning_rate": 5.552327318467645e-07,
      "loss": 2.3313,
      "step": 4496
    },
    {
      "epoch": 0.5821924458685309,
      "grad_norm": 1.617461968051022,
      "learning_rate": 5.549520277505922e-07,
      "loss": 2.2388,
      "step": 4497
    },
    {
      "epoch": 0.5823219082758844,
      "grad_norm": 1.7036046199161534,
      "learning_rate": 5.54671337963138e-07,
      "loss": 2.5518,
      "step": 4498
    },
    {
      "epoch": 0.5824513706832378,
      "grad_norm": 1.766900389112874,
      "learning_rate": 5.543906625417389e-07,
      "loss": 2.3081,
      "step": 4499
    },
    {
      "epoch": 0.5825808330905913,
      "grad_norm": 2.142857262001564,
      "learning_rate": 5.54110001543728e-07,
      "loss": 2.4951,
      "step": 4500
    },
    {
      "epoch": 0.5825808330905913,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.52880859375,
      "eval_runtime": 14.8016,
      "eval_samples_per_second": 2.973,
      "eval_steps_per_second": 0.203,
      "step": 4500
    },
    {
      "epoch": 0.5827102954979447,
      "grad_norm": 2.099203147267288,
      "learning_rate": 5.538293550264361e-07,
      "loss": 2.3979,
      "step": 4501
    },
    {
      "epoch": 0.5828397579052983,
      "grad_norm": 1.7216967856082526,
      "learning_rate": 5.535487230471908e-07,
      "loss": 2.5493,
      "step": 4502
    },
    {
      "epoch": 0.5829692203126517,
      "grad_norm": 2.983331481206411,
      "learning_rate": 5.532681056633173e-07,
      "loss": 2.4946,
      "step": 4503
    },
    {
      "epoch": 0.5830986827200052,
      "grad_norm": 1.5175393776414896,
      "learning_rate": 5.529875029321371e-07,
      "loss": 2.5122,
      "step": 4504
    },
    {
      "epoch": 0.5832281451273587,
      "grad_norm": 1.6911540484770793,
      "learning_rate": 5.527069149109693e-07,
      "loss": 2.4952,
      "step": 4505
    },
    {
      "epoch": 0.5832281451273587,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5323597192764282,
      "eval_runtime": 13.0095,
      "eval_samples_per_second": 3.382,
      "eval_steps_per_second": 0.231,
      "step": 4505
    },
    {
      "epoch": 0.5833576075347121,
      "grad_norm": 1.2548151212969714,
      "learning_rate": 5.524263416571292e-07,
      "loss": 2.376,
      "step": 4506
    },
    {
      "epoch": 0.5834870699420656,
      "grad_norm": 1.7862259878916473,
      "learning_rate": 5.521457832279297e-07,
      "loss": 2.4854,
      "step": 4507
    },
    {
      "epoch": 0.583616532349419,
      "grad_norm": 1.6162308235247107,
      "learning_rate": 5.518652396806808e-07,
      "loss": 2.4661,
      "step": 4508
    },
    {
      "epoch": 0.5837459947567725,
      "grad_norm": 1.7928326604627893,
      "learning_rate": 5.51584711072689e-07,
      "loss": 2.6062,
      "step": 4509
    },
    {
      "epoch": 0.5838754571641259,
      "grad_norm": 1.4103104958052906,
      "learning_rate": 5.513041974612581e-07,
      "loss": 2.2976,
      "step": 4510
    },
    {
      "epoch": 0.5838754571641259,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.532270908355713,
      "eval_runtime": 14.1208,
      "eval_samples_per_second": 3.116,
      "eval_steps_per_second": 0.212,
      "step": 4510
    },
    {
      "epoch": 0.5840049195714795,
      "grad_norm": 1.1372773037517108,
      "learning_rate": 5.510236989036887e-07,
      "loss": 2.3416,
      "step": 4511
    },
    {
      "epoch": 0.5841343819788329,
      "grad_norm": 1.4074753345065598,
      "learning_rate": 5.507432154572783e-07,
      "loss": 2.4895,
      "step": 4512
    },
    {
      "epoch": 0.5842638443861864,
      "grad_norm": 1.9941149831722298,
      "learning_rate": 5.50462747179321e-07,
      "loss": 2.5974,
      "step": 4513
    },
    {
      "epoch": 0.5843933067935398,
      "grad_norm": 1.339234224469598,
      "learning_rate": 5.501822941271084e-07,
      "loss": 2.3149,
      "step": 4514
    },
    {
      "epoch": 0.5845227692008933,
      "grad_norm": 2.9578816349066503,
      "learning_rate": 5.499018563579289e-07,
      "loss": 2.4099,
      "step": 4515
    },
    {
      "epoch": 0.5845227692008933,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.532270908355713,
      "eval_runtime": 14.5651,
      "eval_samples_per_second": 3.021,
      "eval_steps_per_second": 0.206,
      "step": 4515
    },
    {
      "epoch": 0.5846522316082468,
      "grad_norm": 1.3470179539550475,
      "learning_rate": 5.496214339290674e-07,
      "loss": 2.2275,
      "step": 4516
    },
    {
      "epoch": 0.5847816940156002,
      "grad_norm": 2.039081342749023,
      "learning_rate": 5.493410268978057e-07,
      "loss": 2.425,
      "step": 4517
    },
    {
      "epoch": 0.5849111564229537,
      "grad_norm": 1.6992957388751944,
      "learning_rate": 5.49060635321423e-07,
      "loss": 2.2307,
      "step": 4518
    },
    {
      "epoch": 0.5850406188303071,
      "grad_norm": 1.7952098988362484,
      "learning_rate": 5.487802592571946e-07,
      "loss": 2.283,
      "step": 4519
    },
    {
      "epoch": 0.5851700812376606,
      "grad_norm": 1.1277076751073452,
      "learning_rate": 5.484998987623931e-07,
      "loss": 2.2073,
      "step": 4520
    },
    {
      "epoch": 0.5851700812376606,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.530806064605713,
      "eval_runtime": 13.876,
      "eval_samples_per_second": 3.171,
      "eval_steps_per_second": 0.216,
      "step": 4520
    },
    {
      "epoch": 0.5852995436450141,
      "grad_norm": 1.7475169747668098,
      "learning_rate": 5.482195538942877e-07,
      "loss": 2.396,
      "step": 4521
    },
    {
      "epoch": 0.5854290060523676,
      "grad_norm": 1.560242831665156,
      "learning_rate": 5.479392247101453e-07,
      "loss": 2.3115,
      "step": 4522
    },
    {
      "epoch": 0.585558468459721,
      "grad_norm": 1.9106406246836625,
      "learning_rate": 5.476589112672278e-07,
      "loss": 2.4971,
      "step": 4523
    },
    {
      "epoch": 0.5856879308670745,
      "grad_norm": 2.2506451805460013,
      "learning_rate": 5.473786136227953e-07,
      "loss": 2.6135,
      "step": 4524
    },
    {
      "epoch": 0.5858173932744279,
      "grad_norm": 1.4933714712992732,
      "learning_rate": 5.470983318341043e-07,
      "loss": 2.4385,
      "step": 4525
    },
    {
      "epoch": 0.5858173932744279,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5323597192764282,
      "eval_runtime": 14.1022,
      "eval_samples_per_second": 3.12,
      "eval_steps_per_second": 0.213,
      "step": 4525
    },
    {
      "epoch": 0.5859468556817814,
      "grad_norm": 1.5941742999729245,
      "learning_rate": 5.468180659584079e-07,
      "loss": 2.3247,
      "step": 4526
    },
    {
      "epoch": 0.5860763180891349,
      "grad_norm": 2.6250188567711357,
      "learning_rate": 5.465378160529566e-07,
      "loss": 2.2317,
      "step": 4527
    },
    {
      "epoch": 0.5862057804964883,
      "grad_norm": 1.4908669546125441,
      "learning_rate": 5.462575821749964e-07,
      "loss": 2.2434,
      "step": 4528
    },
    {
      "epoch": 0.5863352429038418,
      "grad_norm": 1.943581959195338,
      "learning_rate": 5.459773643817711e-07,
      "loss": 2.5122,
      "step": 4529
    },
    {
      "epoch": 0.5864647053111952,
      "grad_norm": 1.6901467499274367,
      "learning_rate": 5.456971627305209e-07,
      "loss": 2.3955,
      "step": 4530
    },
    {
      "epoch": 0.5864647053111952,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.530229091644287,
      "eval_runtime": 13.8803,
      "eval_samples_per_second": 3.17,
      "eval_steps_per_second": 0.216,
      "step": 4530
    },
    {
      "epoch": 0.5865941677185488,
      "grad_norm": 1.9029360906678596,
      "learning_rate": 5.454169772784826e-07,
      "loss": 2.4451,
      "step": 4531
    },
    {
      "epoch": 0.5867236301259022,
      "grad_norm": 1.4909048976066253,
      "learning_rate": 5.451368080828898e-07,
      "loss": 2.2102,
      "step": 4532
    },
    {
      "epoch": 0.5868530925332557,
      "grad_norm": 2.6305479007837027,
      "learning_rate": 5.448566552009727e-07,
      "loss": 2.5796,
      "step": 4533
    },
    {
      "epoch": 0.5869825549406091,
      "grad_norm": 1.3453952978541655,
      "learning_rate": 5.445765186899582e-07,
      "loss": 2.6185,
      "step": 4534
    },
    {
      "epoch": 0.5871120173479626,
      "grad_norm": 1.5782189018611543,
      "learning_rate": 5.442963986070699e-07,
      "loss": 2.459,
      "step": 4535
    },
    {
      "epoch": 0.5871120173479626,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5301402807235718,
      "eval_runtime": 14.0918,
      "eval_samples_per_second": 3.122,
      "eval_steps_per_second": 0.213,
      "step": 4535
    },
    {
      "epoch": 0.587241479755316,
      "grad_norm": 1.4889708788943745,
      "learning_rate": 5.440162950095278e-07,
      "loss": 2.2698,
      "step": 4536
    },
    {
      "epoch": 0.5873709421626695,
      "grad_norm": 1.5397531188443934,
      "learning_rate": 5.43736207954549e-07,
      "loss": 2.1682,
      "step": 4537
    },
    {
      "epoch": 0.587500404570023,
      "grad_norm": 1.6268127818767149,
      "learning_rate": 5.43456137499347e-07,
      "loss": 2.458,
      "step": 4538
    },
    {
      "epoch": 0.5876298669773764,
      "grad_norm": 1.2645818867959602,
      "learning_rate": 5.431760837011316e-07,
      "loss": 2.5474,
      "step": 4539
    },
    {
      "epoch": 0.5877593293847299,
      "grad_norm": 3.0135533226952314,
      "learning_rate": 5.428960466171096e-07,
      "loss": 2.3164,
      "step": 4540
    },
    {
      "epoch": 0.5877593293847299,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5308948755264282,
      "eval_runtime": 14.4124,
      "eval_samples_per_second": 3.053,
      "eval_steps_per_second": 0.208,
      "step": 4540
    },
    {
      "epoch": 0.5878887917920834,
      "grad_norm": 2.0835073690757033,
      "learning_rate": 5.426160263044842e-07,
      "loss": 2.4875,
      "step": 4541
    },
    {
      "epoch": 0.5880182541994369,
      "grad_norm": 2.0805213822656685,
      "learning_rate": 5.423360228204555e-07,
      "loss": 2.5332,
      "step": 4542
    },
    {
      "epoch": 0.5881477166067903,
      "grad_norm": 1.5671322682078817,
      "learning_rate": 5.420560362222192e-07,
      "loss": 2.3904,
      "step": 4543
    },
    {
      "epoch": 0.5882771790141438,
      "grad_norm": 2.343000694093633,
      "learning_rate": 5.417760665669694e-07,
      "loss": 2.8076,
      "step": 4544
    },
    {
      "epoch": 0.5884066414214972,
      "grad_norm": 1.5512598699330886,
      "learning_rate": 5.414961139118944e-07,
      "loss": 2.4185,
      "step": 4545
    },
    {
      "epoch": 0.5884066414214972,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5315829515457153,
      "eval_runtime": 14.9634,
      "eval_samples_per_second": 2.94,
      "eval_steps_per_second": 0.2,
      "step": 4545
    },
    {
      "epoch": 0.5885361038288507,
      "grad_norm": 1.774725817234927,
      "learning_rate": 5.412161783141808e-07,
      "loss": 2.5615,
      "step": 4546
    },
    {
      "epoch": 0.5886655662362041,
      "grad_norm": 1.467001266115686,
      "learning_rate": 5.409362598310108e-07,
      "loss": 2.2417,
      "step": 4547
    },
    {
      "epoch": 0.5887950286435576,
      "grad_norm": 1.1325015476708955,
      "learning_rate": 5.406563585195637e-07,
      "loss": 2.2734,
      "step": 4548
    },
    {
      "epoch": 0.588924491050911,
      "grad_norm": 1.3267313788790156,
      "learning_rate": 5.40376474437015e-07,
      "loss": 2.3374,
      "step": 4549
    },
    {
      "epoch": 0.5890539534582645,
      "grad_norm": 2.6555908176038026,
      "learning_rate": 5.40096607640537e-07,
      "loss": 2.1421,
      "step": 4550
    },
    {
      "epoch": 0.5890539534582645,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5258567333221436,
      "eval_runtime": 14.6807,
      "eval_samples_per_second": 2.997,
      "eval_steps_per_second": 0.204,
      "step": 4550
    },
    {
      "epoch": 0.5891834158656181,
      "grad_norm": 2.3566320557455303,
      "learning_rate": 5.398167581872973e-07,
      "loss": 2.4551,
      "step": 4551
    },
    {
      "epoch": 0.5893128782729715,
      "grad_norm": 2.3051287006257217,
      "learning_rate": 5.395369261344615e-07,
      "loss": 2.3357,
      "step": 4552
    },
    {
      "epoch": 0.589442340680325,
      "grad_norm": 1.3492685482677123,
      "learning_rate": 5.392571115391908e-07,
      "loss": 2.4094,
      "step": 4553
    },
    {
      "epoch": 0.5895718030876784,
      "grad_norm": 3.098803676934089,
      "learning_rate": 5.38977314458643e-07,
      "loss": 2.6055,
      "step": 4554
    },
    {
      "epoch": 0.5897012654950319,
      "grad_norm": 1.2884599551743159,
      "learning_rate": 5.386975349499723e-07,
      "loss": 2.3872,
      "step": 4555
    },
    {
      "epoch": 0.5897012654950319,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5323597192764282,
      "eval_runtime": 13.8689,
      "eval_samples_per_second": 3.173,
      "eval_steps_per_second": 0.216,
      "step": 4555
    },
    {
      "epoch": 0.5898307279023853,
      "grad_norm": 1.8597443712744164,
      "learning_rate": 5.384177730703294e-07,
      "loss": 2.2432,
      "step": 4556
    },
    {
      "epoch": 0.5899601903097388,
      "grad_norm": 2.7717197013909702,
      "learning_rate": 5.381380288768613e-07,
      "loss": 2.4131,
      "step": 4557
    },
    {
      "epoch": 0.5900896527170922,
      "grad_norm": 3.6581479424697143,
      "learning_rate": 5.378583024267113e-07,
      "loss": 2.3792,
      "step": 4558
    },
    {
      "epoch": 0.5902191151244457,
      "grad_norm": 1.6425422550060185,
      "learning_rate": 5.375785937770193e-07,
      "loss": 2.4292,
      "step": 4559
    },
    {
      "epoch": 0.5903485775317993,
      "grad_norm": 1.6910719812708628,
      "learning_rate": 5.372989029849216e-07,
      "loss": 2.4316,
      "step": 4560
    },
    {
      "epoch": 0.5903485775317993,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.528031826019287,
      "eval_runtime": 14.4014,
      "eval_samples_per_second": 3.055,
      "eval_steps_per_second": 0.208,
      "step": 4560
    },
    {
      "epoch": 0.5904780399391527,
      "grad_norm": 1.0865596742087698,
      "learning_rate": 5.370192301075506e-07,
      "loss": 2.23,
      "step": 4561
    },
    {
      "epoch": 0.5906075023465062,
      "grad_norm": 3.975924273463063,
      "learning_rate": 5.367395752020352e-07,
      "loss": 2.6497,
      "step": 4562
    },
    {
      "epoch": 0.5907369647538596,
      "grad_norm": 1.4755731215712613,
      "learning_rate": 5.364599383255004e-07,
      "loss": 2.3538,
      "step": 4563
    },
    {
      "epoch": 0.5908664271612131,
      "grad_norm": 1.2471566253016024,
      "learning_rate": 5.361803195350678e-07,
      "loss": 2.281,
      "step": 4564
    },
    {
      "epoch": 0.5909958895685665,
      "grad_norm": 1.95543319406126,
      "learning_rate": 5.359007188878552e-07,
      "loss": 2.4631,
      "step": 4565
    },
    {
      "epoch": 0.5909958895685665,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5294522047042847,
      "eval_runtime": 13.6069,
      "eval_samples_per_second": 3.234,
      "eval_steps_per_second": 0.22,
      "step": 4565
    },
    {
      "epoch": 0.59112535197592,
      "grad_norm": 2.7763762507188328,
      "learning_rate": 5.356211364409769e-07,
      "loss": 2.5386,
      "step": 4566
    },
    {
      "epoch": 0.5912548143832734,
      "grad_norm": 2.124939975581646,
      "learning_rate": 5.35341572251543e-07,
      "loss": 2.582,
      "step": 4567
    },
    {
      "epoch": 0.5913842767906269,
      "grad_norm": 2.303041812296135,
      "learning_rate": 5.350620263766602e-07,
      "loss": 2.2463,
      "step": 4568
    },
    {
      "epoch": 0.5915137391979803,
      "grad_norm": 1.3075283894819436,
      "learning_rate": 5.347824988734316e-07,
      "loss": 2.3184,
      "step": 4569
    },
    {
      "epoch": 0.5916432016053339,
      "grad_norm": 2.53119611922096,
      "learning_rate": 5.34502989798956e-07,
      "loss": 2.3137,
      "step": 4570
    },
    {
      "epoch": 0.5916432016053339,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5280095338821411,
      "eval_runtime": 14.256,
      "eval_samples_per_second": 3.086,
      "eval_steps_per_second": 0.21,
      "step": 4570
    },
    {
      "epoch": 0.5917726640126874,
      "grad_norm": 2.0042020168297725,
      "learning_rate": 5.342234992103291e-07,
      "loss": 2.5232,
      "step": 4571
    },
    {
      "epoch": 0.5919021264200408,
      "grad_norm": 1.3431989711007388,
      "learning_rate": 5.339440271646428e-07,
      "loss": 2.3379,
      "step": 4572
    },
    {
      "epoch": 0.5920315888273943,
      "grad_norm": 2.2317047631597746,
      "learning_rate": 5.336645737189842e-07,
      "loss": 2.6182,
      "step": 4573
    },
    {
      "epoch": 0.5921610512347477,
      "grad_norm": 2.3487473612135683,
      "learning_rate": 5.333851389304377e-07,
      "loss": 2.3916,
      "step": 4574
    },
    {
      "epoch": 0.5922905136421012,
      "grad_norm": 1.3129285018836916,
      "learning_rate": 5.331057228560835e-07,
      "loss": 2.4641,
      "step": 4575
    },
    {
      "epoch": 0.5922905136421012,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.531693935394287,
      "eval_runtime": 14.9173,
      "eval_samples_per_second": 2.95,
      "eval_steps_per_second": 0.201,
      "step": 4575
    },
    {
      "epoch": 0.5924199760494546,
      "grad_norm": 1.6490118758297527,
      "learning_rate": 5.328263255529982e-07,
      "loss": 2.5322,
      "step": 4576
    },
    {
      "epoch": 0.5925494384568081,
      "grad_norm": 2.241472373951339,
      "learning_rate": 5.325469470782543e-07,
      "loss": 2.3315,
      "step": 4577
    },
    {
      "epoch": 0.5926789008641615,
      "grad_norm": 2.5842823660227574,
      "learning_rate": 5.322675874889201e-07,
      "loss": 2.6646,
      "step": 4578
    },
    {
      "epoch": 0.592808363271515,
      "grad_norm": 1.3435537354347897,
      "learning_rate": 5.31988246842061e-07,
      "loss": 2.2693,
      "step": 4579
    },
    {
      "epoch": 0.5929378256788685,
      "grad_norm": 2.8056663725240205,
      "learning_rate": 5.317089251947374e-07,
      "loss": 2.4087,
      "step": 4580
    },
    {
      "epoch": 0.5929378256788685,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5300959348678589,
      "eval_runtime": 14.056,
      "eval_samples_per_second": 3.13,
      "eval_steps_per_second": 0.213,
      "step": 4580
    },
    {
      "epoch": 0.593067288086222,
      "grad_norm": 1.9504389631172634,
      "learning_rate": 5.31429622604007e-07,
      "loss": 2.4089,
      "step": 4581
    },
    {
      "epoch": 0.5931967504935755,
      "grad_norm": 1.1873545907786829,
      "learning_rate": 5.311503391269228e-07,
      "loss": 2.2095,
      "step": 4582
    },
    {
      "epoch": 0.5933262129009289,
      "grad_norm": 1.612026875731236,
      "learning_rate": 5.308710748205341e-07,
      "loss": 2.5134,
      "step": 4583
    },
    {
      "epoch": 0.5934556753082824,
      "grad_norm": 1.5047253345107092,
      "learning_rate": 5.305918297418861e-07,
      "loss": 2.2947,
      "step": 4584
    },
    {
      "epoch": 0.5935851377156358,
      "grad_norm": 2.2305966698018795,
      "learning_rate": 5.303126039480206e-07,
      "loss": 2.5244,
      "step": 4585
    },
    {
      "epoch": 0.5935851377156358,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5287197828292847,
      "eval_runtime": 13.7206,
      "eval_samples_per_second": 3.207,
      "eval_steps_per_second": 0.219,
      "step": 4585
    },
    {
      "epoch": 0.5937146001229893,
      "grad_norm": 4.636491843576217,
      "learning_rate": 5.300333974959748e-07,
      "loss": 2.332,
      "step": 4586
    },
    {
      "epoch": 0.5938440625303427,
      "grad_norm": 3.005161868842217,
      "learning_rate": 5.297542104427826e-07,
      "loss": 2.6899,
      "step": 4587
    },
    {
      "epoch": 0.5939735249376962,
      "grad_norm": 1.4455363886980812,
      "learning_rate": 5.294750428454735e-07,
      "loss": 2.3621,
      "step": 4588
    },
    {
      "epoch": 0.5941029873450496,
      "grad_norm": 2.812861240201342,
      "learning_rate": 5.29195894761073e-07,
      "loss": 2.501,
      "step": 4589
    },
    {
      "epoch": 0.5942324497524032,
      "grad_norm": 1.6576516312922545,
      "learning_rate": 5.289167662466028e-07,
      "loss": 2.2856,
      "step": 4590
    },
    {
      "epoch": 0.5942324497524032,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5265891551971436,
      "eval_runtime": 13.9118,
      "eval_samples_per_second": 3.163,
      "eval_steps_per_second": 0.216,
      "step": 4590
    },
    {
      "epoch": 0.5943619121597566,
      "grad_norm": 1.9889810033875162,
      "learning_rate": 5.286376573590808e-07,
      "loss": 2.4609,
      "step": 4591
    },
    {
      "epoch": 0.5944913745671101,
      "grad_norm": 1.7997372337438762,
      "learning_rate": 5.283585681555203e-07,
      "loss": 2.3315,
      "step": 4592
    },
    {
      "epoch": 0.5946208369744636,
      "grad_norm": 1.0811327630514662,
      "learning_rate": 5.280794986929309e-07,
      "loss": 2.2791,
      "step": 4593
    },
    {
      "epoch": 0.594750299381817,
      "grad_norm": 2.974688795523065,
      "learning_rate": 5.278004490283188e-07,
      "loss": 3.0698,
      "step": 4594
    },
    {
      "epoch": 0.5948797617891705,
      "grad_norm": 1.6274579464824417,
      "learning_rate": 5.275214192186851e-07,
      "loss": 2.4763,
      "step": 4595
    },
    {
      "epoch": 0.5948797617891705,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5245028734207153,
      "eval_runtime": 13.8864,
      "eval_samples_per_second": 3.169,
      "eval_steps_per_second": 0.216,
      "step": 4595
    },
    {
      "epoch": 0.5950092241965239,
      "grad_norm": 1.7309548657835876,
      "learning_rate": 5.272424093210271e-07,
      "loss": 2.1794,
      "step": 4596
    },
    {
      "epoch": 0.5951386866038774,
      "grad_norm": 1.3635214500888755,
      "learning_rate": 5.269634193923383e-07,
      "loss": 2.5359,
      "step": 4597
    },
    {
      "epoch": 0.5952681490112308,
      "grad_norm": 1.8784978101975007,
      "learning_rate": 5.266844494896084e-07,
      "loss": 2.3447,
      "step": 4598
    },
    {
      "epoch": 0.5953976114185844,
      "grad_norm": 1.656027506952059,
      "learning_rate": 5.264054996698225e-07,
      "loss": 2.2454,
      "step": 4599
    },
    {
      "epoch": 0.5955270738259378,
      "grad_norm": 1.5922391202571475,
      "learning_rate": 5.261265699899619e-07,
      "loss": 2.6094,
      "step": 4600
    },
    {
      "epoch": 0.5955270738259378,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5288307666778564,
      "eval_runtime": 13.8094,
      "eval_samples_per_second": 3.186,
      "eval_steps_per_second": 0.217,
      "step": 4600
    },
    {
      "epoch": 0.5956565362332913,
      "grad_norm": 2.117081614822697,
      "learning_rate": 5.258476605070031e-07,
      "loss": 2.2891,
      "step": 4601
    },
    {
      "epoch": 0.5957859986406447,
      "grad_norm": 1.545987931517156,
      "learning_rate": 5.255687712779195e-07,
      "loss": 2.4189,
      "step": 4602
    },
    {
      "epoch": 0.5959154610479982,
      "grad_norm": 1.5694924469205487,
      "learning_rate": 5.252899023596797e-07,
      "loss": 2.4517,
      "step": 4603
    },
    {
      "epoch": 0.5960449234553516,
      "grad_norm": 1.4191781487243231,
      "learning_rate": 5.250110538092484e-07,
      "loss": 2.3979,
      "step": 4604
    },
    {
      "epoch": 0.5961743858627051,
      "grad_norm": 1.4470733821411759,
      "learning_rate": 5.247322256835862e-07,
      "loss": 2.3953,
      "step": 4605
    },
    {
      "epoch": 0.5961743858627051,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5259677171707153,
      "eval_runtime": 13.7145,
      "eval_samples_per_second": 3.208,
      "eval_steps_per_second": 0.219,
      "step": 4605
    },
    {
      "epoch": 0.5963038482700586,
      "grad_norm": 2.267121571811472,
      "learning_rate": 5.244534180396491e-07,
      "loss": 2.6289,
      "step": 4606
    },
    {
      "epoch": 0.596433310677412,
      "grad_norm": 3.695027870870121,
      "learning_rate": 5.241746309343894e-07,
      "loss": 2.1094,
      "step": 4607
    },
    {
      "epoch": 0.5965627730847655,
      "grad_norm": 1.1885729297145644,
      "learning_rate": 5.238958644247548e-07,
      "loss": 2.1128,
      "step": 4608
    },
    {
      "epoch": 0.596692235492119,
      "grad_norm": 1.4346680200384383,
      "learning_rate": 5.236171185676893e-07,
      "loss": 2.5024,
      "step": 4609
    },
    {
      "epoch": 0.5968216978994725,
      "grad_norm": 1.4999705119485447,
      "learning_rate": 5.233383934201323e-07,
      "loss": 2.3523,
      "step": 4610
    },
    {
      "epoch": 0.5968216978994725,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5244140625,
      "eval_runtime": 14.3408,
      "eval_samples_per_second": 3.068,
      "eval_steps_per_second": 0.209,
      "step": 4610
    },
    {
      "epoch": 0.5969511603068259,
      "grad_norm": 1.9085544082141734,
      "learning_rate": 5.230596890390192e-07,
      "loss": 2.5918,
      "step": 4611
    },
    {
      "epoch": 0.5970806227141794,
      "grad_norm": 2.3159003718591302,
      "learning_rate": 5.227810054812806e-07,
      "loss": 2.3501,
      "step": 4612
    },
    {
      "epoch": 0.5972100851215328,
      "grad_norm": 1.611334343104449,
      "learning_rate": 5.225023428038436e-07,
      "loss": 2.3511,
      "step": 4613
    },
    {
      "epoch": 0.5973395475288863,
      "grad_norm": 2.0769223658275995,
      "learning_rate": 5.222237010636305e-07,
      "loss": 2.3425,
      "step": 4614
    },
    {
      "epoch": 0.5974690099362397,
      "grad_norm": 2.26214875008948,
      "learning_rate": 5.219450803175597e-07,
      "loss": 2.4568,
      "step": 4615
    },
    {
      "epoch": 0.5974690099362397,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.523881435394287,
      "eval_runtime": 15.0633,
      "eval_samples_per_second": 2.921,
      "eval_steps_per_second": 0.199,
      "step": 4615
    },
    {
      "epoch": 0.5975984723435932,
      "grad_norm": 4.060852651972172,
      "learning_rate": 5.216664806225453e-07,
      "loss": 3.1616,
      "step": 4616
    },
    {
      "epoch": 0.5977279347509467,
      "grad_norm": 6.099827594108327,
      "learning_rate": 5.213879020354964e-07,
      "loss": 3.6753,
      "step": 4617
    },
    {
      "epoch": 0.5978573971583001,
      "grad_norm": 2.3847608943746015,
      "learning_rate": 5.211093446133188e-07,
      "loss": 2.4512,
      "step": 4618
    },
    {
      "epoch": 0.5979868595656537,
      "grad_norm": 1.3047068829912052,
      "learning_rate": 5.208308084129132e-07,
      "loss": 2.4102,
      "step": 4619
    },
    {
      "epoch": 0.5981163219730071,
      "grad_norm": 2.111973967418819,
      "learning_rate": 5.205522934911764e-07,
      "loss": 2.7397,
      "step": 4620
    },
    {
      "epoch": 0.5981163219730071,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5302067995071411,
      "eval_runtime": 14.4888,
      "eval_samples_per_second": 3.037,
      "eval_steps_per_second": 0.207,
      "step": 4620
    },
    {
      "epoch": 0.5982457843803606,
      "grad_norm": 1.832821757461587,
      "learning_rate": 5.202737999050008e-07,
      "loss": 2.4314,
      "step": 4621
    },
    {
      "epoch": 0.598375246787714,
      "grad_norm": 2.27613097589973,
      "learning_rate": 5.199953277112744e-07,
      "loss": 2.4407,
      "step": 4622
    },
    {
      "epoch": 0.5985047091950675,
      "grad_norm": 1.9501545744282807,
      "learning_rate": 5.197168769668804e-07,
      "loss": 2.4131,
      "step": 4623
    },
    {
      "epoch": 0.5986341716024209,
      "grad_norm": 2.674527877803433,
      "learning_rate": 5.194384477286983e-07,
      "loss": 2.5442,
      "step": 4624
    },
    {
      "epoch": 0.5987636340097744,
      "grad_norm": 2.2290724334911305,
      "learning_rate": 5.19160040053603e-07,
      "loss": 2.2948,
      "step": 4625
    },
    {
      "epoch": 0.5987636340097744,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5244362354278564,
      "eval_runtime": 15.0832,
      "eval_samples_per_second": 2.917,
      "eval_steps_per_second": 0.199,
      "step": 4625
    },
    {
      "epoch": 0.5988930964171278,
      "grad_norm": 2.004967525042569,
      "learning_rate": 5.188816539984647e-07,
      "loss": 2.6919,
      "step": 4626
    },
    {
      "epoch": 0.5990225588244813,
      "grad_norm": 1.3715985336297802,
      "learning_rate": 5.186032896201497e-07,
      "loss": 2.5889,
      "step": 4627
    },
    {
      "epoch": 0.5991520212318348,
      "grad_norm": 2.1451904940263034,
      "learning_rate": 5.183249469755192e-07,
      "loss": 2.4714,
      "step": 4628
    },
    {
      "epoch": 0.5992814836391883,
      "grad_norm": 1.3036521755728756,
      "learning_rate": 5.180466261214305e-07,
      "loss": 2.4104,
      "step": 4629
    },
    {
      "epoch": 0.5994109460465418,
      "grad_norm": 2.495382581100563,
      "learning_rate": 5.177683271147363e-07,
      "loss": 2.4368,
      "step": 4630
    },
    {
      "epoch": 0.5994109460465418,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5273659229278564,
      "eval_runtime": 14.6113,
      "eval_samples_per_second": 3.011,
      "eval_steps_per_second": 0.205,
      "step": 4630
    },
    {
      "epoch": 0.5995404084538952,
      "grad_norm": 1.6448220493446621,
      "learning_rate": 5.174900500122847e-07,
      "loss": 2.5645,
      "step": 4631
    },
    {
      "epoch": 0.5996698708612487,
      "grad_norm": 1.8138697533795158,
      "learning_rate": 5.172117948709195e-07,
      "loss": 2.6167,
      "step": 4632
    },
    {
      "epoch": 0.5997993332686021,
      "grad_norm": 1.38946825721266,
      "learning_rate": 5.169335617474802e-07,
      "loss": 2.2793,
      "step": 4633
    },
    {
      "epoch": 0.5999287956759556,
      "grad_norm": 2.6254813229972984,
      "learning_rate": 5.16655350698801e-07,
      "loss": 2.4722,
      "step": 4634
    },
    {
      "epoch": 0.600058258083309,
      "grad_norm": 1.87188404688427,
      "learning_rate": 5.163771617817126e-07,
      "loss": 2.5745,
      "step": 4635
    },
    {
      "epoch": 0.600058258083309,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5244140625,
      "eval_runtime": 14.5075,
      "eval_samples_per_second": 3.033,
      "eval_steps_per_second": 0.207,
      "step": 4635
    },
    {
      "epoch": 0.6001877204906625,
      "grad_norm": 1.838709481065822,
      "learning_rate": 5.160989950530407e-07,
      "loss": 2.6743,
      "step": 4636
    },
    {
      "epoch": 0.6003171828980159,
      "grad_norm": 1.588143706621534,
      "learning_rate": 5.158208505696062e-07,
      "loss": 2.28,
      "step": 4637
    },
    {
      "epoch": 0.6004466453053695,
      "grad_norm": 2.723775472877594,
      "learning_rate": 5.155427283882259e-07,
      "loss": 2.5955,
      "step": 4638
    },
    {
      "epoch": 0.600576107712723,
      "grad_norm": 1.4587974773226866,
      "learning_rate": 5.152646285657122e-07,
      "loss": 2.3396,
      "step": 4639
    },
    {
      "epoch": 0.6007055701200764,
      "grad_norm": 2.1468450141601196,
      "learning_rate": 5.149865511588722e-07,
      "loss": 2.3606,
      "step": 4640
    },
    {
      "epoch": 0.6007055701200764,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5236594676971436,
      "eval_runtime": 13.6146,
      "eval_samples_per_second": 3.232,
      "eval_steps_per_second": 0.22,
      "step": 4640
    },
    {
      "epoch": 0.6008350325274299,
      "grad_norm": 1.631428397490873,
      "learning_rate": 5.14708496224509e-07,
      "loss": 2.384,
      "step": 4641
    },
    {
      "epoch": 0.6009644949347833,
      "grad_norm": 2.04017470320332,
      "learning_rate": 5.144304638194207e-07,
      "loss": 2.4341,
      "step": 4642
    },
    {
      "epoch": 0.6010939573421368,
      "grad_norm": 3.08374706151955,
      "learning_rate": 5.141524540004016e-07,
      "loss": 2.592,
      "step": 4643
    },
    {
      "epoch": 0.6012234197494902,
      "grad_norm": 2.4446964845637176,
      "learning_rate": 5.138744668242406e-07,
      "loss": 2.3721,
      "step": 4644
    },
    {
      "epoch": 0.6013528821568437,
      "grad_norm": 2.140978667714405,
      "learning_rate": 5.135965023477221e-07,
      "loss": 2.5002,
      "step": 4645
    },
    {
      "epoch": 0.6013528821568437,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5265225172042847,
      "eval_runtime": 15.364,
      "eval_samples_per_second": 2.864,
      "eval_steps_per_second": 0.195,
      "step": 4645
    },
    {
      "epoch": 0.6014823445641971,
      "grad_norm": 2.2089871866958752,
      "learning_rate": 5.133185606276259e-07,
      "loss": 2.387,
      "step": 4646
    },
    {
      "epoch": 0.6016118069715506,
      "grad_norm": 1.530984281393375,
      "learning_rate": 5.130406417207275e-07,
      "loss": 2.2078,
      "step": 4647
    },
    {
      "epoch": 0.6017412693789042,
      "grad_norm": 1.3950766769353655,
      "learning_rate": 5.127627456837972e-07,
      "loss": 2.3506,
      "step": 4648
    },
    {
      "epoch": 0.6018707317862576,
      "grad_norm": 1.2981808075495236,
      "learning_rate": 5.124848725736011e-07,
      "loss": 2.3804,
      "step": 4649
    },
    {
      "epoch": 0.6020001941936111,
      "grad_norm": 1.5037332989894194,
      "learning_rate": 5.122070224469006e-07,
      "loss": 2.4231,
      "step": 4650
    },
    {
      "epoch": 0.6020001941936111,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5238147974014282,
      "eval_runtime": 14.7859,
      "eval_samples_per_second": 2.976,
      "eval_steps_per_second": 0.203,
      "step": 4650
    },
    {
      "epoch": 0.6021296566009645,
      "grad_norm": 1.6181061187170833,
      "learning_rate": 5.119291953604517e-07,
      "loss": 2.3038,
      "step": 4651
    },
    {
      "epoch": 0.602259119008318,
      "grad_norm": 1.739981653719173,
      "learning_rate": 5.116513913710065e-07,
      "loss": 2.5127,
      "step": 4652
    },
    {
      "epoch": 0.6023885814156714,
      "grad_norm": 1.894022114440706,
      "learning_rate": 5.113736105353122e-07,
      "loss": 2.366,
      "step": 4653
    },
    {
      "epoch": 0.6025180438230249,
      "grad_norm": 2.1046784772743226,
      "learning_rate": 5.110958529101109e-07,
      "loss": 2.4773,
      "step": 4654
    },
    {
      "epoch": 0.6026475062303783,
      "grad_norm": 1.503307335334289,
      "learning_rate": 5.108181185521406e-07,
      "loss": 2.4241,
      "step": 4655
    },
    {
      "epoch": 0.6026475062303783,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5253018140792847,
      "eval_runtime": 14.0511,
      "eval_samples_per_second": 3.131,
      "eval_steps_per_second": 0.214,
      "step": 4655
    },
    {
      "epoch": 0.6027769686377318,
      "grad_norm": 2.133787942458243,
      "learning_rate": 5.105404075181339e-07,
      "loss": 2.5842,
      "step": 4656
    },
    {
      "epoch": 0.6029064310450852,
      "grad_norm": 1.4029683230836063,
      "learning_rate": 5.102627198648188e-07,
      "loss": 2.3403,
      "step": 4657
    },
    {
      "epoch": 0.6030358934524388,
      "grad_norm": 1.7951964676183856,
      "learning_rate": 5.099850556489189e-07,
      "loss": 2.6255,
      "step": 4658
    },
    {
      "epoch": 0.6031653558597923,
      "grad_norm": 2.8571003538869313,
      "learning_rate": 5.097074149271525e-07,
      "loss": 2.5029,
      "step": 4659
    },
    {
      "epoch": 0.6032948182671457,
      "grad_norm": 2.496719260071145,
      "learning_rate": 5.094297977562336e-07,
      "loss": 2.5282,
      "step": 4660
    },
    {
      "epoch": 0.6032948182671457,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5237704515457153,
      "eval_runtime": 14.9916,
      "eval_samples_per_second": 2.935,
      "eval_steps_per_second": 0.2,
      "step": 4660
    },
    {
      "epoch": 0.6034242806744992,
      "grad_norm": 1.6105813367172834,
      "learning_rate": 5.091522041928711e-07,
      "loss": 2.6514,
      "step": 4661
    },
    {
      "epoch": 0.6035537430818526,
      "grad_norm": 2.362672891883585,
      "learning_rate": 5.088746342937688e-07,
      "loss": 2.4695,
      "step": 4662
    },
    {
      "epoch": 0.6036832054892061,
      "grad_norm": 2.1457448459448503,
      "learning_rate": 5.085970881156263e-07,
      "loss": 2.7498,
      "step": 4663
    },
    {
      "epoch": 0.6038126678965595,
      "grad_norm": 2.379829811618664,
      "learning_rate": 5.08319565715138e-07,
      "loss": 2.2891,
      "step": 4664
    },
    {
      "epoch": 0.603942130303913,
      "grad_norm": 2.8416147110487446,
      "learning_rate": 5.080420671489932e-07,
      "loss": 2.3105,
      "step": 4665
    },
    {
      "epoch": 0.603942130303913,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5244362354278564,
      "eval_runtime": 14.1915,
      "eval_samples_per_second": 3.1,
      "eval_steps_per_second": 0.211,
      "step": 4665
    },
    {
      "epoch": 0.6040715927112664,
      "grad_norm": 3.345384722278484,
      "learning_rate": 5.07764592473877e-07,
      "loss": 2.6643,
      "step": 4666
    },
    {
      "epoch": 0.6042010551186199,
      "grad_norm": 1.6942408309354136,
      "learning_rate": 5.074871417464691e-07,
      "loss": 2.3813,
      "step": 4667
    },
    {
      "epoch": 0.6043305175259734,
      "grad_norm": 1.702581507361869,
      "learning_rate": 5.072097150234443e-07,
      "loss": 2.4702,
      "step": 4668
    },
    {
      "epoch": 0.6044599799333269,
      "grad_norm": 1.7233448643202374,
      "learning_rate": 5.069323123614724e-07,
      "loss": 2.4658,
      "step": 4669
    },
    {
      "epoch": 0.6045894423406803,
      "grad_norm": 1.6505845428953232,
      "learning_rate": 5.066549338172191e-07,
      "loss": 2.4409,
      "step": 4670
    },
    {
      "epoch": 0.6045894423406803,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5245472192764282,
      "eval_runtime": 13.5303,
      "eval_samples_per_second": 3.252,
      "eval_steps_per_second": 0.222,
      "step": 4670
    },
    {
      "epoch": 0.6047189047480338,
      "grad_norm": 2.3037104899856513,
      "learning_rate": 5.063775794473444e-07,
      "loss": 2.4553,
      "step": 4671
    },
    {
      "epoch": 0.6048483671553873,
      "grad_norm": 1.6503649776414486,
      "learning_rate": 5.061002493085035e-07,
      "loss": 2.3635,
      "step": 4672
    },
    {
      "epoch": 0.6049778295627407,
      "grad_norm": 1.4696451757013305,
      "learning_rate": 5.058229434573464e-07,
      "loss": 2.3857,
      "step": 4673
    },
    {
      "epoch": 0.6051072919700942,
      "grad_norm": 2.1155925853249884,
      "learning_rate": 5.055456619505189e-07,
      "loss": 2.5779,
      "step": 4674
    },
    {
      "epoch": 0.6052367543774476,
      "grad_norm": 2.1112521486211593,
      "learning_rate": 5.05268404844661e-07,
      "loss": 2.3508,
      "step": 4675
    },
    {
      "epoch": 0.6052367543774476,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5249911546707153,
      "eval_runtime": 14.5312,
      "eval_samples_per_second": 3.028,
      "eval_steps_per_second": 0.206,
      "step": 4675
    },
    {
      "epoch": 0.6053662167848011,
      "grad_norm": 1.5120186966826514,
      "learning_rate": 5.049911721964085e-07,
      "loss": 2.4761,
      "step": 4676
    },
    {
      "epoch": 0.6054956791921545,
      "grad_norm": 1.6578703636929901,
      "learning_rate": 5.047139640623913e-07,
      "loss": 2.2661,
      "step": 4677
    },
    {
      "epoch": 0.6056251415995081,
      "grad_norm": 1.779035320032976,
      "learning_rate": 5.044367804992353e-07,
      "loss": 2.2852,
      "step": 4678
    },
    {
      "epoch": 0.6057546040068615,
      "grad_norm": 2.469595952228615,
      "learning_rate": 5.041596215635605e-07,
      "loss": 2.5381,
      "step": 4679
    },
    {
      "epoch": 0.605884066414215,
      "grad_norm": 1.8107849013043145,
      "learning_rate": 5.038824873119822e-07,
      "loss": 2.4197,
      "step": 4680
    },
    {
      "epoch": 0.605884066414215,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5250798463821411,
      "eval_runtime": 14.0415,
      "eval_samples_per_second": 3.134,
      "eval_steps_per_second": 0.214,
      "step": 4680
    },
    {
      "epoch": 0.6060135288215684,
      "grad_norm": 1.320772742112339,
      "learning_rate": 5.036053778011109e-07,
      "loss": 2.2725,
      "step": 4681
    },
    {
      "epoch": 0.6061429912289219,
      "grad_norm": 1.4795393559454117,
      "learning_rate": 5.033282930875516e-07,
      "loss": 2.2974,
      "step": 4682
    },
    {
      "epoch": 0.6062724536362754,
      "grad_norm": 2.079559755992505,
      "learning_rate": 5.030512332279048e-07,
      "loss": 2.4663,
      "step": 4683
    },
    {
      "epoch": 0.6064019160436288,
      "grad_norm": 2.485058863213783,
      "learning_rate": 5.027741982787654e-07,
      "loss": 2.4838,
      "step": 4684
    },
    {
      "epoch": 0.6065313784509823,
      "grad_norm": 1.6635237073267555,
      "learning_rate": 5.024971882967233e-07,
      "loss": 2.4609,
      "step": 4685
    },
    {
      "epoch": 0.6065313784509823,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5223277807235718,
      "eval_runtime": 15.008,
      "eval_samples_per_second": 2.932,
      "eval_steps_per_second": 0.2,
      "step": 4685
    },
    {
      "epoch": 0.6066608408583357,
      "grad_norm": 1.4755626621283848,
      "learning_rate": 5.022202033383635e-07,
      "loss": 2.3833,
      "step": 4686
    },
    {
      "epoch": 0.6067903032656893,
      "grad_norm": 1.7108417747180253,
      "learning_rate": 5.019432434602658e-07,
      "loss": 2.4229,
      "step": 4687
    },
    {
      "epoch": 0.6069197656730427,
      "grad_norm": 1.59787545243264,
      "learning_rate": 5.016663087190047e-07,
      "loss": 2.5488,
      "step": 4688
    },
    {
      "epoch": 0.6070492280803962,
      "grad_norm": 1.4573855717201856,
      "learning_rate": 5.013893991711505e-07,
      "loss": 2.5049,
      "step": 4689
    },
    {
      "epoch": 0.6071786904877496,
      "grad_norm": 1.6417766257399868,
      "learning_rate": 5.011125148732665e-07,
      "loss": 2.286,
      "step": 4690
    },
    {
      "epoch": 0.6071786904877496,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5223721265792847,
      "eval_runtime": 13.8892,
      "eval_samples_per_second": 3.168,
      "eval_steps_per_second": 0.216,
      "step": 4690
    },
    {
      "epoch": 0.6073081528951031,
      "grad_norm": 2.5543156760116266,
      "learning_rate": 5.008356558819126e-07,
      "loss": 2.4971,
      "step": 4691
    },
    {
      "epoch": 0.6074376153024565,
      "grad_norm": 2.3547782596763853,
      "learning_rate": 5.005588222536424e-07,
      "loss": 2.1682,
      "step": 4692
    },
    {
      "epoch": 0.60756707770981,
      "grad_norm": 2.160235813262494,
      "learning_rate": 5.002820140450053e-07,
      "loss": 2.5562,
      "step": 4693
    },
    {
      "epoch": 0.6076965401171635,
      "grad_norm": 2.1001118100678355,
      "learning_rate": 5.00005231312545e-07,
      "loss": 2.4353,
      "step": 4694
    },
    {
      "epoch": 0.6078260025245169,
      "grad_norm": 1.4496618795030938,
      "learning_rate": 4.997284741127994e-07,
      "loss": 2.4705,
      "step": 4695
    },
    {
      "epoch": 0.6078260025245169,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5223721265792847,
      "eval_runtime": 14.2468,
      "eval_samples_per_second": 3.088,
      "eval_steps_per_second": 0.211,
      "step": 4695
    },
    {
      "epoch": 0.6079554649318704,
      "grad_norm": 2.2755849782315423,
      "learning_rate": 4.99451742502302e-07,
      "loss": 2.5769,
      "step": 4696
    },
    {
      "epoch": 0.6080849273392239,
      "grad_norm": 2.265958651444216,
      "learning_rate": 4.991750365375811e-07,
      "loss": 2.3123,
      "step": 4697
    },
    {
      "epoch": 0.6082143897465774,
      "grad_norm": 1.331216176682153,
      "learning_rate": 4.988983562751593e-07,
      "loss": 2.344,
      "step": 4698
    },
    {
      "epoch": 0.6083438521539308,
      "grad_norm": 1.7095427885210395,
      "learning_rate": 4.986217017715542e-07,
      "loss": 2.301,
      "step": 4699
    },
    {
      "epoch": 0.6084733145612843,
      "grad_norm": 3.010453449503125,
      "learning_rate": 4.983450730832782e-07,
      "loss": 2.448,
      "step": 4700
    },
    {
      "epoch": 0.6084733145612843,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.520219326019287,
      "eval_runtime": 14.845,
      "eval_samples_per_second": 2.964,
      "eval_steps_per_second": 0.202,
      "step": 4700
    },
    {
      "epoch": 0.6086027769686377,
      "grad_norm": 1.9129005848885028,
      "learning_rate": 4.980684702668382e-07,
      "loss": 2.2495,
      "step": 4701
    },
    {
      "epoch": 0.6087322393759912,
      "grad_norm": 1.3336931624060584,
      "learning_rate": 4.977918933787358e-07,
      "loss": 2.1235,
      "step": 4702
    },
    {
      "epoch": 0.6088617017833446,
      "grad_norm": 1.5081497629896268,
      "learning_rate": 4.975153424754676e-07,
      "loss": 2.3838,
      "step": 4703
    },
    {
      "epoch": 0.6089911641906981,
      "grad_norm": 2.200208238409872,
      "learning_rate": 4.972388176135247e-07,
      "loss": 2.6011,
      "step": 4704
    },
    {
      "epoch": 0.6091206265980516,
      "grad_norm": 1.2959982090814168,
      "learning_rate": 4.96962318849393e-07,
      "loss": 2.3472,
      "step": 4705
    },
    {
      "epoch": 0.6091206265980516,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.52099609375,
      "eval_runtime": 14.9253,
      "eval_samples_per_second": 2.948,
      "eval_steps_per_second": 0.201,
      "step": 4705
    },
    {
      "epoch": 0.609250089005405,
      "grad_norm": 2.176592992288619,
      "learning_rate": 4.966858462395528e-07,
      "loss": 2.4204,
      "step": 4706
    },
    {
      "epoch": 0.6093795514127586,
      "grad_norm": 1.5114345155083193,
      "learning_rate": 4.964093998404794e-07,
      "loss": 2.4133,
      "step": 4707
    },
    {
      "epoch": 0.609509013820112,
      "grad_norm": 1.423788770868988,
      "learning_rate": 4.961329797086424e-07,
      "loss": 2.425,
      "step": 4708
    },
    {
      "epoch": 0.6096384762274655,
      "grad_norm": 1.4431521140185195,
      "learning_rate": 4.958565859005062e-07,
      "loss": 2.343,
      "step": 4709
    },
    {
      "epoch": 0.6097679386348189,
      "grad_norm": 1.9442781312892368,
      "learning_rate": 4.955802184725301e-07,
      "loss": 2.6077,
      "step": 4710
    },
    {
      "epoch": 0.6097679386348189,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5224387645721436,
      "eval_runtime": 14.1263,
      "eval_samples_per_second": 3.115,
      "eval_steps_per_second": 0.212,
      "step": 4710
    },
    {
      "epoch": 0.6098974010421724,
      "grad_norm": 1.6960878604346235,
      "learning_rate": 4.953038774811676e-07,
      "loss": 2.4141,
      "step": 4711
    },
    {
      "epoch": 0.6100268634495258,
      "grad_norm": 2.373923875113945,
      "learning_rate": 4.950275629828669e-07,
      "loss": 2.3877,
      "step": 4712
    },
    {
      "epoch": 0.6101563258568793,
      "grad_norm": 1.8879097568974776,
      "learning_rate": 4.947512750340709e-07,
      "loss": 2.2178,
      "step": 4713
    },
    {
      "epoch": 0.6102857882642327,
      "grad_norm": 3.109372372826221,
      "learning_rate": 4.944750136912169e-07,
      "loss": 2.3252,
      "step": 4714
    },
    {
      "epoch": 0.6104152506715862,
      "grad_norm": 1.8143025150647485,
      "learning_rate": 4.94198779010737e-07,
      "loss": 2.2766,
      "step": 4715
    },
    {
      "epoch": 0.6104152506715862,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.519486904144287,
      "eval_runtime": 14.951,
      "eval_samples_per_second": 2.943,
      "eval_steps_per_second": 0.201,
      "step": 4715
    },
    {
      "epoch": 0.6105447130789396,
      "grad_norm": 2.542847044686102,
      "learning_rate": 4.93922571049058e-07,
      "loss": 2.2988,
      "step": 4716
    },
    {
      "epoch": 0.6106741754862932,
      "grad_norm": 2.183170608361862,
      "learning_rate": 4.936463898626004e-07,
      "loss": 2.3713,
      "step": 4717
    },
    {
      "epoch": 0.6108036378936467,
      "grad_norm": 1.986950915634438,
      "learning_rate": 4.933702355077801e-07,
      "loss": 2.4187,
      "step": 4718
    },
    {
      "epoch": 0.6109331003010001,
      "grad_norm": 1.620949551176615,
      "learning_rate": 4.930941080410072e-07,
      "loss": 2.5188,
      "step": 4719
    },
    {
      "epoch": 0.6110625627083536,
      "grad_norm": 1.3108482149525855,
      "learning_rate": 4.928180075186863e-07,
      "loss": 2.5205,
      "step": 4720
    },
    {
      "epoch": 0.6110625627083536,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5209072828292847,
      "eval_runtime": 14.3563,
      "eval_samples_per_second": 3.065,
      "eval_steps_per_second": 0.209,
      "step": 4720
    },
    {
      "epoch": 0.611192025115707,
      "grad_norm": 1.259033212685834,
      "learning_rate": 4.925419339972167e-07,
      "loss": 2.2778,
      "step": 4721
    },
    {
      "epoch": 0.6113214875230605,
      "grad_norm": 1.5158302861879727,
      "learning_rate": 4.922658875329919e-07,
      "loss": 2.542,
      "step": 4722
    },
    {
      "epoch": 0.6114509499304139,
      "grad_norm": 2.760070698018847,
      "learning_rate": 4.919898681824e-07,
      "loss": 2.3582,
      "step": 4723
    },
    {
      "epoch": 0.6115804123377674,
      "grad_norm": 1.3578681654429485,
      "learning_rate": 4.917138760018234e-07,
      "loss": 2.3206,
      "step": 4724
    },
    {
      "epoch": 0.6117098747451208,
      "grad_norm": 1.585325343866118,
      "learning_rate": 4.914379110476392e-07,
      "loss": 2.3787,
      "step": 4725
    },
    {
      "epoch": 0.6117098747451208,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.522416591644287,
      "eval_runtime": 13.6319,
      "eval_samples_per_second": 3.228,
      "eval_steps_per_second": 0.22,
      "step": 4725
    },
    {
      "epoch": 0.6118393371524744,
      "grad_norm": 1.8909363781040283,
      "learning_rate": 4.91161973376219e-07,
      "loss": 2.6426,
      "step": 4726
    },
    {
      "epoch": 0.6119687995598279,
      "grad_norm": 2.0638650017909845,
      "learning_rate": 4.908860630439283e-07,
      "loss": 2.4556,
      "step": 4727
    },
    {
      "epoch": 0.6120982619671813,
      "grad_norm": 2.663084611793344,
      "learning_rate": 4.906101801071278e-07,
      "loss": 2.4624,
      "step": 4728
    },
    {
      "epoch": 0.6122277243745348,
      "grad_norm": 1.5082132492972424,
      "learning_rate": 4.903343246221718e-07,
      "loss": 2.3088,
      "step": 4729
    },
    {
      "epoch": 0.6123571867818882,
      "grad_norm": 2.1357864125512513,
      "learning_rate": 4.900584966454097e-07,
      "loss": 2.2007,
      "step": 4730
    },
    {
      "epoch": 0.6123571867818882,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5245915651321411,
      "eval_runtime": 14.5136,
      "eval_samples_per_second": 3.032,
      "eval_steps_per_second": 0.207,
      "step": 4730
    },
    {
      "epoch": 0.6124866491892417,
      "grad_norm": 2.2370026456792313,
      "learning_rate": 4.897826962331846e-07,
      "loss": 2.6523,
      "step": 4731
    },
    {
      "epoch": 0.6126161115965951,
      "grad_norm": 1.8990172088613841,
      "learning_rate": 4.895069234418347e-07,
      "loss": 2.406,
      "step": 4732
    },
    {
      "epoch": 0.6127455740039486,
      "grad_norm": 1.8626756037156218,
      "learning_rate": 4.892311783276921e-07,
      "loss": 2.7102,
      "step": 4733
    },
    {
      "epoch": 0.612875036411302,
      "grad_norm": 2.4529761174605764,
      "learning_rate": 4.88955460947083e-07,
      "loss": 2.708,
      "step": 4734
    },
    {
      "epoch": 0.6130044988186555,
      "grad_norm": 1.8866674551346743,
      "learning_rate": 4.886797713563285e-07,
      "loss": 2.4607,
      "step": 4735
    },
    {
      "epoch": 0.6130044988186555,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.520751953125,
      "eval_runtime": 15.4496,
      "eval_samples_per_second": 2.848,
      "eval_steps_per_second": 0.194,
      "step": 4735
    },
    {
      "epoch": 0.613133961226009,
      "grad_norm": 2.5905839450252643,
      "learning_rate": 4.884041096117439e-07,
      "loss": 2.6572,
      "step": 4736
    },
    {
      "epoch": 0.6132634236333625,
      "grad_norm": 1.2495388841907706,
      "learning_rate": 4.881284757696384e-07,
      "loss": 2.4167,
      "step": 4737
    },
    {
      "epoch": 0.613392886040716,
      "grad_norm": 2.2134697229289895,
      "learning_rate": 4.878528698863161e-07,
      "loss": 2.4204,
      "step": 4738
    },
    {
      "epoch": 0.6135223484480694,
      "grad_norm": 3.0943325502854524,
      "learning_rate": 4.875772920180753e-07,
      "loss": 2.543,
      "step": 4739
    },
    {
      "epoch": 0.6136518108554229,
      "grad_norm": 1.368700318488873,
      "learning_rate": 4.873017422212077e-07,
      "loss": 2.3909,
      "step": 4740
    },
    {
      "epoch": 0.6136518108554229,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5215953588485718,
      "eval_runtime": 14.3444,
      "eval_samples_per_second": 3.067,
      "eval_steps_per_second": 0.209,
      "step": 4740
    },
    {
      "epoch": 0.6137812732627763,
      "grad_norm": 1.8151115044253032,
      "learning_rate": 4.870262205520001e-07,
      "loss": 2.2769,
      "step": 4741
    },
    {
      "epoch": 0.6139107356701298,
      "grad_norm": 1.4335975396958591,
      "learning_rate": 4.867507270667337e-07,
      "loss": 2.3052,
      "step": 4742
    },
    {
      "epoch": 0.6140401980774832,
      "grad_norm": 1.666463442711422,
      "learning_rate": 4.864752618216835e-07,
      "loss": 2.5496,
      "step": 4743
    },
    {
      "epoch": 0.6141696604848367,
      "grad_norm": 1.4576366531109501,
      "learning_rate": 4.861998248731192e-07,
      "loss": 2.3169,
      "step": 4744
    },
    {
      "epoch": 0.6142991228921901,
      "grad_norm": 1.6329558559479969,
      "learning_rate": 4.859244162773036e-07,
      "loss": 2.2507,
      "step": 4745
    },
    {
      "epoch": 0.6142991228921901,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5245028734207153,
      "eval_runtime": 21.3827,
      "eval_samples_per_second": 2.058,
      "eval_steps_per_second": 0.14,
      "step": 4745
    },
    {
      "epoch": 0.6144285852995437,
      "grad_norm": 2.5510170480415386,
      "learning_rate": 4.856490360904949e-07,
      "loss": 2.5283,
      "step": 4746
    },
    {
      "epoch": 0.6145580477068971,
      "grad_norm": 1.2446692778573354,
      "learning_rate": 4.853736843689453e-07,
      "loss": 2.2031,
      "step": 4747
    },
    {
      "epoch": 0.6146875101142506,
      "grad_norm": 2.025431028056846,
      "learning_rate": 4.850983611689006e-07,
      "loss": 2.6765,
      "step": 4748
    },
    {
      "epoch": 0.614816972521604,
      "grad_norm": 1.8768243474895507,
      "learning_rate": 4.848230665466014e-07,
      "loss": 2.2834,
      "step": 4749
    },
    {
      "epoch": 0.6149464349289575,
      "grad_norm": 1.9271147056566265,
      "learning_rate": 4.845478005582822e-07,
      "loss": 2.7246,
      "step": 4750
    },
    {
      "epoch": 0.6149464349289575,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5237482786178589,
      "eval_runtime": 14.4036,
      "eval_samples_per_second": 3.055,
      "eval_steps_per_second": 0.208,
      "step": 4750
    },
    {
      "epoch": 0.615075897336311,
      "grad_norm": 1.3723666809726824,
      "learning_rate": 4.842725632601714e-07,
      "loss": 2.4421,
      "step": 4751
    },
    {
      "epoch": 0.6152053597436644,
      "grad_norm": 1.6175201050102304,
      "learning_rate": 4.839973547084921e-07,
      "loss": 2.446,
      "step": 4752
    },
    {
      "epoch": 0.6153348221510179,
      "grad_norm": 1.3440873904086832,
      "learning_rate": 4.83722174959461e-07,
      "loss": 2.386,
      "step": 4753
    },
    {
      "epoch": 0.6154642845583713,
      "grad_norm": 1.2641016375646337,
      "learning_rate": 4.834470240692892e-07,
      "loss": 2.2266,
      "step": 4754
    },
    {
      "epoch": 0.6155937469657248,
      "grad_norm": 2.7573098588141898,
      "learning_rate": 4.83171902094182e-07,
      "loss": 2.614,
      "step": 4755
    },
    {
      "epoch": 0.6155937469657248,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5215731859207153,
      "eval_runtime": 14.5065,
      "eval_samples_per_second": 3.033,
      "eval_steps_per_second": 0.207,
      "step": 4755
    },
    {
      "epoch": 0.6157232093730783,
      "grad_norm": 1.7508364413371391,
      "learning_rate": 4.828968090903385e-07,
      "loss": 2.3469,
      "step": 4756
    },
    {
      "epoch": 0.6158526717804318,
      "grad_norm": 2.4697719598778893,
      "learning_rate": 4.82621745113952e-07,
      "loss": 2.3936,
      "step": 4757
    },
    {
      "epoch": 0.6159821341877852,
      "grad_norm": 2.2831495158240886,
      "learning_rate": 4.823467102212099e-07,
      "loss": 2.4353,
      "step": 4758
    },
    {
      "epoch": 0.6161115965951387,
      "grad_norm": 1.6830887810138064,
      "learning_rate": 4.820717044682938e-07,
      "loss": 2.3291,
      "step": 4759
    },
    {
      "epoch": 0.6162410590024922,
      "grad_norm": 2.7405442520114587,
      "learning_rate": 4.817967279113788e-07,
      "loss": 2.4282,
      "step": 4760
    },
    {
      "epoch": 0.6162410590024922,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5223721265792847,
      "eval_runtime": 15.1117,
      "eval_samples_per_second": 2.912,
      "eval_steps_per_second": 0.199,
      "step": 4760
    },
    {
      "epoch": 0.6163705214098456,
      "grad_norm": 1.9482141052448505,
      "learning_rate": 4.815217806066352e-07,
      "loss": 2.4951,
      "step": 4761
    },
    {
      "epoch": 0.6164999838171991,
      "grad_norm": 1.7464143820499973,
      "learning_rate": 4.812468626102258e-07,
      "loss": 2.4041,
      "step": 4762
    },
    {
      "epoch": 0.6166294462245525,
      "grad_norm": 2.6063638923975647,
      "learning_rate": 4.809719739783085e-07,
      "loss": 2.6301,
      "step": 4763
    },
    {
      "epoch": 0.616758908631906,
      "grad_norm": 1.6603579211327602,
      "learning_rate": 4.806971147670346e-07,
      "loss": 2.3296,
      "step": 4764
    },
    {
      "epoch": 0.6168883710392594,
      "grad_norm": 2.724061764919758,
      "learning_rate": 4.804222850325501e-07,
      "loss": 2.5347,
      "step": 4765
    },
    {
      "epoch": 0.6168883710392594,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5223277807235718,
      "eval_runtime": 14.2622,
      "eval_samples_per_second": 3.085,
      "eval_steps_per_second": 0.21,
      "step": 4765
    },
    {
      "epoch": 0.617017833446613,
      "grad_norm": 2.8470569325507626,
      "learning_rate": 4.801474848309944e-07,
      "loss": 2.583,
      "step": 4766
    },
    {
      "epoch": 0.6171472958539664,
      "grad_norm": 1.6009513689529358,
      "learning_rate": 4.798727142185011e-07,
      "loss": 2.4285,
      "step": 4767
    },
    {
      "epoch": 0.6172767582613199,
      "grad_norm": 1.6912766382093678,
      "learning_rate": 4.795979732511972e-07,
      "loss": 2.1677,
      "step": 4768
    },
    {
      "epoch": 0.6174062206686733,
      "grad_norm": 1.933232509225762,
      "learning_rate": 4.793232619852046e-07,
      "loss": 2.668,
      "step": 4769
    },
    {
      "epoch": 0.6175356830760268,
      "grad_norm": 2.6713268216514887,
      "learning_rate": 4.790485804766384e-07,
      "loss": 2.626,
      "step": 4770
    },
    {
      "epoch": 0.6175356830760268,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5208407640457153,
      "eval_runtime": 21.5605,
      "eval_samples_per_second": 2.041,
      "eval_steps_per_second": 0.139,
      "step": 4770
    },
    {
      "epoch": 0.6176651454833803,
      "grad_norm": 2.987021907833521,
      "learning_rate": 4.787739287816079e-07,
      "loss": 2.2104,
      "step": 4771
    },
    {
      "epoch": 0.6177946078907337,
      "grad_norm": 1.8582896853991664,
      "learning_rate": 4.784993069562164e-07,
      "loss": 2.4246,
      "step": 4772
    },
    {
      "epoch": 0.6179240702980872,
      "grad_norm": 1.6868164150794653,
      "learning_rate": 4.782247150565608e-07,
      "loss": 2.3877,
      "step": 4773
    },
    {
      "epoch": 0.6180535327054406,
      "grad_norm": 2.7070872776097934,
      "learning_rate": 4.779501531387323e-07,
      "loss": 2.4199,
      "step": 4774
    },
    {
      "epoch": 0.6181829951127942,
      "grad_norm": 2.6531124114855094,
      "learning_rate": 4.776756212588154e-07,
      "loss": 2.3264,
      "step": 4775
    },
    {
      "epoch": 0.6181829951127942,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5209072828292847,
      "eval_runtime": 14.4832,
      "eval_samples_per_second": 3.038,
      "eval_steps_per_second": 0.207,
      "step": 4775
    },
    {
      "epoch": 0.6183124575201476,
      "grad_norm": 2.1703987969537613,
      "learning_rate": 4.774011194728891e-07,
      "loss": 2.5508,
      "step": 4776
    },
    {
      "epoch": 0.6184419199275011,
      "grad_norm": 1.558599246304356,
      "learning_rate": 4.771266478370258e-07,
      "loss": 2.3804,
      "step": 4777
    },
    {
      "epoch": 0.6185713823348545,
      "grad_norm": 2.221271758443395,
      "learning_rate": 4.768522064072921e-07,
      "loss": 2.4751,
      "step": 4778
    },
    {
      "epoch": 0.618700844742208,
      "grad_norm": 1.1691453568684558,
      "learning_rate": 4.765777952397479e-07,
      "loss": 2.0979,
      "step": 4779
    },
    {
      "epoch": 0.6188303071495614,
      "grad_norm": 2.230813251021605,
      "learning_rate": 4.763034143904474e-07,
      "loss": 2.4996,
      "step": 4780
    },
    {
      "epoch": 0.6188303071495614,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5201970338821411,
      "eval_runtime": 14.9223,
      "eval_samples_per_second": 2.949,
      "eval_steps_per_second": 0.201,
      "step": 4780
    },
    {
      "epoch": 0.6189597695569149,
      "grad_norm": 2.9986899077491795,
      "learning_rate": 4.760290639154384e-07,
      "loss": 2.5708,
      "step": 4781
    },
    {
      "epoch": 0.6190892319642683,
      "grad_norm": 1.4654181095005336,
      "learning_rate": 4.757547438707627e-07,
      "loss": 2.2988,
      "step": 4782
    },
    {
      "epoch": 0.6192186943716218,
      "grad_norm": 2.1809846064587743,
      "learning_rate": 4.7548045431245566e-07,
      "loss": 2.3057,
      "step": 4783
    },
    {
      "epoch": 0.6193481567789753,
      "grad_norm": 1.6972058371645136,
      "learning_rate": 4.7520619529654644e-07,
      "loss": 2.4062,
      "step": 4784
    },
    {
      "epoch": 0.6194776191863288,
      "grad_norm": 1.8874239416501835,
      "learning_rate": 4.7493196687905796e-07,
      "loss": 2.3154,
      "step": 4785
    },
    {
      "epoch": 0.6194776191863288,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.52001953125,
      "eval_runtime": 14.1437,
      "eval_samples_per_second": 3.111,
      "eval_steps_per_second": 0.212,
      "step": 4785
    },
    {
      "epoch": 0.6196070815936823,
      "grad_norm": 1.3553940396386084,
      "learning_rate": 4.746577691160069e-07,
      "loss": 2.1719,
      "step": 4786
    },
    {
      "epoch": 0.6197365440010357,
      "grad_norm": 2.855121805918522,
      "learning_rate": 4.743836020634038e-07,
      "loss": 2.4114,
      "step": 4787
    },
    {
      "epoch": 0.6198660064083892,
      "grad_norm": 1.5379529689841176,
      "learning_rate": 4.7410946577725286e-07,
      "loss": 2.3159,
      "step": 4788
    },
    {
      "epoch": 0.6199954688157426,
      "grad_norm": 1.6290192295483459,
      "learning_rate": 4.7383536031355224e-07,
      "loss": 2.2993,
      "step": 4789
    },
    {
      "epoch": 0.6201249312230961,
      "grad_norm": 3.0386774850661973,
      "learning_rate": 4.7356128572829297e-07,
      "loss": 2.3904,
      "step": 4790
    },
    {
      "epoch": 0.6201249312230961,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5222389698028564,
      "eval_runtime": 14.1221,
      "eval_samples_per_second": 3.116,
      "eval_steps_per_second": 0.212,
      "step": 4790
    },
    {
      "epoch": 0.6202543936304495,
      "grad_norm": 1.681077652333097,
      "learning_rate": 4.732872420774605e-07,
      "loss": 2.3577,
      "step": 4791
    },
    {
      "epoch": 0.620383856037803,
      "grad_norm": 1.7899460274118708,
      "learning_rate": 4.7301322941703406e-07,
      "loss": 2.2529,
      "step": 4792
    },
    {
      "epoch": 0.6205133184451564,
      "grad_norm": 1.739871020013475,
      "learning_rate": 4.727392478029862e-07,
      "loss": 2.5083,
      "step": 4793
    },
    {
      "epoch": 0.6206427808525099,
      "grad_norm": 1.2278170669570239,
      "learning_rate": 4.724652972912833e-07,
      "loss": 2.2422,
      "step": 4794
    },
    {
      "epoch": 0.6207722432598635,
      "grad_norm": 1.0660098313494566,
      "learning_rate": 4.7219137793788476e-07,
      "loss": 2.0244,
      "step": 4795
    },
    {
      "epoch": 0.6207722432598635,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5209072828292847,
      "eval_runtime": 13.5206,
      "eval_samples_per_second": 3.254,
      "eval_steps_per_second": 0.222,
      "step": 4795
    },
    {
      "epoch": 0.6209017056672169,
      "grad_norm": 1.9280486796276555,
      "learning_rate": 4.719174897987448e-07,
      "loss": 2.3804,
      "step": 4796
    },
    {
      "epoch": 0.6210311680745704,
      "grad_norm": 1.3476716292687931,
      "learning_rate": 4.7164363292981035e-07,
      "loss": 2.2083,
      "step": 4797
    },
    {
      "epoch": 0.6211606304819238,
      "grad_norm": 2.2605349563766723,
      "learning_rate": 4.713698073870222e-07,
      "loss": 2.4973,
      "step": 4798
    },
    {
      "epoch": 0.6212900928892773,
      "grad_norm": 1.757603252083686,
      "learning_rate": 4.7109601322631494e-07,
      "loss": 2.5037,
      "step": 4799
    },
    {
      "epoch": 0.6214195552966307,
      "grad_norm": 1.159193969446243,
      "learning_rate": 4.708222505036166e-07,
      "loss": 2.3098,
      "step": 4800
    },
    {
      "epoch": 0.6214195552966307,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5193980932235718,
      "eval_runtime": 13.1993,
      "eval_samples_per_second": 3.334,
      "eval_steps_per_second": 0.227,
      "step": 4800
    },
    {
      "epoch": 0.6215490177039842,
      "grad_norm": 2.2079613778662806,
      "learning_rate": 4.7054851927484844e-07,
      "loss": 2.4604,
      "step": 4801
    },
    {
      "epoch": 0.6216784801113376,
      "grad_norm": 1.558757051689407,
      "learning_rate": 4.7027481959592585e-07,
      "loss": 2.5024,
      "step": 4802
    },
    {
      "epoch": 0.6218079425186911,
      "grad_norm": 2.2722807922430373,
      "learning_rate": 4.7000115152275744e-07,
      "loss": 2.3936,
      "step": 4803
    },
    {
      "epoch": 0.6219374049260445,
      "grad_norm": 1.233923597380604,
      "learning_rate": 4.6972751511124555e-07,
      "loss": 2.5024,
      "step": 4804
    },
    {
      "epoch": 0.6220668673333981,
      "grad_norm": 1.8351769032087282,
      "learning_rate": 4.6945391041728587e-07,
      "loss": 2.4086,
      "step": 4805
    },
    {
      "epoch": 0.6220668673333981,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.52001953125,
      "eval_runtime": 14.0139,
      "eval_samples_per_second": 3.14,
      "eval_steps_per_second": 0.214,
      "step": 4805
    },
    {
      "epoch": 0.6221963297407516,
      "grad_norm": 1.2352334389783513,
      "learning_rate": 4.6918033749676807e-07,
      "loss": 2.4363,
      "step": 4806
    },
    {
      "epoch": 0.622325792148105,
      "grad_norm": 2.037538611892236,
      "learning_rate": 4.689067964055744e-07,
      "loss": 2.4058,
      "step": 4807
    },
    {
      "epoch": 0.6224552545554585,
      "grad_norm": 2.084213723741783,
      "learning_rate": 4.686332871995815e-07,
      "loss": 2.4011,
      "step": 4808
    },
    {
      "epoch": 0.6225847169628119,
      "grad_norm": 1.435217340892084,
      "learning_rate": 4.683598099346588e-07,
      "loss": 2.2743,
      "step": 4809
    },
    {
      "epoch": 0.6227141793701654,
      "grad_norm": 2.0467644162050798,
      "learning_rate": 4.6808636466667014e-07,
      "loss": 2.5183,
      "step": 4810
    },
    {
      "epoch": 0.6227141793701654,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5164684057235718,
      "eval_runtime": 14.4351,
      "eval_samples_per_second": 3.048,
      "eval_steps_per_second": 0.208,
      "step": 4810
    },
    {
      "epoch": 0.6228436417775188,
      "grad_norm": 1.7589953915309569,
      "learning_rate": 4.6781295145147223e-07,
      "loss": 2.3408,
      "step": 4811
    },
    {
      "epoch": 0.6229731041848723,
      "grad_norm": 2.2787829702733955,
      "learning_rate": 4.6753957034491477e-07,
      "loss": 2.512,
      "step": 4812
    },
    {
      "epoch": 0.6231025665922257,
      "grad_norm": 2.576545186007895,
      "learning_rate": 4.6726622140284144e-07,
      "loss": 2.7393,
      "step": 4813
    },
    {
      "epoch": 0.6232320289995793,
      "grad_norm": 1.9079266224686722,
      "learning_rate": 4.6699290468108974e-07,
      "loss": 2.708,
      "step": 4814
    },
    {
      "epoch": 0.6233614914069328,
      "grad_norm": 2.5384012167468795,
      "learning_rate": 4.667196202354898e-07,
      "loss": 2.3809,
      "step": 4815
    },
    {
      "epoch": 0.6233614914069328,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5188876390457153,
      "eval_runtime": 14.6165,
      "eval_samples_per_second": 3.01,
      "eval_steps_per_second": 0.205,
      "step": 4815
    },
    {
      "epoch": 0.6234909538142862,
      "grad_norm": 1.8016570878633815,
      "learning_rate": 4.6644636812186576e-07,
      "loss": 2.5266,
      "step": 4816
    },
    {
      "epoch": 0.6236204162216397,
      "grad_norm": 1.4780838002469523,
      "learning_rate": 4.661731483960348e-07,
      "loss": 2.3491,
      "step": 4817
    },
    {
      "epoch": 0.6237498786289931,
      "grad_norm": 3.041106649985947,
      "learning_rate": 4.658999611138075e-07,
      "loss": 2.3999,
      "step": 4818
    },
    {
      "epoch": 0.6238793410363466,
      "grad_norm": 2.8284981027694616,
      "learning_rate": 4.6562680633098795e-07,
      "loss": 2.3655,
      "step": 4819
    },
    {
      "epoch": 0.6240088034437,
      "grad_norm": 2.3524105281768466,
      "learning_rate": 4.6535368410337343e-07,
      "loss": 2.3608,
      "step": 4820
    },
    {
      "epoch": 0.6240088034437,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.520263671875,
      "eval_runtime": 15.1054,
      "eval_samples_per_second": 2.913,
      "eval_steps_per_second": 0.199,
      "step": 4820
    },
    {
      "epoch": 0.6241382658510535,
      "grad_norm": 1.9015511151215727,
      "learning_rate": 4.6508059448675496e-07,
      "loss": 2.4133,
      "step": 4821
    },
    {
      "epoch": 0.6242677282584069,
      "grad_norm": 3.860684255110897,
      "learning_rate": 4.6480753753691656e-07,
      "loss": 2.7827,
      "step": 4822
    },
    {
      "epoch": 0.6243971906657604,
      "grad_norm": 3.5419426782611447,
      "learning_rate": 4.6453451330963544e-07,
      "loss": 2.4639,
      "step": 4823
    },
    {
      "epoch": 0.624526653073114,
      "grad_norm": 2.119160921924202,
      "learning_rate": 4.642615218606825e-07,
      "loss": 2.4426,
      "step": 4824
    },
    {
      "epoch": 0.6246561154804674,
      "grad_norm": 1.1694802042887913,
      "learning_rate": 4.639885632458216e-07,
      "loss": 2.3538,
      "step": 4825
    },
    {
      "epoch": 0.6246561154804674,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5216175317764282,
      "eval_runtime": 13.9388,
      "eval_samples_per_second": 3.157,
      "eval_steps_per_second": 0.215,
      "step": 4825
    },
    {
      "epoch": 0.6247855778878209,
      "grad_norm": 1.6075606763313617,
      "learning_rate": 4.637156375208103e-07,
      "loss": 2.4294,
      "step": 4826
    },
    {
      "epoch": 0.6249150402951743,
      "grad_norm": 2.3912746323611125,
      "learning_rate": 4.634427447413992e-07,
      "loss": 2.571,
      "step": 4827
    },
    {
      "epoch": 0.6250445027025278,
      "grad_norm": 2.0187086760074844,
      "learning_rate": 4.631698849633322e-07,
      "loss": 2.3652,
      "step": 4828
    },
    {
      "epoch": 0.6251739651098812,
      "grad_norm": 2.267014675992171,
      "learning_rate": 4.6289705824234615e-07,
      "loss": 2.5625,
      "step": 4829
    },
    {
      "epoch": 0.6253034275172347,
      "grad_norm": 2.434189885511156,
      "learning_rate": 4.6262426463417167e-07,
      "loss": 2.3181,
      "step": 4830
    },
    {
      "epoch": 0.6253034275172347,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5201305150985718,
      "eval_runtime": 13.8735,
      "eval_samples_per_second": 3.172,
      "eval_steps_per_second": 0.216,
      "step": 4830
    },
    {
      "epoch": 0.6254328899245881,
      "grad_norm": 2.1247362090535042,
      "learning_rate": 4.623515041945323e-07,
      "loss": 2.4526,
      "step": 4831
    },
    {
      "epoch": 0.6255623523319416,
      "grad_norm": 2.494096550000849,
      "learning_rate": 4.620787769791448e-07,
      "loss": 2.6392,
      "step": 4832
    },
    {
      "epoch": 0.625691814739295,
      "grad_norm": 1.3406017798921561,
      "learning_rate": 4.6180608304371983e-07,
      "loss": 2.2561,
      "step": 4833
    },
    {
      "epoch": 0.6258212771466486,
      "grad_norm": 4.4356331579398,
      "learning_rate": 4.615334224439599e-07,
      "loss": 2.759,
      "step": 4834
    },
    {
      "epoch": 0.625950739554002,
      "grad_norm": 1.912969323175227,
      "learning_rate": 4.6126079523556173e-07,
      "loss": 2.5732,
      "step": 4835
    },
    {
      "epoch": 0.625950739554002,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5207741260528564,
      "eval_runtime": 15.1412,
      "eval_samples_per_second": 2.906,
      "eval_steps_per_second": 0.198,
      "step": 4835
    },
    {
      "epoch": 0.6260802019613555,
      "grad_norm": 1.8638434461433748,
      "learning_rate": 4.609882014742148e-07,
      "loss": 2.4902,
      "step": 4836
    },
    {
      "epoch": 0.626209664368709,
      "grad_norm": 1.6006112483697674,
      "learning_rate": 4.607156412156023e-07,
      "loss": 2.3518,
      "step": 4837
    },
    {
      "epoch": 0.6263391267760624,
      "grad_norm": 3.882380563014391,
      "learning_rate": 4.6044311451539994e-07,
      "loss": 2.7146,
      "step": 4838
    },
    {
      "epoch": 0.6264685891834159,
      "grad_norm": 1.344442264192774,
      "learning_rate": 4.6017062142927714e-07,
      "loss": 2.3457,
      "step": 4839
    },
    {
      "epoch": 0.6265980515907693,
      "grad_norm": 1.1417007193096187,
      "learning_rate": 4.598981620128955e-07,
      "loss": 2.3555,
      "step": 4840
    },
    {
      "epoch": 0.6265980515907693,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.521484375,
      "eval_runtime": 15.089,
      "eval_samples_per_second": 2.916,
      "eval_steps_per_second": 0.199,
      "step": 4840
    },
    {
      "epoch": 0.6267275139981228,
      "grad_norm": 1.4621411876346493,
      "learning_rate": 4.5962573632191073e-07,
      "loss": 2.3552,
      "step": 4841
    },
    {
      "epoch": 0.6268569764054762,
      "grad_norm": 1.9302695523348672,
      "learning_rate": 4.593533444119714e-07,
      "loss": 2.4299,
      "step": 4842
    },
    {
      "epoch": 0.6269864388128297,
      "grad_norm": 2.5896879199647973,
      "learning_rate": 4.59080986338719e-07,
      "loss": 2.5593,
      "step": 4843
    },
    {
      "epoch": 0.6271159012201832,
      "grad_norm": 2.0031093227754937,
      "learning_rate": 4.588086621577882e-07,
      "loss": 2.5408,
      "step": 4844
    },
    {
      "epoch": 0.6272453636275367,
      "grad_norm": 2.0309483538022555,
      "learning_rate": 4.585363719248066e-07,
      "loss": 2.3303,
      "step": 4845
    },
    {
      "epoch": 0.6272453636275367,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.519975185394287,
      "eval_runtime": 15.0488,
      "eval_samples_per_second": 2.924,
      "eval_steps_per_second": 0.199,
      "step": 4845
    },
    {
      "epoch": 0.6273748260348901,
      "grad_norm": 1.470026964245225,
      "learning_rate": 4.582641156953951e-07,
      "loss": 2.4177,
      "step": 4846
    },
    {
      "epoch": 0.6275042884422436,
      "grad_norm": 1.6707857748702866,
      "learning_rate": 4.579918935251676e-07,
      "loss": 2.2476,
      "step": 4847
    },
    {
      "epoch": 0.627633750849597,
      "grad_norm": 1.4629334567991232,
      "learning_rate": 4.577197054697309e-07,
      "loss": 2.2788,
      "step": 4848
    },
    {
      "epoch": 0.6277632132569505,
      "grad_norm": 2.0432469102657667,
      "learning_rate": 4.5744755158468505e-07,
      "loss": 2.3682,
      "step": 4849
    },
    {
      "epoch": 0.627892675664304,
      "grad_norm": 1.6376989615019812,
      "learning_rate": 4.57175431925623e-07,
      "loss": 2.2832,
      "step": 4850
    },
    {
      "epoch": 0.627892675664304,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.522261142730713,
      "eval_runtime": 14.7608,
      "eval_samples_per_second": 2.981,
      "eval_steps_per_second": 0.203,
      "step": 4850
    },
    {
      "epoch": 0.6280221380716574,
      "grad_norm": 1.3450595916188441,
      "learning_rate": 4.5690334654813046e-07,
      "loss": 2.2729,
      "step": 4851
    },
    {
      "epoch": 0.6281516004790109,
      "grad_norm": 2.08195208996776,
      "learning_rate": 4.566312955077866e-07,
      "loss": 2.5872,
      "step": 4852
    },
    {
      "epoch": 0.6282810628863643,
      "grad_norm": 2.5382215092308305,
      "learning_rate": 4.5635927886016325e-07,
      "loss": 2.4219,
      "step": 4853
    },
    {
      "epoch": 0.6284105252937179,
      "grad_norm": 2.8375273261736185,
      "learning_rate": 4.560872966608255e-07,
      "loss": 2.6938,
      "step": 4854
    },
    {
      "epoch": 0.6285399877010713,
      "grad_norm": 1.4412592849440522,
      "learning_rate": 4.5581534896533087e-07,
      "loss": 2.1096,
      "step": 4855
    },
    {
      "epoch": 0.6285399877010713,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5201083421707153,
      "eval_runtime": 14.7195,
      "eval_samples_per_second": 2.989,
      "eval_steps_per_second": 0.204,
      "step": 4855
    },
    {
      "epoch": 0.6286694501084248,
      "grad_norm": 1.3479452564618954,
      "learning_rate": 4.5554343582923084e-07,
      "loss": 2.0356,
      "step": 4856
    },
    {
      "epoch": 0.6287989125157782,
      "grad_norm": 2.9529239607144957,
      "learning_rate": 4.5527155730806854e-07,
      "loss": 2.563,
      "step": 4857
    },
    {
      "epoch": 0.6289283749231317,
      "grad_norm": 2.0425900307853304,
      "learning_rate": 4.5499971345738096e-07,
      "loss": 2.4158,
      "step": 4858
    },
    {
      "epoch": 0.6290578373304851,
      "grad_norm": 1.7757806699839356,
      "learning_rate": 4.547279043326974e-07,
      "loss": 2.4375,
      "step": 4859
    },
    {
      "epoch": 0.6291872997378386,
      "grad_norm": 1.6479052189831416,
      "learning_rate": 4.5445612998954084e-07,
      "loss": 2.4028,
      "step": 4860
    },
    {
      "epoch": 0.6291872997378386,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5215953588485718,
      "eval_runtime": 14.4909,
      "eval_samples_per_second": 3.036,
      "eval_steps_per_second": 0.207,
      "step": 4860
    },
    {
      "epoch": 0.629316762145192,
      "grad_norm": 1.4078552494197605,
      "learning_rate": 4.5418439048342674e-07,
      "loss": 2.5203,
      "step": 4861
    },
    {
      "epoch": 0.6294462245525455,
      "grad_norm": 4.175734639026459,
      "learning_rate": 4.5391268586986294e-07,
      "loss": 2.3047,
      "step": 4862
    },
    {
      "epoch": 0.6295756869598991,
      "grad_norm": 2.8345499193118435,
      "learning_rate": 4.536410162043506e-07,
      "loss": 2.5217,
      "step": 4863
    },
    {
      "epoch": 0.6297051493672525,
      "grad_norm": 1.8386520517259233,
      "learning_rate": 4.5336938154238423e-07,
      "loss": 2.7283,
      "step": 4864
    },
    {
      "epoch": 0.629834611774606,
      "grad_norm": 2.119911620922354,
      "learning_rate": 4.530977819394504e-07,
      "loss": 2.561,
      "step": 4865
    },
    {
      "epoch": 0.629834611774606,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.522904872894287,
      "eval_runtime": 14.09,
      "eval_samples_per_second": 3.123,
      "eval_steps_per_second": 0.213,
      "step": 4865
    },
    {
      "epoch": 0.6299640741819594,
      "grad_norm": 4.370710963730192,
      "learning_rate": 4.5282621745102895e-07,
      "loss": 2.8455,
      "step": 4866
    },
    {
      "epoch": 0.6300935365893129,
      "grad_norm": 1.2519320279493165,
      "learning_rate": 4.525546881325925e-07,
      "loss": 2.2441,
      "step": 4867
    },
    {
      "epoch": 0.6302229989966663,
      "grad_norm": 2.0071450680600056,
      "learning_rate": 4.5228319403960617e-07,
      "loss": 2.1821,
      "step": 4868
    },
    {
      "epoch": 0.6303524614040198,
      "grad_norm": 1.8092020055640403,
      "learning_rate": 4.5201173522752827e-07,
      "loss": 2.7168,
      "step": 4869
    },
    {
      "epoch": 0.6304819238113732,
      "grad_norm": 2.3422770266869737,
      "learning_rate": 4.517403117518097e-07,
      "loss": 2.3696,
      "step": 4870
    },
    {
      "epoch": 0.6304819238113732,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.522904872894287,
      "eval_runtime": 14.8162,
      "eval_samples_per_second": 2.97,
      "eval_steps_per_second": 0.202,
      "step": 4870
    },
    {
      "epoch": 0.6306113862187267,
      "grad_norm": 1.6612477468068756,
      "learning_rate": 4.5146892366789444e-07,
      "loss": 2.2202,
      "step": 4871
    },
    {
      "epoch": 0.6307408486260802,
      "grad_norm": 1.9041070693228226,
      "learning_rate": 4.511975710312189e-07,
      "loss": 2.4409,
      "step": 4872
    },
    {
      "epoch": 0.6308703110334337,
      "grad_norm": 1.6135515238372142,
      "learning_rate": 4.50926253897212e-07,
      "loss": 2.3828,
      "step": 4873
    },
    {
      "epoch": 0.6309997734407872,
      "grad_norm": 2.5766874356418836,
      "learning_rate": 4.5065497232129623e-07,
      "loss": 2.5208,
      "step": 4874
    },
    {
      "epoch": 0.6311292358481406,
      "grad_norm": 1.3072003413914515,
      "learning_rate": 4.5038372635888617e-07,
      "loss": 2.3862,
      "step": 4875
    },
    {
      "epoch": 0.6311292358481406,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5228604078292847,
      "eval_runtime": 13.4608,
      "eval_samples_per_second": 3.269,
      "eval_steps_per_second": 0.223,
      "step": 4875
    },
    {
      "epoch": 0.6312586982554941,
      "grad_norm": 1.1163169241071866,
      "learning_rate": 4.501125160653892e-07,
      "loss": 2.3691,
      "step": 4876
    },
    {
      "epoch": 0.6313881606628475,
      "grad_norm": 1.7738073788940885,
      "learning_rate": 4.498413414962057e-07,
      "loss": 2.3398,
      "step": 4877
    },
    {
      "epoch": 0.631517623070201,
      "grad_norm": 1.7532421085884224,
      "learning_rate": 4.495702027067286e-07,
      "loss": 2.3523,
      "step": 4878
    },
    {
      "epoch": 0.6316470854775544,
      "grad_norm": 1.3631621761569281,
      "learning_rate": 4.4929909975234334e-07,
      "loss": 2.2739,
      "step": 4879
    },
    {
      "epoch": 0.6317765478849079,
      "grad_norm": 1.23106790159973,
      "learning_rate": 4.490280326884282e-07,
      "loss": 2.2249,
      "step": 4880
    },
    {
      "epoch": 0.6317765478849079,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.520063877105713,
      "eval_runtime": 14.1868,
      "eval_samples_per_second": 3.101,
      "eval_steps_per_second": 0.211,
      "step": 4880
    },
    {
      "epoch": 0.6319060102922613,
      "grad_norm": 1.9067979784124736,
      "learning_rate": 4.487570015703542e-07,
      "loss": 2.5671,
      "step": 4881
    },
    {
      "epoch": 0.6320354726996148,
      "grad_norm": 1.4096318787599522,
      "learning_rate": 4.4848600645348474e-07,
      "loss": 2.375,
      "step": 4882
    },
    {
      "epoch": 0.6321649351069684,
      "grad_norm": 1.3021306240637123,
      "learning_rate": 4.4821504739317666e-07,
      "loss": 2.3713,
      "step": 4883
    },
    {
      "epoch": 0.6322943975143218,
      "grad_norm": 2.075831541558824,
      "learning_rate": 4.4794412444477817e-07,
      "loss": 2.4082,
      "step": 4884
    },
    {
      "epoch": 0.6324238599216753,
      "grad_norm": 2.5741685033592425,
      "learning_rate": 4.47673237663631e-07,
      "loss": 2.5193,
      "step": 4885
    },
    {
      "epoch": 0.6324238599216753,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5221946239471436,
      "eval_runtime": 14.7117,
      "eval_samples_per_second": 2.991,
      "eval_steps_per_second": 0.204,
      "step": 4885
    },
    {
      "epoch": 0.6325533223290287,
      "grad_norm": 1.5109783677818003,
      "learning_rate": 4.4740238710506917e-07,
      "loss": 2.3381,
      "step": 4886
    },
    {
      "epoch": 0.6326827847363822,
      "grad_norm": 2.416775206314579,
      "learning_rate": 4.4713157282441965e-07,
      "loss": 2.3679,
      "step": 4887
    },
    {
      "epoch": 0.6328122471437356,
      "grad_norm": 3.028285312017075,
      "learning_rate": 4.468607948770015e-07,
      "loss": 2.5254,
      "step": 4888
    },
    {
      "epoch": 0.6329417095510891,
      "grad_norm": 1.644037736028637,
      "learning_rate": 4.4659005331812706e-07,
      "loss": 2.3445,
      "step": 4889
    },
    {
      "epoch": 0.6330711719584425,
      "grad_norm": 1.8347472144076518,
      "learning_rate": 4.4631934820310014e-07,
      "loss": 2.3625,
      "step": 4890
    },
    {
      "epoch": 0.6330711719584425,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5208407640457153,
      "eval_runtime": 14.3911,
      "eval_samples_per_second": 3.057,
      "eval_steps_per_second": 0.208,
      "step": 4890
    },
    {
      "epoch": 0.633200634365796,
      "grad_norm": 1.484978945256261,
      "learning_rate": 4.4604867958721806e-07,
      "loss": 2.4421,
      "step": 4891
    },
    {
      "epoch": 0.6333300967731494,
      "grad_norm": 1.9952727359984612,
      "learning_rate": 4.4577804752577037e-07,
      "loss": 2.4604,
      "step": 4892
    },
    {
      "epoch": 0.633459559180503,
      "grad_norm": 2.5379883235014344,
      "learning_rate": 4.4550745207403914e-07,
      "loss": 2.5916,
      "step": 4893
    },
    {
      "epoch": 0.6335890215878565,
      "grad_norm": 2.0554329518908108,
      "learning_rate": 4.45236893287299e-07,
      "loss": 2.4639,
      "step": 4894
    },
    {
      "epoch": 0.6337184839952099,
      "grad_norm": 1.1657284365021892,
      "learning_rate": 4.4496637122081716e-07,
      "loss": 2.2905,
      "step": 4895
    },
    {
      "epoch": 0.6337184839952099,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5194202661514282,
      "eval_runtime": 13.593,
      "eval_samples_per_second": 3.237,
      "eval_steps_per_second": 0.221,
      "step": 4895
    },
    {
      "epoch": 0.6338479464025634,
      "grad_norm": 1.6218614425099527,
      "learning_rate": 4.446958859298529e-07,
      "loss": 2.563,
      "step": 4896
    },
    {
      "epoch": 0.6339774088099168,
      "grad_norm": 2.9661731196381607,
      "learning_rate": 4.4442543746965863e-07,
      "loss": 2.4685,
      "step": 4897
    },
    {
      "epoch": 0.6341068712172703,
      "grad_norm": 4.202311927808688,
      "learning_rate": 4.441550258954788e-07,
      "loss": 2.8179,
      "step": 4898
    },
    {
      "epoch": 0.6342363336246237,
      "grad_norm": 1.1573631384110417,
      "learning_rate": 4.4388465126255047e-07,
      "loss": 2.2051,
      "step": 4899
    },
    {
      "epoch": 0.6343657960319772,
      "grad_norm": 2.923240870348519,
      "learning_rate": 4.436143136261033e-07,
      "loss": 2.4626,
      "step": 4900
    },
    {
      "epoch": 0.6343657960319772,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5236594676971436,
      "eval_runtime": 14.8005,
      "eval_samples_per_second": 2.973,
      "eval_steps_per_second": 0.203,
      "step": 4900
    },
    {
      "epoch": 0.6344952584393306,
      "grad_norm": 1.4021242530710845,
      "learning_rate": 4.4334401304135903e-07,
      "loss": 2.2727,
      "step": 4901
    },
    {
      "epoch": 0.6346247208466842,
      "grad_norm": 1.9417369598608372,
      "learning_rate": 4.4307374956353203e-07,
      "loss": 2.6392,
      "step": 4902
    },
    {
      "epoch": 0.6347541832540377,
      "grad_norm": 2.3763954857292533,
      "learning_rate": 4.4280352324782923e-07,
      "loss": 2.3638,
      "step": 4903
    },
    {
      "epoch": 0.6348836456613911,
      "grad_norm": 1.3421804016438619,
      "learning_rate": 4.425333341494495e-07,
      "loss": 2.3342,
      "step": 4904
    },
    {
      "epoch": 0.6350131080687446,
      "grad_norm": 2.8811795394829733,
      "learning_rate": 4.422631823235849e-07,
      "loss": 2.6611,
      "step": 4905
    },
    {
      "epoch": 0.6350131080687446,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5207741260528564,
      "eval_runtime": 13.8131,
      "eval_samples_per_second": 3.185,
      "eval_steps_per_second": 0.217,
      "step": 4905
    },
    {
      "epoch": 0.635142570476098,
      "grad_norm": 2.8304521690964584,
      "learning_rate": 4.4199306782541954e-07,
      "loss": 2.4204,
      "step": 4906
    },
    {
      "epoch": 0.6352720328834515,
      "grad_norm": 2.4503525274064804,
      "learning_rate": 4.4172299071012914e-07,
      "loss": 2.5066,
      "step": 4907
    },
    {
      "epoch": 0.6354014952908049,
      "grad_norm": 2.4271166400527906,
      "learning_rate": 4.414529510328826e-07,
      "loss": 2.7051,
      "step": 4908
    },
    {
      "epoch": 0.6355309576981584,
      "grad_norm": 3.8373859933626724,
      "learning_rate": 4.411829488488413e-07,
      "loss": 2.5947,
      "step": 4909
    },
    {
      "epoch": 0.6356604201055118,
      "grad_norm": 1.6524350019024767,
      "learning_rate": 4.4091298421315846e-07,
      "loss": 2.4502,
      "step": 4910
    },
    {
      "epoch": 0.6356604201055118,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5230380296707153,
      "eval_runtime": 14.8343,
      "eval_samples_per_second": 2.966,
      "eval_steps_per_second": 0.202,
      "step": 4910
    },
    {
      "epoch": 0.6357898825128653,
      "grad_norm": 1.658254507787194,
      "learning_rate": 4.406430571809801e-07,
      "loss": 2.3638,
      "step": 4911
    },
    {
      "epoch": 0.6359193449202188,
      "grad_norm": 4.302768239573851,
      "learning_rate": 4.4037316780744376e-07,
      "loss": 2.5718,
      "step": 4912
    },
    {
      "epoch": 0.6360488073275723,
      "grad_norm": 1.571992586524972,
      "learning_rate": 4.401033161476802e-07,
      "loss": 2.3667,
      "step": 4913
    },
    {
      "epoch": 0.6361782697349257,
      "grad_norm": 1.206824624194397,
      "learning_rate": 4.3983350225681206e-07,
      "loss": 2.3049,
      "step": 4914
    },
    {
      "epoch": 0.6363077321422792,
      "grad_norm": 2.5301002294313037,
      "learning_rate": 4.3956372618995414e-07,
      "loss": 2.3196,
      "step": 4915
    },
    {
      "epoch": 0.6363077321422792,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5236594676971436,
      "eval_runtime": 14.1277,
      "eval_samples_per_second": 3.114,
      "eval_steps_per_second": 0.212,
      "step": 4915
    },
    {
      "epoch": 0.6364371945496327,
      "grad_norm": 2.390147937707176,
      "learning_rate": 4.392939880022138e-07,
      "loss": 2.5623,
      "step": 4916
    },
    {
      "epoch": 0.6365666569569861,
      "grad_norm": 1.5702398438911267,
      "learning_rate": 4.390242877486907e-07,
      "loss": 2.3467,
      "step": 4917
    },
    {
      "epoch": 0.6366961193643396,
      "grad_norm": 2.076575848619408,
      "learning_rate": 4.3875462548447614e-07,
      "loss": 2.1763,
      "step": 4918
    },
    {
      "epoch": 0.636825581771693,
      "grad_norm": 1.5970703942136362,
      "learning_rate": 4.3848500126465436e-07,
      "loss": 2.5168,
      "step": 4919
    },
    {
      "epoch": 0.6369550441790465,
      "grad_norm": 2.244949992435683,
      "learning_rate": 4.382154151443016e-07,
      "loss": 2.4556,
      "step": 4920
    },
    {
      "epoch": 0.6369550441790465,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.522993564605713,
      "eval_runtime": 15.8074,
      "eval_samples_per_second": 2.784,
      "eval_steps_per_second": 0.19,
      "step": 4920
    },
    {
      "epoch": 0.6370845065863999,
      "grad_norm": 1.8530543422427423,
      "learning_rate": 4.3794586717848627e-07,
      "loss": 2.2808,
      "step": 4921
    },
    {
      "epoch": 0.6372139689937535,
      "grad_norm": 2.5359321422104766,
      "learning_rate": 4.3767635742226906e-07,
      "loss": 2.2993,
      "step": 4922
    },
    {
      "epoch": 0.6373434314011069,
      "grad_norm": 1.920511937105126,
      "learning_rate": 4.374068859307026e-07,
      "loss": 2.4138,
      "step": 4923
    },
    {
      "epoch": 0.6374728938084604,
      "grad_norm": 2.3511951860651803,
      "learning_rate": 4.3713745275883207e-07,
      "loss": 2.6345,
      "step": 4924
    },
    {
      "epoch": 0.6376023562158138,
      "grad_norm": 2.26202948333815,
      "learning_rate": 4.3686805796169465e-07,
      "loss": 2.2808,
      "step": 4925
    },
    {
      "epoch": 0.6376023562158138,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5207741260528564,
      "eval_runtime": 14.3424,
      "eval_samples_per_second": 3.068,
      "eval_steps_per_second": 0.209,
      "step": 4925
    },
    {
      "epoch": 0.6377318186231673,
      "grad_norm": 1.635161546432112,
      "learning_rate": 4.365987015943196e-07,
      "loss": 2.5249,
      "step": 4926
    },
    {
      "epoch": 0.6378612810305208,
      "grad_norm": 1.777679941754722,
      "learning_rate": 4.363293837117283e-07,
      "loss": 2.3967,
      "step": 4927
    },
    {
      "epoch": 0.6379907434378742,
      "grad_norm": 2.956598113073973,
      "learning_rate": 4.36060104368935e-07,
      "loss": 2.5676,
      "step": 4928
    },
    {
      "epoch": 0.6381202058452277,
      "grad_norm": 4.4859268668379455,
      "learning_rate": 4.3579086362094473e-07,
      "loss": 2.3379,
      "step": 4929
    },
    {
      "epoch": 0.6382496682525811,
      "grad_norm": 1.6794487436934151,
      "learning_rate": 4.3552166152275566e-07,
      "loss": 2.3805,
      "step": 4930
    },
    {
      "epoch": 0.6382496682525811,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5193980932235718,
      "eval_runtime": 14.1138,
      "eval_samples_per_second": 3.118,
      "eval_steps_per_second": 0.213,
      "step": 4930
    },
    {
      "epoch": 0.6383791306599346,
      "grad_norm": 2.0779766566999416,
      "learning_rate": 4.3525249812935765e-07,
      "loss": 2.3147,
      "step": 4931
    },
    {
      "epoch": 0.6385085930672881,
      "grad_norm": 1.0951967491804933,
      "learning_rate": 4.34983373495733e-07,
      "loss": 2.2136,
      "step": 4932
    },
    {
      "epoch": 0.6386380554746416,
      "grad_norm": 1.5708383255210587,
      "learning_rate": 4.3471428767685587e-07,
      "loss": 2.5371,
      "step": 4933
    },
    {
      "epoch": 0.638767517881995,
      "grad_norm": 2.083396496913675,
      "learning_rate": 4.344452407276926e-07,
      "loss": 2.6357,
      "step": 4934
    },
    {
      "epoch": 0.6388969802893485,
      "grad_norm": 1.5939262098479803,
      "learning_rate": 4.341762327032009e-07,
      "loss": 2.2949,
      "step": 4935
    },
    {
      "epoch": 0.6388969802893485,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5222389698028564,
      "eval_runtime": 14.0285,
      "eval_samples_per_second": 3.136,
      "eval_steps_per_second": 0.214,
      "step": 4935
    },
    {
      "epoch": 0.639026442696702,
      "grad_norm": 1.3598935016189486,
      "learning_rate": 4.3390726365833166e-07,
      "loss": 2.7097,
      "step": 4936
    },
    {
      "epoch": 0.6391559051040554,
      "grad_norm": 1.0933249256915052,
      "learning_rate": 4.3363833364802695e-07,
      "loss": 2.2151,
      "step": 4937
    },
    {
      "epoch": 0.6392853675114089,
      "grad_norm": 2.0923128993160884,
      "learning_rate": 4.333694427272214e-07,
      "loss": 2.377,
      "step": 4938
    },
    {
      "epoch": 0.6394148299187623,
      "grad_norm": 3.5353005267093045,
      "learning_rate": 4.331005909508415e-07,
      "loss": 2.6572,
      "step": 4939
    },
    {
      "epoch": 0.6395442923261158,
      "grad_norm": 2.795589913100964,
      "learning_rate": 4.328317783738054e-07,
      "loss": 2.3752,
      "step": 4940
    },
    {
      "epoch": 0.6395442923261158,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5207297801971436,
      "eval_runtime": 14.966,
      "eval_samples_per_second": 2.94,
      "eval_steps_per_second": 0.2,
      "step": 4940
    },
    {
      "epoch": 0.6396737547334693,
      "grad_norm": 2.285371083908442,
      "learning_rate": 4.325630050510236e-07,
      "loss": 2.4373,
      "step": 4941
    },
    {
      "epoch": 0.6398032171408228,
      "grad_norm": 1.3942275189089453,
      "learning_rate": 4.322942710373986e-07,
      "loss": 2.3584,
      "step": 4942
    },
    {
      "epoch": 0.6399326795481762,
      "grad_norm": 1.6898399958472543,
      "learning_rate": 4.3202557638782465e-07,
      "loss": 2.3667,
      "step": 4943
    },
    {
      "epoch": 0.6400621419555297,
      "grad_norm": 1.4460994997464984,
      "learning_rate": 4.3175692115718814e-07,
      "loss": 2.3154,
      "step": 4944
    },
    {
      "epoch": 0.6401916043628831,
      "grad_norm": 1.451576549912737,
      "learning_rate": 4.3148830540036754e-07,
      "loss": 2.3804,
      "step": 4945
    },
    {
      "epoch": 0.6401916043628831,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.522172451019287,
      "eval_runtime": 14.2568,
      "eval_samples_per_second": 3.086,
      "eval_steps_per_second": 0.21,
      "step": 4945
    },
    {
      "epoch": 0.6403210667702366,
      "grad_norm": 4.345837209415813,
      "learning_rate": 4.3121972917223265e-07,
      "loss": 2.5698,
      "step": 4946
    },
    {
      "epoch": 0.64045052917759,
      "grad_norm": 1.606721975497307,
      "learning_rate": 4.3095119252764585e-07,
      "loss": 2.272,
      "step": 4947
    },
    {
      "epoch": 0.6405799915849435,
      "grad_norm": 1.2256735042129443,
      "learning_rate": 4.3068269552146114e-07,
      "loss": 2.3613,
      "step": 4948
    },
    {
      "epoch": 0.640709453992297,
      "grad_norm": 1.4752863202939852,
      "learning_rate": 4.3041423820852455e-07,
      "loss": 2.4041,
      "step": 4949
    },
    {
      "epoch": 0.6408389163996504,
      "grad_norm": 2.9786664964810474,
      "learning_rate": 4.30145820643674e-07,
      "loss": 2.9348,
      "step": 4950
    },
    {
      "epoch": 0.6408389163996504,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5193759202957153,
      "eval_runtime": 15.5295,
      "eval_samples_per_second": 2.833,
      "eval_steps_per_second": 0.193,
      "step": 4950
    },
    {
      "epoch": 0.640968378807004,
      "grad_norm": 1.2978239288269562,
      "learning_rate": 4.29877442881739e-07,
      "loss": 2.3005,
      "step": 4951
    },
    {
      "epoch": 0.6410978412143574,
      "grad_norm": 2.184264788065103,
      "learning_rate": 4.296091049775412e-07,
      "loss": 2.4124,
      "step": 4952
    },
    {
      "epoch": 0.6412273036217109,
      "grad_norm": 1.3962816740289716,
      "learning_rate": 4.293408069858941e-07,
      "loss": 2.2791,
      "step": 4953
    },
    {
      "epoch": 0.6413567660290643,
      "grad_norm": 2.0052974936005965,
      "learning_rate": 4.290725489616029e-07,
      "loss": 2.4958,
      "step": 4954
    },
    {
      "epoch": 0.6414862284364178,
      "grad_norm": 2.2378135269057946,
      "learning_rate": 4.2880433095946494e-07,
      "loss": 2.6021,
      "step": 4955
    },
    {
      "epoch": 0.6414862284364178,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5187100172042847,
      "eval_runtime": 14.6229,
      "eval_samples_per_second": 3.009,
      "eval_steps_per_second": 0.205,
      "step": 4955
    },
    {
      "epoch": 0.6416156908437712,
      "grad_norm": 1.9479978529502049,
      "learning_rate": 4.285361530342694e-07,
      "loss": 2.3157,
      "step": 4956
    },
    {
      "epoch": 0.6417451532511247,
      "grad_norm": 1.5151373476144803,
      "learning_rate": 4.2826801524079644e-07,
      "loss": 2.5103,
      "step": 4957
    },
    {
      "epoch": 0.6418746156584781,
      "grad_norm": 2.5521479923206782,
      "learning_rate": 4.2799991763381875e-07,
      "loss": 2.4475,
      "step": 4958
    },
    {
      "epoch": 0.6420040780658316,
      "grad_norm": 1.5085737957913836,
      "learning_rate": 4.277318602681011e-07,
      "loss": 2.6428,
      "step": 4959
    },
    {
      "epoch": 0.642133540473185,
      "grad_norm": 2.0426052348627017,
      "learning_rate": 4.274638431983995e-07,
      "loss": 2.3787,
      "step": 4960
    },
    {
      "epoch": 0.642133540473185,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5193092823028564,
      "eval_runtime": 14.9381,
      "eval_samples_per_second": 2.945,
      "eval_steps_per_second": 0.201,
      "step": 4960
    },
    {
      "epoch": 0.6422630028805386,
      "grad_norm": 1.9005883489478663,
      "learning_rate": 4.271958664794619e-07,
      "loss": 2.6172,
      "step": 4961
    },
    {
      "epoch": 0.6423924652878921,
      "grad_norm": 1.3883497758863885,
      "learning_rate": 4.269279301660276e-07,
      "loss": 2.5564,
      "step": 4962
    },
    {
      "epoch": 0.6425219276952455,
      "grad_norm": 1.3788152288368425,
      "learning_rate": 4.266600343128284e-07,
      "loss": 2.376,
      "step": 4963
    },
    {
      "epoch": 0.642651390102599,
      "grad_norm": 1.8944665527086342,
      "learning_rate": 4.2639217897458743e-07,
      "loss": 2.5703,
      "step": 4964
    },
    {
      "epoch": 0.6427808525099524,
      "grad_norm": 1.3930917307540436,
      "learning_rate": 4.2612436420601933e-07,
      "loss": 2.2278,
      "step": 4965
    },
    {
      "epoch": 0.6427808525099524,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5193537473678589,
      "eval_runtime": 14.2742,
      "eval_samples_per_second": 3.082,
      "eval_steps_per_second": 0.21,
      "step": 4965
    },
    {
      "epoch": 0.6429103149173059,
      "grad_norm": 3.668165570018318,
      "learning_rate": 4.25856590061831e-07,
      "loss": 2.5049,
      "step": 4966
    },
    {
      "epoch": 0.6430397773246593,
      "grad_norm": 1.5791329369018345,
      "learning_rate": 4.2558885659672076e-07,
      "loss": 2.3518,
      "step": 4967
    },
    {
      "epoch": 0.6431692397320128,
      "grad_norm": 1.308936334945954,
      "learning_rate": 4.253211638653783e-07,
      "loss": 2.2686,
      "step": 4968
    },
    {
      "epoch": 0.6432987021393662,
      "grad_norm": 1.9552846725998698,
      "learning_rate": 4.2505351192248536e-07,
      "loss": 2.272,
      "step": 4969
    },
    {
      "epoch": 0.6434281645467197,
      "grad_norm": 1.4064037493957136,
      "learning_rate": 4.2478590082271536e-07,
      "loss": 2.3115,
      "step": 4970
    },
    {
      "epoch": 0.6434281645467197,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.519975185394287,
      "eval_runtime": 15.0198,
      "eval_samples_per_second": 2.929,
      "eval_steps_per_second": 0.2,
      "step": 4970
    },
    {
      "epoch": 0.6435576269540733,
      "grad_norm": 1.514505670954632,
      "learning_rate": 4.2451833062073324e-07,
      "loss": 2.4558,
      "step": 4971
    },
    {
      "epoch": 0.6436870893614267,
      "grad_norm": 1.4113416528167442,
      "learning_rate": 4.242508013711959e-07,
      "loss": 2.241,
      "step": 4972
    },
    {
      "epoch": 0.6438165517687802,
      "grad_norm": 1.6449012745926082,
      "learning_rate": 4.2398331312875114e-07,
      "loss": 2.4839,
      "step": 4973
    },
    {
      "epoch": 0.6439460141761336,
      "grad_norm": 4.059713504578763,
      "learning_rate": 4.2371586594803916e-07,
      "loss": 2.5552,
      "step": 4974
    },
    {
      "epoch": 0.6440754765834871,
      "grad_norm": 1.5577016706588052,
      "learning_rate": 4.2344845988369133e-07,
      "loss": 2.4517,
      "step": 4975
    },
    {
      "epoch": 0.6440754765834871,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.519486904144287,
      "eval_runtime": 14.8622,
      "eval_samples_per_second": 2.961,
      "eval_steps_per_second": 0.202,
      "step": 4975
    },
    {
      "epoch": 0.6442049389908405,
      "grad_norm": 2.178007375575733,
      "learning_rate": 4.2318109499033086e-07,
      "loss": 2.1646,
      "step": 4976
    },
    {
      "epoch": 0.644334401398194,
      "grad_norm": 3.1869711669259595,
      "learning_rate": 4.229137713225724e-07,
      "loss": 2.7422,
      "step": 4977
    },
    {
      "epoch": 0.6444638638055474,
      "grad_norm": 2.0844338991285354,
      "learning_rate": 4.2264648893502247e-07,
      "loss": 2.4128,
      "step": 4978
    },
    {
      "epoch": 0.6445933262129009,
      "grad_norm": 1.9875051853425896,
      "learning_rate": 4.223792478822786e-07,
      "loss": 2.1685,
      "step": 4979
    },
    {
      "epoch": 0.6447227886202543,
      "grad_norm": 2.4305407465946973,
      "learning_rate": 4.221120482189302e-07,
      "loss": 2.3511,
      "step": 4980
    },
    {
      "epoch": 0.6447227886202543,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5171786546707153,
      "eval_runtime": 14.3067,
      "eval_samples_per_second": 3.075,
      "eval_steps_per_second": 0.21,
      "step": 4980
    },
    {
      "epoch": 0.6448522510276079,
      "grad_norm": 1.5319464305702553,
      "learning_rate": 4.2184488999955814e-07,
      "loss": 2.4819,
      "step": 4981
    },
    {
      "epoch": 0.6449817134349614,
      "grad_norm": 2.1200934321157523,
      "learning_rate": 4.215777732787353e-07,
      "loss": 2.438,
      "step": 4982
    },
    {
      "epoch": 0.6451111758423148,
      "grad_norm": 2.5104283913079084,
      "learning_rate": 4.213106981110253e-07,
      "loss": 2.4199,
      "step": 4983
    },
    {
      "epoch": 0.6452406382496683,
      "grad_norm": 2.8242708093084508,
      "learning_rate": 4.210436645509842e-07,
      "loss": 2.6672,
      "step": 4984
    },
    {
      "epoch": 0.6453701006570217,
      "grad_norm": 2.0260523237758394,
      "learning_rate": 4.2077667265315815e-07,
      "loss": 2.2629,
      "step": 4985
    },
    {
      "epoch": 0.6453701006570217,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5165793895721436,
      "eval_runtime": 14.6831,
      "eval_samples_per_second": 2.997,
      "eval_steps_per_second": 0.204,
      "step": 4985
    },
    {
      "epoch": 0.6454995630643752,
      "grad_norm": 1.7376328174174853,
      "learning_rate": 4.205097224720863e-07,
      "loss": 2.2319,
      "step": 4986
    },
    {
      "epoch": 0.6456290254717286,
      "grad_norm": 3.325165477359538,
      "learning_rate": 4.202428140622984e-07,
      "loss": 2.3313,
      "step": 4987
    },
    {
      "epoch": 0.6457584878790821,
      "grad_norm": 2.104664033510982,
      "learning_rate": 4.1997594747831606e-07,
      "loss": 2.3867,
      "step": 4988
    },
    {
      "epoch": 0.6458879502864355,
      "grad_norm": 2.410161047294339,
      "learning_rate": 4.197091227746522e-07,
      "loss": 2.3816,
      "step": 4989
    },
    {
      "epoch": 0.6460174126937891,
      "grad_norm": 2.8756805439705,
      "learning_rate": 4.1944234000581093e-07,
      "loss": 2.7629,
      "step": 4990
    },
    {
      "epoch": 0.6460174126937891,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5208629369735718,
      "eval_runtime": 14.457,
      "eval_samples_per_second": 3.044,
      "eval_steps_per_second": 0.208,
      "step": 4990
    },
    {
      "epoch": 0.6461468751011425,
      "grad_norm": 2.346182580128464,
      "learning_rate": 4.191755992262882e-07,
      "loss": 2.5061,
      "step": 4991
    },
    {
      "epoch": 0.646276337508496,
      "grad_norm": 1.3889098894000738,
      "learning_rate": 4.189089004905711e-07,
      "loss": 2.3145,
      "step": 4992
    },
    {
      "epoch": 0.6464057999158495,
      "grad_norm": 1.4824765381743759,
      "learning_rate": 4.1864224385313845e-07,
      "loss": 2.219,
      "step": 4993
    },
    {
      "epoch": 0.6465352623232029,
      "grad_norm": 1.9514998596828175,
      "learning_rate": 4.1837562936846017e-07,
      "loss": 2.3149,
      "step": 4994
    },
    {
      "epoch": 0.6466647247305564,
      "grad_norm": 3.0065328230364323,
      "learning_rate": 4.1810905709099775e-07,
      "loss": 2.446,
      "step": 4995
    },
    {
      "epoch": 0.6466647247305564,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5193537473678589,
      "eval_runtime": 14.1913,
      "eval_samples_per_second": 3.101,
      "eval_steps_per_second": 0.211,
      "step": 4995
    },
    {
      "epoch": 0.6467941871379098,
      "grad_norm": 1.2117412755033656,
      "learning_rate": 4.1784252707520387e-07,
      "loss": 2.2966,
      "step": 4996
    },
    {
      "epoch": 0.6469236495452633,
      "grad_norm": 1.5046607177680527,
      "learning_rate": 4.175760393755227e-07,
      "loss": 2.3955,
      "step": 4997
    },
    {
      "epoch": 0.6470531119526167,
      "grad_norm": 2.2201490531030763,
      "learning_rate": 4.1730959404638964e-07,
      "loss": 2.5449,
      "step": 4998
    },
    {
      "epoch": 0.6471825743599702,
      "grad_norm": 2.0009148280349387,
      "learning_rate": 4.1704319114223167e-07,
      "loss": 2.3579,
      "step": 4999
    },
    {
      "epoch": 0.6473120367673237,
      "grad_norm": 2.619664473274341,
      "learning_rate": 4.1677683071746744e-07,
      "loss": 2.4207,
      "step": 5000
    },
    {
      "epoch": 0.6473120367673237,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5201083421707153,
      "eval_runtime": 14.6853,
      "eval_samples_per_second": 2.996,
      "eval_steps_per_second": 0.204,
      "step": 5000
    },
    {
      "epoch": 0.6474414991746772,
      "grad_norm": 1.5423378270164758,
      "learning_rate": 4.1651051282650565e-07,
      "loss": 2.23,
      "step": 5001
    },
    {
      "epoch": 0.6475709615820306,
      "grad_norm": 2.2168790713588766,
      "learning_rate": 4.1624423752374757e-07,
      "loss": 2.467,
      "step": 5002
    },
    {
      "epoch": 0.6477004239893841,
      "grad_norm": 4.238044830500742,
      "learning_rate": 4.1597800486358514e-07,
      "loss": 2.7502,
      "step": 5003
    },
    {
      "epoch": 0.6478298863967376,
      "grad_norm": 1.315030738518405,
      "learning_rate": 4.1571181490040203e-07,
      "loss": 2.2556,
      "step": 5004
    },
    {
      "epoch": 0.647959348804091,
      "grad_norm": 2.0476048994716485,
      "learning_rate": 4.1544566768857276e-07,
      "loss": 2.5449,
      "step": 5005
    },
    {
      "epoch": 0.647959348804091,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5178444385528564,
      "eval_runtime": 13.1657,
      "eval_samples_per_second": 3.342,
      "eval_steps_per_second": 0.228,
      "step": 5005
    },
    {
      "epoch": 0.6480888112114445,
      "grad_norm": 4.1503996450971306,
      "learning_rate": 4.151795632824637e-07,
      "loss": 2.7852,
      "step": 5006
    },
    {
      "epoch": 0.6482182736187979,
      "grad_norm": 2.3969681986438585,
      "learning_rate": 4.149135017364313e-07,
      "loss": 2.4185,
      "step": 5007
    },
    {
      "epoch": 0.6483477360261514,
      "grad_norm": 5.467870955809488,
      "learning_rate": 4.1464748310482446e-07,
      "loss": 2.9004,
      "step": 5008
    },
    {
      "epoch": 0.6484771984335048,
      "grad_norm": 3.302516453557103,
      "learning_rate": 4.1438150744198285e-07,
      "loss": 2.5051,
      "step": 5009
    },
    {
      "epoch": 0.6486066608408584,
      "grad_norm": 1.9473029119630008,
      "learning_rate": 4.141155748022375e-07,
      "loss": 2.2058,
      "step": 5010
    },
    {
      "epoch": 0.6486066608408584,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5214622020721436,
      "eval_runtime": 15.6205,
      "eval_samples_per_second": 2.817,
      "eval_steps_per_second": 0.192,
      "step": 5010
    },
    {
      "epoch": 0.6487361232482118,
      "grad_norm": 1.5225694385522397,
      "learning_rate": 4.138496852399104e-07,
      "loss": 2.3213,
      "step": 5011
    },
    {
      "epoch": 0.6488655856555653,
      "grad_norm": 1.29497540285883,
      "learning_rate": 4.1358383880931497e-07,
      "loss": 2.3738,
      "step": 5012
    },
    {
      "epoch": 0.6489950480629187,
      "grad_norm": 2.210618806339647,
      "learning_rate": 4.133180355647555e-07,
      "loss": 2.4683,
      "step": 5013
    },
    {
      "epoch": 0.6491245104702722,
      "grad_norm": 1.7513332411139912,
      "learning_rate": 4.13052275560528e-07,
      "loss": 2.3081,
      "step": 5014
    },
    {
      "epoch": 0.6492539728776257,
      "grad_norm": 1.8535856629197902,
      "learning_rate": 4.127865588509191e-07,
      "loss": 2.4546,
      "step": 5015
    },
    {
      "epoch": 0.6492539728776257,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5193759202957153,
      "eval_runtime": 14.0118,
      "eval_samples_per_second": 3.14,
      "eval_steps_per_second": 0.214,
      "step": 5015
    },
    {
      "epoch": 0.6493834352849791,
      "grad_norm": 2.7257107635063034,
      "learning_rate": 4.12520885490207e-07,
      "loss": 2.5786,
      "step": 5016
    },
    {
      "epoch": 0.6495128976923326,
      "grad_norm": 1.7039836552107441,
      "learning_rate": 4.1225525553266086e-07,
      "loss": 2.3533,
      "step": 5017
    },
    {
      "epoch": 0.649642360099686,
      "grad_norm": 1.9939476435919625,
      "learning_rate": 4.119896690325408e-07,
      "loss": 2.5183,
      "step": 5018
    },
    {
      "epoch": 0.6497718225070395,
      "grad_norm": 1.7527247350777142,
      "learning_rate": 4.1172412604409836e-07,
      "loss": 2.4221,
      "step": 5019
    },
    {
      "epoch": 0.649901284914393,
      "grad_norm": 1.333675338909329,
      "learning_rate": 4.114586266215762e-07,
      "loss": 2.269,
      "step": 5020
    },
    {
      "epoch": 0.649901284914393,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5201305150985718,
      "eval_runtime": 13.7462,
      "eval_samples_per_second": 3.201,
      "eval_steps_per_second": 0.218,
      "step": 5020
    },
    {
      "epoch": 0.6500307473217465,
      "grad_norm": 1.734064725440254,
      "learning_rate": 4.111931708192077e-07,
      "loss": 2.5037,
      "step": 5021
    },
    {
      "epoch": 0.6501602097290999,
      "grad_norm": 1.4368427898116118,
      "learning_rate": 4.1092775869121764e-07,
      "loss": 2.4485,
      "step": 5022
    },
    {
      "epoch": 0.6502896721364534,
      "grad_norm": 1.9345469087588762,
      "learning_rate": 4.106623902918223e-07,
      "loss": 2.2983,
      "step": 5023
    },
    {
      "epoch": 0.6504191345438068,
      "grad_norm": 1.3994814096798716,
      "learning_rate": 4.103970656752279e-07,
      "loss": 2.3792,
      "step": 5024
    },
    {
      "epoch": 0.6505485969511603,
      "grad_norm": 1.7151358282869378,
      "learning_rate": 4.101317848956327e-07,
      "loss": 2.2712,
      "step": 5025
    },
    {
      "epoch": 0.6505485969511603,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5166015625,
      "eval_runtime": 13.7232,
      "eval_samples_per_second": 3.206,
      "eval_steps_per_second": 0.219,
      "step": 5025
    },
    {
      "epoch": 0.6506780593585137,
      "grad_norm": 1.343725655668395,
      "learning_rate": 4.098665480072254e-07,
      "loss": 2.2896,
      "step": 5026
    },
    {
      "epoch": 0.6508075217658672,
      "grad_norm": 1.2760485238642996,
      "learning_rate": 4.096013550641864e-07,
      "loss": 2.2854,
      "step": 5027
    },
    {
      "epoch": 0.6509369841732207,
      "grad_norm": 2.0562967627072375,
      "learning_rate": 4.0933620612068683e-07,
      "loss": 2.4966,
      "step": 5028
    },
    {
      "epoch": 0.6510664465805742,
      "grad_norm": 2.464894842372267,
      "learning_rate": 4.090711012308882e-07,
      "loss": 2.509,
      "step": 5029
    },
    {
      "epoch": 0.6511959089879277,
      "grad_norm": 1.3179228354792176,
      "learning_rate": 4.088060404489436e-07,
      "loss": 2.3752,
      "step": 5030
    },
    {
      "epoch": 0.6511959089879277,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5228382349014282,
      "eval_runtime": 13.6436,
      "eval_samples_per_second": 3.225,
      "eval_steps_per_second": 0.22,
      "step": 5030
    },
    {
      "epoch": 0.6513253713952811,
      "grad_norm": 3.858383770113083,
      "learning_rate": 4.0854102382899753e-07,
      "loss": 2.7832,
      "step": 5031
    },
    {
      "epoch": 0.6514548338026346,
      "grad_norm": 2.513984411162413,
      "learning_rate": 4.082760514251848e-07,
      "loss": 2.4937,
      "step": 5032
    },
    {
      "epoch": 0.651584296209988,
      "grad_norm": 2.5458047701558995,
      "learning_rate": 4.0801112329163123e-07,
      "loss": 2.2312,
      "step": 5033
    },
    {
      "epoch": 0.6517137586173415,
      "grad_norm": 1.6480911628021269,
      "learning_rate": 4.0774623948245414e-07,
      "loss": 2.3281,
      "step": 5034
    },
    {
      "epoch": 0.6518432210246949,
      "grad_norm": 1.4769260925158803,
      "learning_rate": 4.0748140005176096e-07,
      "loss": 2.4873,
      "step": 5035
    },
    {
      "epoch": 0.6518432210246949,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5186434984207153,
      "eval_runtime": 14.5776,
      "eval_samples_per_second": 3.018,
      "eval_steps_per_second": 0.206,
      "step": 5035
    },
    {
      "epoch": 0.6519726834320484,
      "grad_norm": 1.378453756377644,
      "learning_rate": 4.0721660505365073e-07,
      "loss": 2.5588,
      "step": 5036
    },
    {
      "epoch": 0.6521021458394018,
      "grad_norm": 2.730489453808209,
      "learning_rate": 4.069518545422132e-07,
      "loss": 2.4067,
      "step": 5037
    },
    {
      "epoch": 0.6522316082467553,
      "grad_norm": 3.346448500996533,
      "learning_rate": 4.06687148571529e-07,
      "loss": 2.5771,
      "step": 5038
    },
    {
      "epoch": 0.6523610706541089,
      "grad_norm": 1.189941279327231,
      "learning_rate": 4.0642248719566983e-07,
      "loss": 2.3987,
      "step": 5039
    },
    {
      "epoch": 0.6524905330614623,
      "grad_norm": 1.934587604175701,
      "learning_rate": 4.06157870468698e-07,
      "loss": 2.5571,
      "step": 5040
    },
    {
      "epoch": 0.6524905330614623,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5215731859207153,
      "eval_runtime": 14.5379,
      "eval_samples_per_second": 3.027,
      "eval_steps_per_second": 0.206,
      "step": 5040
    },
    {
      "epoch": 0.6526199954688158,
      "grad_norm": 2.6843035624478992,
      "learning_rate": 4.058932984446667e-07,
      "loss": 2.3965,
      "step": 5041
    },
    {
      "epoch": 0.6527494578761692,
      "grad_norm": 2.2576884183651367,
      "learning_rate": 4.056287711776204e-07,
      "loss": 2.5122,
      "step": 5042
    },
    {
      "epoch": 0.6528789202835227,
      "grad_norm": 1.3509512030833801,
      "learning_rate": 4.053642887215939e-07,
      "loss": 2.3586,
      "step": 5043
    },
    {
      "epoch": 0.6530083826908761,
      "grad_norm": 1.5034457112503246,
      "learning_rate": 4.050998511306134e-07,
      "loss": 2.5308,
      "step": 5044
    },
    {
      "epoch": 0.6531378450982296,
      "grad_norm": 2.443284834086762,
      "learning_rate": 4.0483545845869556e-07,
      "loss": 2.3098,
      "step": 5045
    },
    {
      "epoch": 0.6531378450982296,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5187100172042847,
      "eval_runtime": 13.6516,
      "eval_samples_per_second": 3.223,
      "eval_steps_per_second": 0.22,
      "step": 5045
    },
    {
      "epoch": 0.653267307505583,
      "grad_norm": 1.4599420498327067,
      "learning_rate": 4.0457111075984767e-07,
      "loss": 2.6724,
      "step": 5046
    },
    {
      "epoch": 0.6533967699129365,
      "grad_norm": 5.067183238761789,
      "learning_rate": 4.0430680808806825e-07,
      "loss": 2.7002,
      "step": 5047
    },
    {
      "epoch": 0.65352623232029,
      "grad_norm": 1.670059141004386,
      "learning_rate": 4.0404255049734667e-07,
      "loss": 2.3376,
      "step": 5048
    },
    {
      "epoch": 0.6536556947276435,
      "grad_norm": 2.5271320683805483,
      "learning_rate": 4.037783380416625e-07,
      "loss": 2.2805,
      "step": 5049
    },
    {
      "epoch": 0.653785157134997,
      "grad_norm": 2.481168647184336,
      "learning_rate": 4.0351417077498727e-07,
      "loss": 2.4534,
      "step": 5050
    },
    {
      "epoch": 0.653785157134997,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5179775953292847,
      "eval_runtime": 13.5056,
      "eval_samples_per_second": 3.258,
      "eval_steps_per_second": 0.222,
      "step": 5050
    },
    {
      "epoch": 0.6539146195423504,
      "grad_norm": 3.3264588868860114,
      "learning_rate": 4.0325004875128167e-07,
      "loss": 2.7644,
      "step": 5051
    },
    {
      "epoch": 0.6540440819497039,
      "grad_norm": 2.2105953991858085,
      "learning_rate": 4.0298597202449824e-07,
      "loss": 2.4448,
      "step": 5052
    },
    {
      "epoch": 0.6541735443570573,
      "grad_norm": 2.9751967752117374,
      "learning_rate": 4.0272194064857995e-07,
      "loss": 2.5493,
      "step": 5053
    },
    {
      "epoch": 0.6543030067644108,
      "grad_norm": 1.6447415487331738,
      "learning_rate": 4.0245795467746067e-07,
      "loss": 2.1528,
      "step": 5054
    },
    {
      "epoch": 0.6544324691717642,
      "grad_norm": 1.6295542111171875,
      "learning_rate": 4.02194014165065e-07,
      "loss": 2.4771,
      "step": 5055
    },
    {
      "epoch": 0.6544324691717642,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5200417041778564,
      "eval_runtime": 13.9813,
      "eval_samples_per_second": 3.147,
      "eval_steps_per_second": 0.215,
      "step": 5055
    },
    {
      "epoch": 0.6545619315791177,
      "grad_norm": 1.7161823327021302,
      "learning_rate": 4.0193011916530823e-07,
      "loss": 2.2747,
      "step": 5056
    },
    {
      "epoch": 0.6546913939864711,
      "grad_norm": 1.8586556236181142,
      "learning_rate": 4.016662697320956e-07,
      "loss": 2.3313,
      "step": 5057
    },
    {
      "epoch": 0.6548208563938246,
      "grad_norm": 2.810041862092068,
      "learning_rate": 4.0140246591932433e-07,
      "loss": 2.4409,
      "step": 5058
    },
    {
      "epoch": 0.6549503188011782,
      "grad_norm": 2.674635196176011,
      "learning_rate": 4.011387077808815e-07,
      "loss": 2.2977,
      "step": 5059
    },
    {
      "epoch": 0.6550797812085316,
      "grad_norm": 2.0085069906240887,
      "learning_rate": 4.008749953706451e-07,
      "loss": 2.606,
      "step": 5060
    },
    {
      "epoch": 0.6550797812085316,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5186434984207153,
      "eval_runtime": 14.77,
      "eval_samples_per_second": 2.979,
      "eval_steps_per_second": 0.203,
      "step": 5060
    },
    {
      "epoch": 0.6552092436158851,
      "grad_norm": 1.5851830352913399,
      "learning_rate": 4.006113287424837e-07,
      "loss": 2.4102,
      "step": 5061
    },
    {
      "epoch": 0.6553387060232385,
      "grad_norm": 1.4390640008720585,
      "learning_rate": 4.003477079502566e-07,
      "loss": 2.3667,
      "step": 5062
    },
    {
      "epoch": 0.655468168430592,
      "grad_norm": 1.7512951961351804,
      "learning_rate": 4.000841330478136e-07,
      "loss": 2.3301,
      "step": 5063
    },
    {
      "epoch": 0.6555976308379454,
      "grad_norm": 1.3290034689401424,
      "learning_rate": 3.998206040889952e-07,
      "loss": 2.1821,
      "step": 5064
    },
    {
      "epoch": 0.6557270932452989,
      "grad_norm": 2.2412698819182113,
      "learning_rate": 3.9955712112763267e-07,
      "loss": 2.47,
      "step": 5065
    },
    {
      "epoch": 0.6557270932452989,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5164684057235718,
      "eval_runtime": 14.411,
      "eval_samples_per_second": 3.053,
      "eval_steps_per_second": 0.208,
      "step": 5065
    },
    {
      "epoch": 0.6558565556526523,
      "grad_norm": 3.6718753992178654,
      "learning_rate": 3.9929368421754765e-07,
      "loss": 2.5491,
      "step": 5066
    },
    {
      "epoch": 0.6559860180600058,
      "grad_norm": 1.228562881685534,
      "learning_rate": 3.990302934125526e-07,
      "loss": 2.3064,
      "step": 5067
    },
    {
      "epoch": 0.6561154804673592,
      "grad_norm": 1.5334290980141807,
      "learning_rate": 3.9876694876645016e-07,
      "loss": 2.3853,
      "step": 5068
    },
    {
      "epoch": 0.6562449428747128,
      "grad_norm": 2.1654949654858404,
      "learning_rate": 3.9850365033303396e-07,
      "loss": 2.4165,
      "step": 5069
    },
    {
      "epoch": 0.6563744052820663,
      "grad_norm": 2.4019823823346673,
      "learning_rate": 3.982403981660881e-07,
      "loss": 2.3071,
      "step": 5070
    },
    {
      "epoch": 0.6563744052820663,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5206631422042847,
      "eval_runtime": 15.7309,
      "eval_samples_per_second": 2.797,
      "eval_steps_per_second": 0.191,
      "step": 5070
    },
    {
      "epoch": 0.6565038676894197,
      "grad_norm": 3.4869967089434932,
      "learning_rate": 3.979771923193871e-07,
      "loss": 2.8848,
      "step": 5071
    },
    {
      "epoch": 0.6566333300967732,
      "grad_norm": 2.1100427911813036,
      "learning_rate": 3.97714032846696e-07,
      "loss": 2.8606,
      "step": 5072
    },
    {
      "epoch": 0.6567627925041266,
      "grad_norm": 1.3733993212918951,
      "learning_rate": 3.974509198017709e-07,
      "loss": 2.3398,
      "step": 5073
    },
    {
      "epoch": 0.6568922549114801,
      "grad_norm": 2.6078588587997893,
      "learning_rate": 3.9718785323835747e-07,
      "loss": 2.3777,
      "step": 5074
    },
    {
      "epoch": 0.6570217173188335,
      "grad_norm": 1.8913946025395285,
      "learning_rate": 3.969248332101925e-07,
      "loss": 2.3518,
      "step": 5075
    },
    {
      "epoch": 0.6570217173188335,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.52001953125,
      "eval_runtime": 13.3703,
      "eval_samples_per_second": 3.291,
      "eval_steps_per_second": 0.224,
      "step": 5075
    },
    {
      "epoch": 0.657151179726187,
      "grad_norm": 2.120538980944094,
      "learning_rate": 3.9666185977100314e-07,
      "loss": 2.5015,
      "step": 5076
    },
    {
      "epoch": 0.6572806421335404,
      "grad_norm": 1.7462444312062881,
      "learning_rate": 3.9639893297450714e-07,
      "loss": 2.3416,
      "step": 5077
    },
    {
      "epoch": 0.657410104540894,
      "grad_norm": 1.3142990010739095,
      "learning_rate": 3.96136052874413e-07,
      "loss": 2.4033,
      "step": 5078
    },
    {
      "epoch": 0.6575395669482474,
      "grad_norm": 2.775288309021671,
      "learning_rate": 3.9587321952441857e-07,
      "loss": 2.3828,
      "step": 5079
    },
    {
      "epoch": 0.6576690293556009,
      "grad_norm": 1.5540206084955568,
      "learning_rate": 3.9561043297821316e-07,
      "loss": 2.5361,
      "step": 5080
    },
    {
      "epoch": 0.6576690293556009,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5213512182235718,
      "eval_runtime": 14.2628,
      "eval_samples_per_second": 3.085,
      "eval_steps_per_second": 0.21,
      "step": 5080
    },
    {
      "epoch": 0.6577984917629544,
      "grad_norm": 1.2668582158650334,
      "learning_rate": 3.9534769328947647e-07,
      "loss": 2.3352,
      "step": 5081
    },
    {
      "epoch": 0.6579279541703078,
      "grad_norm": 1.3170165531362004,
      "learning_rate": 3.9508500051187826e-07,
      "loss": 2.3354,
      "step": 5082
    },
    {
      "epoch": 0.6580574165776613,
      "grad_norm": 3.303702335316497,
      "learning_rate": 3.9482235469907884e-07,
      "loss": 2.1265,
      "step": 5083
    },
    {
      "epoch": 0.6581868789850147,
      "grad_norm": 1.808692046422752,
      "learning_rate": 3.9455975590472914e-07,
      "loss": 2.4451,
      "step": 5084
    },
    {
      "epoch": 0.6583163413923682,
      "grad_norm": 1.3869193689002375,
      "learning_rate": 3.942972041824699e-07,
      "loss": 2.2463,
      "step": 5085
    },
    {
      "epoch": 0.6583163413923682,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5206853151321411,
      "eval_runtime": 15.0582,
      "eval_samples_per_second": 2.922,
      "eval_steps_per_second": 0.199,
      "step": 5085
    },
    {
      "epoch": 0.6584458037997216,
      "grad_norm": 2.0581482240865334,
      "learning_rate": 3.9403469958593286e-07,
      "loss": 2.51,
      "step": 5086
    },
    {
      "epoch": 0.6585752662070751,
      "grad_norm": 1.6978150213233507,
      "learning_rate": 3.937722421687399e-07,
      "loss": 2.4026,
      "step": 5087
    },
    {
      "epoch": 0.6587047286144286,
      "grad_norm": 1.5749043485762106,
      "learning_rate": 3.935098319845032e-07,
      "loss": 2.5483,
      "step": 5088
    },
    {
      "epoch": 0.6588341910217821,
      "grad_norm": 1.7346630258114264,
      "learning_rate": 3.932474690868256e-07,
      "loss": 2.6348,
      "step": 5089
    },
    {
      "epoch": 0.6589636534291355,
      "grad_norm": 1.8385748672661781,
      "learning_rate": 3.9298515352929974e-07,
      "loss": 2.396,
      "step": 5090
    },
    {
      "epoch": 0.6589636534291355,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.520063877105713,
      "eval_runtime": 14.2696,
      "eval_samples_per_second": 3.083,
      "eval_steps_per_second": 0.21,
      "step": 5090
    },
    {
      "epoch": 0.659093115836489,
      "grad_norm": 1.7919133923344202,
      "learning_rate": 3.927228853655089e-07,
      "loss": 2.2737,
      "step": 5091
    },
    {
      "epoch": 0.6592225782438424,
      "grad_norm": 2.472952432846999,
      "learning_rate": 3.9246066464902685e-07,
      "loss": 2.3416,
      "step": 5092
    },
    {
      "epoch": 0.6593520406511959,
      "grad_norm": 1.8409635578852772,
      "learning_rate": 3.9219849143341745e-07,
      "loss": 2.3484,
      "step": 5093
    },
    {
      "epoch": 0.6594815030585494,
      "grad_norm": 1.5354990759147005,
      "learning_rate": 3.9193636577223464e-07,
      "loss": 2.2981,
      "step": 5094
    },
    {
      "epoch": 0.6596109654659028,
      "grad_norm": 2.8406258733537837,
      "learning_rate": 3.916742877190235e-07,
      "loss": 2.4692,
      "step": 5095
    },
    {
      "epoch": 0.6596109654659028,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.521484375,
      "eval_runtime": 14.1393,
      "eval_samples_per_second": 3.112,
      "eval_steps_per_second": 0.212,
      "step": 5095
    },
    {
      "epoch": 0.6597404278732563,
      "grad_norm": 1.3670449957574278,
      "learning_rate": 3.914122573273181e-07,
      "loss": 2.2891,
      "step": 5096
    },
    {
      "epoch": 0.6598698902806097,
      "grad_norm": 2.506823950901503,
      "learning_rate": 3.9115027465064375e-07,
      "loss": 2.5471,
      "step": 5097
    },
    {
      "epoch": 0.6599993526879633,
      "grad_norm": 1.757220271206005,
      "learning_rate": 3.908883397425157e-07,
      "loss": 2.4419,
      "step": 5098
    },
    {
      "epoch": 0.6601288150953167,
      "grad_norm": 1.1859617161818887,
      "learning_rate": 3.9062645265643946e-07,
      "loss": 2.2988,
      "step": 5099
    },
    {
      "epoch": 0.6602582775026702,
      "grad_norm": 2.067553285362183,
      "learning_rate": 3.903646134459112e-07,
      "loss": 2.4841,
      "step": 5100
    },
    {
      "epoch": 0.6602582775026702,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.521440029144287,
      "eval_runtime": 14.6948,
      "eval_samples_per_second": 2.994,
      "eval_steps_per_second": 0.204,
      "step": 5100
    },
    {
      "epoch": 0.6603877399100236,
      "grad_norm": 1.4590953613474575,
      "learning_rate": 3.901028221644162e-07,
      "loss": 2.3407,
      "step": 5101
    },
    {
      "epoch": 0.6605172023173771,
      "grad_norm": 2.3624622050943693,
      "learning_rate": 3.898410788654307e-07,
      "loss": 2.4902,
      "step": 5102
    },
    {
      "epoch": 0.6606466647247305,
      "grad_norm": 1.796253910186038,
      "learning_rate": 3.895793836024215e-07,
      "loss": 2.5266,
      "step": 5103
    },
    {
      "epoch": 0.660776127132084,
      "grad_norm": 4.028303478472632,
      "learning_rate": 3.8931773642884493e-07,
      "loss": 2.3796,
      "step": 5104
    },
    {
      "epoch": 0.6609055895394375,
      "grad_norm": 1.6015025071514128,
      "learning_rate": 3.890561373981477e-07,
      "loss": 2.5498,
      "step": 5105
    },
    {
      "epoch": 0.6609055895394375,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5191982984542847,
      "eval_runtime": 14.5094,
      "eval_samples_per_second": 3.033,
      "eval_steps_per_second": 0.207,
      "step": 5105
    },
    {
      "epoch": 0.6610350519467909,
      "grad_norm": 1.6189397534895302,
      "learning_rate": 3.8879458656376707e-07,
      "loss": 2.5557,
      "step": 5106
    },
    {
      "epoch": 0.6611645143541444,
      "grad_norm": 1.7465514629863457,
      "learning_rate": 3.885330839791296e-07,
      "loss": 2.2998,
      "step": 5107
    },
    {
      "epoch": 0.6612939767614979,
      "grad_norm": 2.872800723127391,
      "learning_rate": 3.8827162969765265e-07,
      "loss": 2.7788,
      "step": 5108
    },
    {
      "epoch": 0.6614234391688514,
      "grad_norm": 1.1646514624490423,
      "learning_rate": 3.8801022377274356e-07,
      "loss": 2.3994,
      "step": 5109
    },
    {
      "epoch": 0.6615529015762048,
      "grad_norm": 3.8993895184748006,
      "learning_rate": 3.8774886625779985e-07,
      "loss": 2.7563,
      "step": 5110
    },
    {
      "epoch": 0.6615529015762048,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.520552158355713,
      "eval_runtime": 14.2053,
      "eval_samples_per_second": 3.097,
      "eval_steps_per_second": 0.211,
      "step": 5110
    },
    {
      "epoch": 0.6616823639835583,
      "grad_norm": 1.6761707718581333,
      "learning_rate": 3.874875572062091e-07,
      "loss": 2.5566,
      "step": 5111
    },
    {
      "epoch": 0.6618118263909117,
      "grad_norm": 2.42446852769922,
      "learning_rate": 3.872262966713491e-07,
      "loss": 2.5781,
      "step": 5112
    },
    {
      "epoch": 0.6619412887982652,
      "grad_norm": 1.4991038214552679,
      "learning_rate": 3.869650847065873e-07,
      "loss": 2.4695,
      "step": 5113
    },
    {
      "epoch": 0.6620707512056186,
      "grad_norm": 1.3298492183229564,
      "learning_rate": 3.867039213652817e-07,
      "loss": 2.4062,
      "step": 5114
    },
    {
      "epoch": 0.6622002136129721,
      "grad_norm": 2.0585475192212566,
      "learning_rate": 3.864428067007803e-07,
      "loss": 2.396,
      "step": 5115
    },
    {
      "epoch": 0.6622002136129721,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5179775953292847,
      "eval_runtime": 14.5302,
      "eval_samples_per_second": 3.028,
      "eval_steps_per_second": 0.206,
      "step": 5115
    },
    {
      "epoch": 0.6623296760203256,
      "grad_norm": 1.962227896635834,
      "learning_rate": 3.8618174076642076e-07,
      "loss": 2.3892,
      "step": 5116
    },
    {
      "epoch": 0.6624591384276791,
      "grad_norm": 2.287359298205945,
      "learning_rate": 3.8592072361553157e-07,
      "loss": 2.4265,
      "step": 5117
    },
    {
      "epoch": 0.6625886008350326,
      "grad_norm": 1.4751271525261889,
      "learning_rate": 3.856597553014303e-07,
      "loss": 2.3013,
      "step": 5118
    },
    {
      "epoch": 0.662718063242386,
      "grad_norm": 1.54005879839691,
      "learning_rate": 3.853988358774253e-07,
      "loss": 2.3831,
      "step": 5119
    },
    {
      "epoch": 0.6628475256497395,
      "grad_norm": 3.7889978247804117,
      "learning_rate": 3.8513796539681444e-07,
      "loss": 2.4536,
      "step": 5120
    },
    {
      "epoch": 0.6628475256497395,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5213955640792847,
      "eval_runtime": 14.1483,
      "eval_samples_per_second": 3.11,
      "eval_steps_per_second": 0.212,
      "step": 5120
    },
    {
      "epoch": 0.6629769880570929,
      "grad_norm": 1.4772823111355498,
      "learning_rate": 3.8487714391288595e-07,
      "loss": 2.3752,
      "step": 5121
    },
    {
      "epoch": 0.6631064504644464,
      "grad_norm": 1.8540814294947863,
      "learning_rate": 3.846163714789179e-07,
      "loss": 2.395,
      "step": 5122
    },
    {
      "epoch": 0.6632359128717998,
      "grad_norm": 1.5728833590556046,
      "learning_rate": 3.843556481481786e-07,
      "loss": 2.3787,
      "step": 5123
    },
    {
      "epoch": 0.6633653752791533,
      "grad_norm": 1.8407201326668305,
      "learning_rate": 3.840949739739256e-07,
      "loss": 2.5516,
      "step": 5124
    },
    {
      "epoch": 0.6634948376865067,
      "grad_norm": 1.1281453230800251,
      "learning_rate": 3.8383434900940695e-07,
      "loss": 2.1985,
      "step": 5125
    },
    {
      "epoch": 0.6634948376865067,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5184437036514282,
      "eval_runtime": 15.7301,
      "eval_samples_per_second": 2.797,
      "eval_steps_per_second": 0.191,
      "step": 5125
    },
    {
      "epoch": 0.6636243000938602,
      "grad_norm": 1.6718655606211663,
      "learning_rate": 3.835737733078608e-07,
      "loss": 2.6182,
      "step": 5126
    },
    {
      "epoch": 0.6637537625012138,
      "grad_norm": 1.2695176838103426,
      "learning_rate": 3.8331324692251494e-07,
      "loss": 2.177,
      "step": 5127
    },
    {
      "epoch": 0.6638832249085672,
      "grad_norm": 4.087295517944597,
      "learning_rate": 3.8305276990658753e-07,
      "loss": 2.4541,
      "step": 5128
    },
    {
      "epoch": 0.6640126873159207,
      "grad_norm": 1.8454832240698456,
      "learning_rate": 3.827923423132854e-07,
      "loss": 2.5847,
      "step": 5129
    },
    {
      "epoch": 0.6641421497232741,
      "grad_norm": 1.839616472821154,
      "learning_rate": 3.825319641958068e-07,
      "loss": 2.2156,
      "step": 5130
    },
    {
      "epoch": 0.6641421497232741,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5199307203292847,
      "eval_runtime": 13.9844,
      "eval_samples_per_second": 3.146,
      "eval_steps_per_second": 0.215,
      "step": 5130
    },
    {
      "epoch": 0.6642716121306276,
      "grad_norm": 3.2461673156177198,
      "learning_rate": 3.8227163560733924e-07,
      "loss": 2.7466,
      "step": 5131
    },
    {
      "epoch": 0.664401074537981,
      "grad_norm": 1.7380332264186027,
      "learning_rate": 3.820113566010599e-07,
      "loss": 2.3745,
      "step": 5132
    },
    {
      "epoch": 0.6645305369453345,
      "grad_norm": 2.344446791536525,
      "learning_rate": 3.8175112723013615e-07,
      "loss": 2.4751,
      "step": 5133
    },
    {
      "epoch": 0.6646599993526879,
      "grad_norm": 1.2471046012903564,
      "learning_rate": 3.814909475477252e-07,
      "loss": 2.1833,
      "step": 5134
    },
    {
      "epoch": 0.6647894617600414,
      "grad_norm": 2.8199112596985865,
      "learning_rate": 3.812308176069739e-07,
      "loss": 2.7944,
      "step": 5135
    },
    {
      "epoch": 0.6647894617600414,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.519242763519287,
      "eval_runtime": 14.1271,
      "eval_samples_per_second": 3.115,
      "eval_steps_per_second": 0.212,
      "step": 5135
    },
    {
      "epoch": 0.6649189241673948,
      "grad_norm": 1.5752387910162489,
      "learning_rate": 3.809707374610189e-07,
      "loss": 2.4453,
      "step": 5136
    },
    {
      "epoch": 0.6650483865747484,
      "grad_norm": 4.743906889161572,
      "learning_rate": 3.807107071629871e-07,
      "loss": 2.7966,
      "step": 5137
    },
    {
      "epoch": 0.6651778489821019,
      "grad_norm": 1.5386285627355307,
      "learning_rate": 3.8045072676599485e-07,
      "loss": 2.5845,
      "step": 5138
    },
    {
      "epoch": 0.6653073113894553,
      "grad_norm": 1.854354073363231,
      "learning_rate": 3.801907963231487e-07,
      "loss": 2.4751,
      "step": 5139
    },
    {
      "epoch": 0.6654367737968088,
      "grad_norm": 2.430741100944076,
      "learning_rate": 3.799309158875442e-07,
      "loss": 2.4009,
      "step": 5140
    },
    {
      "epoch": 0.6654367737968088,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.517777919769287,
      "eval_runtime": 13.9566,
      "eval_samples_per_second": 3.153,
      "eval_steps_per_second": 0.215,
      "step": 5140
    },
    {
      "epoch": 0.6655662362041622,
      "grad_norm": 1.680177128866981,
      "learning_rate": 3.7967108551226745e-07,
      "loss": 2.4585,
      "step": 5141
    },
    {
      "epoch": 0.6656956986115157,
      "grad_norm": 2.0435515407621634,
      "learning_rate": 3.794113052503942e-07,
      "loss": 2.4807,
      "step": 5142
    },
    {
      "epoch": 0.6658251610188691,
      "grad_norm": 2.2911964132494647,
      "learning_rate": 3.7915157515498956e-07,
      "loss": 2.6887,
      "step": 5143
    },
    {
      "epoch": 0.6659546234262226,
      "grad_norm": 1.5835927017498963,
      "learning_rate": 3.7889189527910876e-07,
      "loss": 2.4379,
      "step": 5144
    },
    {
      "epoch": 0.666084085833576,
      "grad_norm": 1.3622481158306292,
      "learning_rate": 3.7863226567579714e-07,
      "loss": 2.3567,
      "step": 5145
    },
    {
      "epoch": 0.666084085833576,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.519975185394287,
      "eval_runtime": 13.927,
      "eval_samples_per_second": 3.159,
      "eval_steps_per_second": 0.215,
      "step": 5145
    },
    {
      "epoch": 0.6662135482409295,
      "grad_norm": 1.8280483503986014,
      "learning_rate": 3.783726863980886e-07,
      "loss": 2.3433,
      "step": 5146
    },
    {
      "epoch": 0.666343010648283,
      "grad_norm": 1.8989326011540517,
      "learning_rate": 3.781131574990078e-07,
      "loss": 2.4128,
      "step": 5147
    },
    {
      "epoch": 0.6664724730556365,
      "grad_norm": 1.1719131934732134,
      "learning_rate": 3.7785367903156856e-07,
      "loss": 2.4124,
      "step": 5148
    },
    {
      "epoch": 0.66660193546299,
      "grad_norm": 2.6265563704726467,
      "learning_rate": 3.7759425104877507e-07,
      "loss": 2.3127,
      "step": 5149
    },
    {
      "epoch": 0.6667313978703434,
      "grad_norm": 1.9314737227858774,
      "learning_rate": 3.773348736036204e-07,
      "loss": 2.6221,
      "step": 5150
    },
    {
      "epoch": 0.6667313978703434,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5184437036514282,
      "eval_runtime": 14.5884,
      "eval_samples_per_second": 3.016,
      "eval_steps_per_second": 0.206,
      "step": 5150
    },
    {
      "epoch": 0.6668608602776969,
      "grad_norm": 1.4916318311367764,
      "learning_rate": 3.77075546749088e-07,
      "loss": 2.4309,
      "step": 5151
    },
    {
      "epoch": 0.6669903226850503,
      "grad_norm": 1.1933919085341167,
      "learning_rate": 3.768162705381499e-07,
      "loss": 2.3296,
      "step": 5152
    },
    {
      "epoch": 0.6671197850924038,
      "grad_norm": 2.0277253775961017,
      "learning_rate": 3.7655704502376934e-07,
      "loss": 2.3223,
      "step": 5153
    },
    {
      "epoch": 0.6672492474997572,
      "grad_norm": 1.9699504523553544,
      "learning_rate": 3.762978702588979e-07,
      "loss": 2.4688,
      "step": 5154
    },
    {
      "epoch": 0.6673787099071107,
      "grad_norm": 1.2773131139319953,
      "learning_rate": 3.7603874629647744e-07,
      "loss": 2.2078,
      "step": 5155
    },
    {
      "epoch": 0.6673787099071107,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5199528932571411,
      "eval_runtime": 14.8951,
      "eval_samples_per_second": 2.954,
      "eval_steps_per_second": 0.201,
      "step": 5155
    },
    {
      "epoch": 0.6675081723144641,
      "grad_norm": 1.5437265689498072,
      "learning_rate": 3.757796731894393e-07,
      "loss": 2.3879,
      "step": 5156
    },
    {
      "epoch": 0.6676376347218177,
      "grad_norm": 2.414722426905868,
      "learning_rate": 3.7552065099070444e-07,
      "loss": 2.5288,
      "step": 5157
    },
    {
      "epoch": 0.6677670971291711,
      "grad_norm": 1.8267964633701326,
      "learning_rate": 3.752616797531832e-07,
      "loss": 2.4658,
      "step": 5158
    },
    {
      "epoch": 0.6678965595365246,
      "grad_norm": 1.5019852449538464,
      "learning_rate": 3.7500275952977587e-07,
      "loss": 2.2144,
      "step": 5159
    },
    {
      "epoch": 0.6680260219438781,
      "grad_norm": 1.954640115840008,
      "learning_rate": 3.747438903733722e-07,
      "loss": 2.4492,
      "step": 5160
    },
    {
      "epoch": 0.6680260219438781,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5178444385528564,
      "eval_runtime": 14.0502,
      "eval_samples_per_second": 3.132,
      "eval_steps_per_second": 0.214,
      "step": 5160
    },
    {
      "epoch": 0.6681554843512315,
      "grad_norm": 1.8310980985668466,
      "learning_rate": 3.744850723368512e-07,
      "loss": 2.5166,
      "step": 5161
    },
    {
      "epoch": 0.668284946758585,
      "grad_norm": 1.6287100846845086,
      "learning_rate": 3.742263054730821e-07,
      "loss": 2.3857,
      "step": 5162
    },
    {
      "epoch": 0.6684144091659384,
      "grad_norm": 3.895947084589344,
      "learning_rate": 3.73967589834923e-07,
      "loss": 2.5234,
      "step": 5163
    },
    {
      "epoch": 0.6685438715732919,
      "grad_norm": 1.9194024118819928,
      "learning_rate": 3.737089254752218e-07,
      "loss": 2.2913,
      "step": 5164
    },
    {
      "epoch": 0.6686733339806453,
      "grad_norm": 2.751060996527409,
      "learning_rate": 3.73450312446816e-07,
      "loss": 2.4575,
      "step": 5165
    },
    {
      "epoch": 0.6686733339806453,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5171786546707153,
      "eval_runtime": 14.0054,
      "eval_samples_per_second": 3.142,
      "eval_steps_per_second": 0.214,
      "step": 5165
    },
    {
      "epoch": 0.6688027963879989,
      "grad_norm": 1.8718206248093348,
      "learning_rate": 3.731917508025326e-07,
      "loss": 2.5067,
      "step": 5166
    },
    {
      "epoch": 0.6689322587953523,
      "grad_norm": 2.315044819729028,
      "learning_rate": 3.7293324059518813e-07,
      "loss": 2.489,
      "step": 5167
    },
    {
      "epoch": 0.6690617212027058,
      "grad_norm": 2.24429911689122,
      "learning_rate": 3.7267478187758833e-07,
      "loss": 2.333,
      "step": 5168
    },
    {
      "epoch": 0.6691911836100592,
      "grad_norm": 1.8112099638411543,
      "learning_rate": 3.724163747025287e-07,
      "loss": 2.3835,
      "step": 5169
    },
    {
      "epoch": 0.6693206460174127,
      "grad_norm": 2.940289527489837,
      "learning_rate": 3.7215801912279413e-07,
      "loss": 2.1499,
      "step": 5170
    },
    {
      "epoch": 0.6693206460174127,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5149592161178589,
      "eval_runtime": 13.3265,
      "eval_samples_per_second": 3.302,
      "eval_steps_per_second": 0.225,
      "step": 5170
    },
    {
      "epoch": 0.6694501084247662,
      "grad_norm": 2.1048840969972122,
      "learning_rate": 3.7189971519115894e-07,
      "loss": 2.3857,
      "step": 5171
    },
    {
      "epoch": 0.6695795708321196,
      "grad_norm": 1.8588938143128404,
      "learning_rate": 3.716414629603872e-07,
      "loss": 2.8069,
      "step": 5172
    },
    {
      "epoch": 0.6697090332394731,
      "grad_norm": 1.4840976175737275,
      "learning_rate": 3.7138326248323225e-07,
      "loss": 2.7664,
      "step": 5173
    },
    {
      "epoch": 0.6698384956468265,
      "grad_norm": 1.4695531501696821,
      "learning_rate": 3.7112511381243623e-07,
      "loss": 2.4041,
      "step": 5174
    },
    {
      "epoch": 0.66996795805418,
      "grad_norm": 1.7802913169297638,
      "learning_rate": 3.7086701700073143e-07,
      "loss": 2.3398,
      "step": 5175
    },
    {
      "epoch": 0.66996795805418,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5157138109207153,
      "eval_runtime": 13.1817,
      "eval_samples_per_second": 3.338,
      "eval_steps_per_second": 0.228,
      "step": 5175
    },
    {
      "epoch": 0.6700974204615335,
      "grad_norm": 4.72140639057189,
      "learning_rate": 3.7060897210083947e-07,
      "loss": 2.8203,
      "step": 5176
    },
    {
      "epoch": 0.670226882868887,
      "grad_norm": 1.382085859640986,
      "learning_rate": 3.703509791654713e-07,
      "loss": 2.2473,
      "step": 5177
    },
    {
      "epoch": 0.6703563452762404,
      "grad_norm": 1.8636648435792593,
      "learning_rate": 3.700930382473272e-07,
      "loss": 2.054,
      "step": 5178
    },
    {
      "epoch": 0.6704858076835939,
      "grad_norm": 5.683917391002986,
      "learning_rate": 3.6983514939909674e-07,
      "loss": 2.9268,
      "step": 5179
    },
    {
      "epoch": 0.6706152700909473,
      "grad_norm": 1.872971604088809,
      "learning_rate": 3.695773126734589e-07,
      "loss": 2.6372,
      "step": 5180
    },
    {
      "epoch": 0.6706152700909473,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5172230005264282,
      "eval_runtime": 14.1143,
      "eval_samples_per_second": 3.117,
      "eval_steps_per_second": 0.213,
      "step": 5180
    },
    {
      "epoch": 0.6707447324983008,
      "grad_norm": 3.841026394448021,
      "learning_rate": 3.693195281230822e-07,
      "loss": 2.7532,
      "step": 5181
    },
    {
      "epoch": 0.6708741949056543,
      "grad_norm": 2.098262756639662,
      "learning_rate": 3.690617958006243e-07,
      "loss": 2.3953,
      "step": 5182
    },
    {
      "epoch": 0.6710036573130077,
      "grad_norm": 2.6526068040980113,
      "learning_rate": 3.6880411575873216e-07,
      "loss": 2.1807,
      "step": 5183
    },
    {
      "epoch": 0.6711331197203612,
      "grad_norm": 1.2352903410283615,
      "learning_rate": 3.6854648805004236e-07,
      "loss": 2.2712,
      "step": 5184
    },
    {
      "epoch": 0.6712625821277146,
      "grad_norm": 1.555460958812336,
      "learning_rate": 3.682889127271803e-07,
      "loss": 2.2493,
      "step": 5185
    },
    {
      "epoch": 0.6712625821277146,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5150257349014282,
      "eval_runtime": 14.3342,
      "eval_samples_per_second": 3.07,
      "eval_steps_per_second": 0.209,
      "step": 5185
    },
    {
      "epoch": 0.6713920445350682,
      "grad_norm": 1.2240766574091408,
      "learning_rate": 3.680313898427612e-07,
      "loss": 2.3901,
      "step": 5186
    },
    {
      "epoch": 0.6715215069424216,
      "grad_norm": 1.616171328978076,
      "learning_rate": 3.677739194493892e-07,
      "loss": 2.4905,
      "step": 5187
    },
    {
      "epoch": 0.6716509693497751,
      "grad_norm": 1.9645193182919447,
      "learning_rate": 3.675165015996579e-07,
      "loss": 2.2568,
      "step": 5188
    },
    {
      "epoch": 0.6717804317571285,
      "grad_norm": 1.3031903170890882,
      "learning_rate": 3.6725913634615003e-07,
      "loss": 2.3292,
      "step": 5189
    },
    {
      "epoch": 0.671909894164482,
      "grad_norm": 2.869303765077146,
      "learning_rate": 3.67001823741438e-07,
      "loss": 2.5576,
      "step": 5190
    },
    {
      "epoch": 0.671909894164482,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5164240598678589,
      "eval_runtime": 14.7051,
      "eval_samples_per_second": 2.992,
      "eval_steps_per_second": 0.204,
      "step": 5190
    },
    {
      "epoch": 0.6720393565718354,
      "grad_norm": 2.2019988217144504,
      "learning_rate": 3.667445638380826e-07,
      "loss": 2.2393,
      "step": 5191
    },
    {
      "epoch": 0.6721688189791889,
      "grad_norm": 2.028643793224081,
      "learning_rate": 3.664873566886347e-07,
      "loss": 2.5603,
      "step": 5192
    },
    {
      "epoch": 0.6722982813865424,
      "grad_norm": 2.340970224544167,
      "learning_rate": 3.662302023456339e-07,
      "loss": 2.4185,
      "step": 5193
    },
    {
      "epoch": 0.6724277437938958,
      "grad_norm": 2.3294127503985136,
      "learning_rate": 3.659731008616094e-07,
      "loss": 2.5471,
      "step": 5194
    },
    {
      "epoch": 0.6725572062012493,
      "grad_norm": 2.0852620322255233,
      "learning_rate": 3.657160522890794e-07,
      "loss": 2.3655,
      "step": 5195
    },
    {
      "epoch": 0.6725572062012493,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5157581567764282,
      "eval_runtime": 14.3533,
      "eval_samples_per_second": 3.066,
      "eval_steps_per_second": 0.209,
      "step": 5195
    },
    {
      "epoch": 0.6726866686086028,
      "grad_norm": 2.6355346694446693,
      "learning_rate": 3.654590566805511e-07,
      "loss": 2.3953,
      "step": 5196
    },
    {
      "epoch": 0.6728161310159563,
      "grad_norm": 3.0306316730134166,
      "learning_rate": 3.6520211408852093e-07,
      "loss": 2.4951,
      "step": 5197
    },
    {
      "epoch": 0.6729455934233097,
      "grad_norm": 1.2556637892813427,
      "learning_rate": 3.64945224565475e-07,
      "loss": 2.2407,
      "step": 5198
    },
    {
      "epoch": 0.6730750558306632,
      "grad_norm": 1.959171994137629,
      "learning_rate": 3.6468838816388795e-07,
      "loss": 2.5046,
      "step": 5199
    },
    {
      "epoch": 0.6732045182380166,
      "grad_norm": 3.654400560571032,
      "learning_rate": 3.6443160493622407e-07,
      "loss": 2.697,
      "step": 5200
    },
    {
      "epoch": 0.6732045182380166,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5178000926971436,
      "eval_runtime": 14.1994,
      "eval_samples_per_second": 3.099,
      "eval_steps_per_second": 0.211,
      "step": 5200
    },
    {
      "epoch": 0.6733339806453701,
      "grad_norm": 1.4555148836636034,
      "learning_rate": 3.6417487493493634e-07,
      "loss": 2.2358,
      "step": 5201
    },
    {
      "epoch": 0.6734634430527235,
      "grad_norm": 1.5211274824094976,
      "learning_rate": 3.6391819821246685e-07,
      "loss": 2.3491,
      "step": 5202
    },
    {
      "epoch": 0.673592905460077,
      "grad_norm": 2.046669640509887,
      "learning_rate": 3.6366157482124747e-07,
      "loss": 2.3335,
      "step": 5203
    },
    {
      "epoch": 0.6737223678674304,
      "grad_norm": 1.767659539458005,
      "learning_rate": 3.634050048136983e-07,
      "loss": 2.4586,
      "step": 5204
    },
    {
      "epoch": 0.673851830274784,
      "grad_norm": 1.7714915188330898,
      "learning_rate": 3.631484882422292e-07,
      "loss": 2.4551,
      "step": 5205
    },
    {
      "epoch": 0.673851830274784,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5164684057235718,
      "eval_runtime": 14.5253,
      "eval_samples_per_second": 3.029,
      "eval_steps_per_second": 0.207,
      "step": 5205
    },
    {
      "epoch": 0.6739812926821375,
      "grad_norm": 2.049300693559449,
      "learning_rate": 3.6289202515923936e-07,
      "loss": 2.3489,
      "step": 5206
    },
    {
      "epoch": 0.6741107550894909,
      "grad_norm": 1.7485913639195079,
      "learning_rate": 3.6263561561711554e-07,
      "loss": 2.4399,
      "step": 5207
    },
    {
      "epoch": 0.6742402174968444,
      "grad_norm": 2.1131034879887527,
      "learning_rate": 3.6237925966823547e-07,
      "loss": 2.3411,
      "step": 5208
    },
    {
      "epoch": 0.6743696799041978,
      "grad_norm": 1.8245187362374382,
      "learning_rate": 3.6212295736496426e-07,
      "loss": 2.3673,
      "step": 5209
    },
    {
      "epoch": 0.6744991423115513,
      "grad_norm": 1.2856899072052732,
      "learning_rate": 3.618667087596574e-07,
      "loss": 2.249,
      "step": 5210
    },
    {
      "epoch": 0.6744991423115513,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5185546875,
      "eval_runtime": 14.5418,
      "eval_samples_per_second": 3.026,
      "eval_steps_per_second": 0.206,
      "step": 5210
    },
    {
      "epoch": 0.6746286047189047,
      "grad_norm": 1.6513603051575372,
      "learning_rate": 3.61610513904659e-07,
      "loss": 2.4089,
      "step": 5211
    },
    {
      "epoch": 0.6747580671262582,
      "grad_norm": 2.258679709221032,
      "learning_rate": 3.613543728523018e-07,
      "loss": 2.625,
      "step": 5212
    },
    {
      "epoch": 0.6748875295336116,
      "grad_norm": 1.426089270343243,
      "learning_rate": 3.610982856549075e-07,
      "loss": 2.3296,
      "step": 5213
    },
    {
      "epoch": 0.6750169919409651,
      "grad_norm": 1.9765930465741033,
      "learning_rate": 3.6084225236478767e-07,
      "loss": 2.1831,
      "step": 5214
    },
    {
      "epoch": 0.6751464543483187,
      "grad_norm": 2.63646888398566,
      "learning_rate": 3.6058627303424174e-07,
      "loss": 2.374,
      "step": 5215
    },
    {
      "epoch": 0.6751464543483187,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.517777919769287,
      "eval_runtime": 13.9994,
      "eval_samples_per_second": 3.143,
      "eval_steps_per_second": 0.214,
      "step": 5215
    },
    {
      "epoch": 0.6752759167556721,
      "grad_norm": 1.5671852668968798,
      "learning_rate": 3.6033034771555923e-07,
      "loss": 2.3293,
      "step": 5216
    },
    {
      "epoch": 0.6754053791630256,
      "grad_norm": 1.359830987904769,
      "learning_rate": 3.6007447646101786e-07,
      "loss": 2.405,
      "step": 5217
    },
    {
      "epoch": 0.675534841570379,
      "grad_norm": 1.7563274816105932,
      "learning_rate": 3.598186593228842e-07,
      "loss": 2.3501,
      "step": 5218
    },
    {
      "epoch": 0.6756643039777325,
      "grad_norm": 2.712288861932896,
      "learning_rate": 3.5956289635341465e-07,
      "loss": 2.8359,
      "step": 5219
    },
    {
      "epoch": 0.6757937663850859,
      "grad_norm": 1.4580046995383933,
      "learning_rate": 3.5930718760485333e-07,
      "loss": 2.3892,
      "step": 5220
    },
    {
      "epoch": 0.6757937663850859,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5184658765792847,
      "eval_runtime": 14.2688,
      "eval_samples_per_second": 3.084,
      "eval_steps_per_second": 0.21,
      "step": 5220
    },
    {
      "epoch": 0.6759232287924394,
      "grad_norm": 2.210650569872333,
      "learning_rate": 3.590515331294346e-07,
      "loss": 2.5549,
      "step": 5221
    },
    {
      "epoch": 0.6760526911997928,
      "grad_norm": 1.5490546548895017,
      "learning_rate": 3.5879593297938055e-07,
      "loss": 2.332,
      "step": 5222
    },
    {
      "epoch": 0.6761821536071463,
      "grad_norm": 2.412738058317075,
      "learning_rate": 3.5854038720690305e-07,
      "loss": 2.5547,
      "step": 5223
    },
    {
      "epoch": 0.6763116160144997,
      "grad_norm": 2.3987816391579657,
      "learning_rate": 3.582848958642024e-07,
      "loss": 2.481,
      "step": 5224
    },
    {
      "epoch": 0.6764410784218533,
      "grad_norm": 2.3761614745260227,
      "learning_rate": 3.5802945900346763e-07,
      "loss": 2.5806,
      "step": 5225
    },
    {
      "epoch": 0.6764410784218533,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.51708984375,
      "eval_runtime": 15.0319,
      "eval_samples_per_second": 2.927,
      "eval_steps_per_second": 0.2,
      "step": 5225
    },
    {
      "epoch": 0.6765705408292068,
      "grad_norm": 1.999288599859236,
      "learning_rate": 3.5777407667687723e-07,
      "loss": 2.3669,
      "step": 5226
    },
    {
      "epoch": 0.6767000032365602,
      "grad_norm": 1.9038521953853287,
      "learning_rate": 3.575187489365979e-07,
      "loss": 2.5393,
      "step": 5227
    },
    {
      "epoch": 0.6768294656439137,
      "grad_norm": 2.655885107538993,
      "learning_rate": 3.572634758347861e-07,
      "loss": 2.3694,
      "step": 5228
    },
    {
      "epoch": 0.6769589280512671,
      "grad_norm": 1.2756381051525916,
      "learning_rate": 3.5700825742358595e-07,
      "loss": 2.2866,
      "step": 5229
    },
    {
      "epoch": 0.6770883904586206,
      "grad_norm": 1.735971958165332,
      "learning_rate": 3.56753093755131e-07,
      "loss": 2.4614,
      "step": 5230
    },
    {
      "epoch": 0.6770883904586206,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5171120166778564,
      "eval_runtime": 12.9492,
      "eval_samples_per_second": 3.398,
      "eval_steps_per_second": 0.232,
      "step": 5230
    },
    {
      "epoch": 0.677217852865974,
      "grad_norm": 1.7071448357533605,
      "learning_rate": 3.5649798488154365e-07,
      "loss": 2.2178,
      "step": 5231
    },
    {
      "epoch": 0.6773473152733275,
      "grad_norm": 3.4925803110194904,
      "learning_rate": 3.5624293085493556e-07,
      "loss": 2.7458,
      "step": 5232
    },
    {
      "epoch": 0.6774767776806809,
      "grad_norm": 2.7030568290693404,
      "learning_rate": 3.5598793172740597e-07,
      "loss": 2.6646,
      "step": 5233
    },
    {
      "epoch": 0.6776062400880344,
      "grad_norm": 2.259110266225499,
      "learning_rate": 3.557329875510445e-07,
      "loss": 2.3389,
      "step": 5234
    },
    {
      "epoch": 0.677735702495388,
      "grad_norm": 2.984263218140695,
      "learning_rate": 3.554780983779276e-07,
      "loss": 2.645,
      "step": 5235
    },
    {
      "epoch": 0.677735702495388,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5136497020721436,
      "eval_runtime": 15.1559,
      "eval_samples_per_second": 2.903,
      "eval_steps_per_second": 0.198,
      "step": 5235
    },
    {
      "epoch": 0.6778651649027414,
      "grad_norm": 1.6363349893759631,
      "learning_rate": 3.552232642601221e-07,
      "loss": 2.4404,
      "step": 5236
    },
    {
      "epoch": 0.6779946273100949,
      "grad_norm": 1.2916843377566822,
      "learning_rate": 3.549684852496832e-07,
      "loss": 2.3035,
      "step": 5237
    },
    {
      "epoch": 0.6781240897174483,
      "grad_norm": 1.2762116165880084,
      "learning_rate": 3.5471376139865427e-07,
      "loss": 2.2085,
      "step": 5238
    },
    {
      "epoch": 0.6782535521248018,
      "grad_norm": 2.239098902198714,
      "learning_rate": 3.5445909275906826e-07,
      "loss": 2.4478,
      "step": 5239
    },
    {
      "epoch": 0.6783830145321552,
      "grad_norm": 2.362653082394167,
      "learning_rate": 3.542044793829461e-07,
      "loss": 2.4512,
      "step": 5240
    },
    {
      "epoch": 0.6783830145321552,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5144709348678589,
      "eval_runtime": 14.2809,
      "eval_samples_per_second": 3.081,
      "eval_steps_per_second": 0.21,
      "step": 5240
    },
    {
      "epoch": 0.6785124769395087,
      "grad_norm": 1.6466598228001548,
      "learning_rate": 3.5394992132229767e-07,
      "loss": 2.3201,
      "step": 5241
    },
    {
      "epoch": 0.6786419393468621,
      "grad_norm": 1.7901975282499403,
      "learning_rate": 3.536954186291219e-07,
      "loss": 2.2375,
      "step": 5242
    },
    {
      "epoch": 0.6787714017542156,
      "grad_norm": 1.4214197741100385,
      "learning_rate": 3.534409713554057e-07,
      "loss": 2.2954,
      "step": 5243
    },
    {
      "epoch": 0.6789008641615691,
      "grad_norm": 2.042846127205418,
      "learning_rate": 3.5318657955312573e-07,
      "loss": 2.7146,
      "step": 5244
    },
    {
      "epoch": 0.6790303265689226,
      "grad_norm": 3.5721648727622117,
      "learning_rate": 3.5293224327424625e-07,
      "loss": 2.3306,
      "step": 5245
    },
    {
      "epoch": 0.6790303265689226,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5171786546707153,
      "eval_runtime": 14.7304,
      "eval_samples_per_second": 2.987,
      "eval_steps_per_second": 0.204,
      "step": 5245
    },
    {
      "epoch": 0.679159788976276,
      "grad_norm": 3.0066869720647116,
      "learning_rate": 3.526779625707204e-07,
      "loss": 2.4009,
      "step": 5246
    },
    {
      "epoch": 0.6792892513836295,
      "grad_norm": 2.1355296737675027,
      "learning_rate": 3.5242373749449056e-07,
      "loss": 2.2749,
      "step": 5247
    },
    {
      "epoch": 0.679418713790983,
      "grad_norm": 2.1017472222072735,
      "learning_rate": 3.52169568097487e-07,
      "loss": 2.3696,
      "step": 5248
    },
    {
      "epoch": 0.6795481761983364,
      "grad_norm": 3.6488145130240808,
      "learning_rate": 3.519154544316294e-07,
      "loss": 2.4302,
      "step": 5249
    },
    {
      "epoch": 0.6796776386056899,
      "grad_norm": 1.880456897453718,
      "learning_rate": 3.5166139654882525e-07,
      "loss": 2.41,
      "step": 5250
    },
    {
      "epoch": 0.6796776386056899,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5179332494735718,
      "eval_runtime": 15.2384,
      "eval_samples_per_second": 2.887,
      "eval_steps_per_second": 0.197,
      "step": 5250
    },
    {
      "epoch": 0.6798071010130433,
      "grad_norm": 3.1613642716518733,
      "learning_rate": 3.514073945009714e-07,
      "loss": 2.4265,
      "step": 5251
    },
    {
      "epoch": 0.6799365634203968,
      "grad_norm": 1.8057933957080892,
      "learning_rate": 3.5115344833995273e-07,
      "loss": 2.1995,
      "step": 5252
    },
    {
      "epoch": 0.6800660258277502,
      "grad_norm": 1.4441837457786255,
      "learning_rate": 3.508995581176425e-07,
      "loss": 2.3928,
      "step": 5253
    },
    {
      "epoch": 0.6801954882351038,
      "grad_norm": 3.874670791895159,
      "learning_rate": 3.506457238859033e-07,
      "loss": 2.4443,
      "step": 5254
    },
    {
      "epoch": 0.6803249506424572,
      "grad_norm": 1.7297980002020237,
      "learning_rate": 3.5039194569658625e-07,
      "loss": 2.3359,
      "step": 5255
    },
    {
      "epoch": 0.6803249506424572,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5163795948028564,
      "eval_runtime": 14.1876,
      "eval_samples_per_second": 3.101,
      "eval_steps_per_second": 0.211,
      "step": 5255
    },
    {
      "epoch": 0.6804544130498107,
      "grad_norm": 1.3142094783061335,
      "learning_rate": 3.5013822360153024e-07,
      "loss": 2.3228,
      "step": 5256
    },
    {
      "epoch": 0.6805838754571641,
      "grad_norm": 1.6543195155186603,
      "learning_rate": 3.4988455765256326e-07,
      "loss": 2.4092,
      "step": 5257
    },
    {
      "epoch": 0.6807133378645176,
      "grad_norm": 1.8122925535389094,
      "learning_rate": 3.4963094790150154e-07,
      "loss": 2.3245,
      "step": 5258
    },
    {
      "epoch": 0.680842800271871,
      "grad_norm": 1.600405426267812,
      "learning_rate": 3.493773944001501e-07,
      "loss": 2.2654,
      "step": 5259
    },
    {
      "epoch": 0.6809722626792245,
      "grad_norm": 2.1278032642846902,
      "learning_rate": 3.4912389720030275e-07,
      "loss": 2.354,
      "step": 5260
    },
    {
      "epoch": 0.6809722626792245,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5129616260528564,
      "eval_runtime": 14.4426,
      "eval_samples_per_second": 3.047,
      "eval_steps_per_second": 0.208,
      "step": 5260
    },
    {
      "epoch": 0.681101725086578,
      "grad_norm": 1.9564449809840208,
      "learning_rate": 3.4887045635374096e-07,
      "loss": 2.5923,
      "step": 5261
    },
    {
      "epoch": 0.6812311874939314,
      "grad_norm": 2.8503831455817883,
      "learning_rate": 3.486170719122356e-07,
      "loss": 2.4429,
      "step": 5262
    },
    {
      "epoch": 0.6813606499012849,
      "grad_norm": 2.6314135290550316,
      "learning_rate": 3.483637439275453e-07,
      "loss": 2.4775,
      "step": 5263
    },
    {
      "epoch": 0.6814901123086384,
      "grad_norm": 4.32835843404821,
      "learning_rate": 3.4811047245141713e-07,
      "loss": 2.4233,
      "step": 5264
    },
    {
      "epoch": 0.6816195747159919,
      "grad_norm": 1.2409452199406177,
      "learning_rate": 3.478572575355876e-07,
      "loss": 2.2716,
      "step": 5265
    },
    {
      "epoch": 0.6816195747159919,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.515669345855713,
      "eval_runtime": 14.0382,
      "eval_samples_per_second": 3.134,
      "eval_steps_per_second": 0.214,
      "step": 5265
    },
    {
      "epoch": 0.6817490371233453,
      "grad_norm": 3.1463823453908804,
      "learning_rate": 3.4760409923178036e-07,
      "loss": 2.7,
      "step": 5266
    },
    {
      "epoch": 0.6818784995306988,
      "grad_norm": 1.4538289416347718,
      "learning_rate": 3.4735099759170864e-07,
      "loss": 2.4497,
      "step": 5267
    },
    {
      "epoch": 0.6820079619380522,
      "grad_norm": 1.3960796627036527,
      "learning_rate": 3.470979526670734e-07,
      "loss": 2.2971,
      "step": 5268
    },
    {
      "epoch": 0.6821374243454057,
      "grad_norm": 1.5680110984654894,
      "learning_rate": 3.4684496450956386e-07,
      "loss": 2.4592,
      "step": 5269
    },
    {
      "epoch": 0.6822668867527591,
      "grad_norm": 1.775844244948726,
      "learning_rate": 3.465920331708586e-07,
      "loss": 2.4331,
      "step": 5270
    },
    {
      "epoch": 0.6822668867527591,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.517045497894287,
      "eval_runtime": 13.8252,
      "eval_samples_per_second": 3.183,
      "eval_steps_per_second": 0.217,
      "step": 5270
    },
    {
      "epoch": 0.6823963491601126,
      "grad_norm": 3.0064942558009884,
      "learning_rate": 3.4633915870262323e-07,
      "loss": 2.4385,
      "step": 5271
    },
    {
      "epoch": 0.6825258115674661,
      "grad_norm": 1.756906264922669,
      "learning_rate": 3.4608634115651333e-07,
      "loss": 2.3374,
      "step": 5272
    },
    {
      "epoch": 0.6826552739748195,
      "grad_norm": 2.038579186850268,
      "learning_rate": 3.4583358058417145e-07,
      "loss": 2.3347,
      "step": 5273
    },
    {
      "epoch": 0.6827847363821731,
      "grad_norm": 1.141392873884542,
      "learning_rate": 3.455808770372289e-07,
      "loss": 2.1926,
      "step": 5274
    },
    {
      "epoch": 0.6829141987895265,
      "grad_norm": 1.6832690108620663,
      "learning_rate": 3.453282305673061e-07,
      "loss": 2.0496,
      "step": 5275
    },
    {
      "epoch": 0.6829141987895265,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.519331455230713,
      "eval_runtime": 13.3241,
      "eval_samples_per_second": 3.302,
      "eval_steps_per_second": 0.225,
      "step": 5275
    },
    {
      "epoch": 0.68304366119688,
      "grad_norm": 1.3132853117623964,
      "learning_rate": 3.450756412260107e-07,
      "loss": 2.2083,
      "step": 5276
    },
    {
      "epoch": 0.6831731236042334,
      "grad_norm": 1.8641173578625694,
      "learning_rate": 3.4482310906493925e-07,
      "loss": 2.4832,
      "step": 5277
    },
    {
      "epoch": 0.6833025860115869,
      "grad_norm": 1.5046968889959005,
      "learning_rate": 3.445706341356771e-07,
      "loss": 2.4243,
      "step": 5278
    },
    {
      "epoch": 0.6834320484189403,
      "grad_norm": 1.3426248272115668,
      "learning_rate": 3.4431821648979684e-07,
      "loss": 2.3423,
      "step": 5279
    },
    {
      "epoch": 0.6835615108262938,
      "grad_norm": 1.7858745758616192,
      "learning_rate": 3.4406585617886015e-07,
      "loss": 2.4058,
      "step": 5280
    },
    {
      "epoch": 0.6835615108262938,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5186434984207153,
      "eval_runtime": 14.1945,
      "eval_samples_per_second": 3.1,
      "eval_steps_per_second": 0.211,
      "step": 5280
    },
    {
      "epoch": 0.6836909732336472,
      "grad_norm": 2.0156811300806248,
      "learning_rate": 3.438135532544162e-07,
      "loss": 2.4004,
      "step": 5281
    },
    {
      "epoch": 0.6838204356410007,
      "grad_norm": 1.2292898591618622,
      "learning_rate": 3.4356130776800347e-07,
      "loss": 2.1897,
      "step": 5282
    },
    {
      "epoch": 0.6839498980483542,
      "grad_norm": 1.4536210374840777,
      "learning_rate": 3.4330911977114843e-07,
      "loss": 2.3479,
      "step": 5283
    },
    {
      "epoch": 0.6840793604557077,
      "grad_norm": 1.5917119251743426,
      "learning_rate": 3.430569893153653e-07,
      "loss": 2.321,
      "step": 5284
    },
    {
      "epoch": 0.6842088228630612,
      "grad_norm": 1.194981712229032,
      "learning_rate": 3.428049164521564e-07,
      "loss": 2.2852,
      "step": 5285
    },
    {
      "epoch": 0.6842088228630612,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.517134189605713,
      "eval_runtime": 14.8292,
      "eval_samples_per_second": 2.967,
      "eval_steps_per_second": 0.202,
      "step": 5285
    },
    {
      "epoch": 0.6843382852704146,
      "grad_norm": 1.4564356124850828,
      "learning_rate": 3.4255290123301357e-07,
      "loss": 2.3604,
      "step": 5286
    },
    {
      "epoch": 0.6844677476777681,
      "grad_norm": 1.5239257772120258,
      "learning_rate": 3.423009437094153e-07,
      "loss": 2.2656,
      "step": 5287
    },
    {
      "epoch": 0.6845972100851215,
      "grad_norm": 2.1551809473036747,
      "learning_rate": 3.420490439328296e-07,
      "loss": 2.6562,
      "step": 5288
    },
    {
      "epoch": 0.684726672492475,
      "grad_norm": 1.4104229389410412,
      "learning_rate": 3.417972019547116e-07,
      "loss": 2.3862,
      "step": 5289
    },
    {
      "epoch": 0.6848561348998284,
      "grad_norm": 1.780957551376093,
      "learning_rate": 3.4154541782650556e-07,
      "loss": 2.3049,
      "step": 5290
    },
    {
      "epoch": 0.6848561348998284,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5157803297042847,
      "eval_runtime": 13.7832,
      "eval_samples_per_second": 3.192,
      "eval_steps_per_second": 0.218,
      "step": 5290
    },
    {
      "epoch": 0.6849855973071819,
      "grad_norm": 1.5735364381034245,
      "learning_rate": 3.4129369159964327e-07,
      "loss": 2.3372,
      "step": 5291
    },
    {
      "epoch": 0.6851150597145353,
      "grad_norm": 1.717616490730774,
      "learning_rate": 3.4104202332554475e-07,
      "loss": 2.4036,
      "step": 5292
    },
    {
      "epoch": 0.6852445221218889,
      "grad_norm": 1.5051676450433018,
      "learning_rate": 3.4079041305561876e-07,
      "loss": 2.3101,
      "step": 5293
    },
    {
      "epoch": 0.6853739845292424,
      "grad_norm": 1.3031780948838116,
      "learning_rate": 3.405388608412613e-07,
      "loss": 2.3641,
      "step": 5294
    },
    {
      "epoch": 0.6855034469365958,
      "grad_norm": 1.5857000372699583,
      "learning_rate": 3.4028736673385755e-07,
      "loss": 2.3579,
      "step": 5295
    },
    {
      "epoch": 0.6855034469365958,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5150257349014282,
      "eval_runtime": 15.0757,
      "eval_samples_per_second": 2.919,
      "eval_steps_per_second": 0.199,
      "step": 5295
    },
    {
      "epoch": 0.6856329093439493,
      "grad_norm": 2.1342613905983585,
      "learning_rate": 3.4003593078478e-07,
      "loss": 2.449,
      "step": 5296
    },
    {
      "epoch": 0.6857623717513027,
      "grad_norm": 1.4041942243521495,
      "learning_rate": 3.397845530453893e-07,
      "loss": 2.3462,
      "step": 5297
    },
    {
      "epoch": 0.6858918341586562,
      "grad_norm": 1.3320730837144454,
      "learning_rate": 3.39533233567035e-07,
      "loss": 2.2778,
      "step": 5298
    },
    {
      "epoch": 0.6860212965660096,
      "grad_norm": 1.7325679053807328,
      "learning_rate": 3.392819724010535e-07,
      "loss": 2.4946,
      "step": 5299
    },
    {
      "epoch": 0.6861507589733631,
      "grad_norm": 1.2735286825120988,
      "learning_rate": 3.3903076959877047e-07,
      "loss": 2.1953,
      "step": 5300
    },
    {
      "epoch": 0.6861507589733631,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5143154859542847,
      "eval_runtime": 14.5467,
      "eval_samples_per_second": 3.025,
      "eval_steps_per_second": 0.206,
      "step": 5300
    },
    {
      "epoch": 0.6862802213807165,
      "grad_norm": 2.384971743131937,
      "learning_rate": 3.3877962521149956e-07,
      "loss": 2.3931,
      "step": 5301
    },
    {
      "epoch": 0.68640968378807,
      "grad_norm": 2.147038904908983,
      "learning_rate": 3.385285392905412e-07,
      "loss": 2.4946,
      "step": 5302
    },
    {
      "epoch": 0.6865391461954236,
      "grad_norm": 1.986882172640929,
      "learning_rate": 3.3827751188718533e-07,
      "loss": 2.2798,
      "step": 5303
    },
    {
      "epoch": 0.686668608602777,
      "grad_norm": 1.5409879420459855,
      "learning_rate": 3.380265430527091e-07,
      "loss": 2.5415,
      "step": 5304
    },
    {
      "epoch": 0.6867980710101305,
      "grad_norm": 1.8687361414984769,
      "learning_rate": 3.3777563283837805e-07,
      "loss": 2.2322,
      "step": 5305
    },
    {
      "epoch": 0.6867980710101305,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5142711400985718,
      "eval_runtime": 18.8418,
      "eval_samples_per_second": 2.335,
      "eval_steps_per_second": 0.159,
      "step": 5305
    },
    {
      "epoch": 0.6869275334174839,
      "grad_norm": 2.361479109078417,
      "learning_rate": 3.375247812954463e-07,
      "loss": 2.3794,
      "step": 5306
    },
    {
      "epoch": 0.6870569958248374,
      "grad_norm": 1.3951456832389841,
      "learning_rate": 3.372739884751543e-07,
      "loss": 2.3733,
      "step": 5307
    },
    {
      "epoch": 0.6871864582321908,
      "grad_norm": 2.028005489437002,
      "learning_rate": 3.3702325442873217e-07,
      "loss": 2.168,
      "step": 5308
    },
    {
      "epoch": 0.6873159206395443,
      "grad_norm": 2.234759749833977,
      "learning_rate": 3.367725792073976e-07,
      "loss": 2.3535,
      "step": 5309
    },
    {
      "epoch": 0.6874453830468977,
      "grad_norm": 1.8700754833814541,
      "learning_rate": 3.3652196286235553e-07,
      "loss": 2.3005,
      "step": 5310
    },
    {
      "epoch": 0.6874453830468977,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5149592161178589,
      "eval_runtime": 14.6834,
      "eval_samples_per_second": 2.997,
      "eval_steps_per_second": 0.204,
      "step": 5310
    },
    {
      "epoch": 0.6875748454542512,
      "grad_norm": 1.8876010473368234,
      "learning_rate": 3.3627140544479997e-07,
      "loss": 2.5183,
      "step": 5311
    },
    {
      "epoch": 0.6877043078616046,
      "grad_norm": 1.6051176000616365,
      "learning_rate": 3.3602090700591215e-07,
      "loss": 2.2588,
      "step": 5312
    },
    {
      "epoch": 0.6878337702689582,
      "grad_norm": 1.4515583689326508,
      "learning_rate": 3.357704675968611e-07,
      "loss": 2.2747,
      "step": 5313
    },
    {
      "epoch": 0.6879632326763117,
      "grad_norm": 1.5844909081318954,
      "learning_rate": 3.355200872688048e-07,
      "loss": 2.4324,
      "step": 5314
    },
    {
      "epoch": 0.6880926950836651,
      "grad_norm": 3.217833584365405,
      "learning_rate": 3.3526976607288793e-07,
      "loss": 2.5137,
      "step": 5315
    },
    {
      "epoch": 0.6880926950836651,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5150035619735718,
      "eval_runtime": 15.5964,
      "eval_samples_per_second": 2.821,
      "eval_steps_per_second": 0.192,
      "step": 5315
    },
    {
      "epoch": 0.6882221574910186,
      "grad_norm": 1.5114742460583046,
      "learning_rate": 3.35019504060244e-07,
      "loss": 2.4365,
      "step": 5316
    },
    {
      "epoch": 0.688351619898372,
      "grad_norm": 3.2874319073919347,
      "learning_rate": 3.3476930128199387e-07,
      "loss": 2.3691,
      "step": 5317
    },
    {
      "epoch": 0.6884810823057255,
      "grad_norm": 2.256535391542108,
      "learning_rate": 3.345191577892469e-07,
      "loss": 2.4149,
      "step": 5318
    },
    {
      "epoch": 0.6886105447130789,
      "grad_norm": 3.429347989522284,
      "learning_rate": 3.3426907363309976e-07,
      "loss": 2.3914,
      "step": 5319
    },
    {
      "epoch": 0.6887400071204324,
      "grad_norm": 1.0874046382361477,
      "learning_rate": 3.3401904886463684e-07,
      "loss": 2.1731,
      "step": 5320
    },
    {
      "epoch": 0.6887400071204324,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5172008275985718,
      "eval_runtime": 19.1102,
      "eval_samples_per_second": 2.302,
      "eval_steps_per_second": 0.157,
      "step": 5320
    },
    {
      "epoch": 0.6888694695277858,
      "grad_norm": 1.3805835847510348,
      "learning_rate": 3.337690835349314e-07,
      "loss": 2.3584,
      "step": 5321
    },
    {
      "epoch": 0.6889989319351393,
      "grad_norm": 1.5954479129584558,
      "learning_rate": 3.335191776950434e-07,
      "loss": 2.3792,
      "step": 5322
    },
    {
      "epoch": 0.6891283943424928,
      "grad_norm": 2.8428686669989243,
      "learning_rate": 3.332693313960216e-07,
      "loss": 2.4805,
      "step": 5323
    },
    {
      "epoch": 0.6892578567498463,
      "grad_norm": 3.0839326029096332,
      "learning_rate": 3.33019544688902e-07,
      "loss": 2.3149,
      "step": 5324
    },
    {
      "epoch": 0.6893873191571998,
      "grad_norm": 1.7666352164113863,
      "learning_rate": 3.3276981762470835e-07,
      "loss": 2.5637,
      "step": 5325
    },
    {
      "epoch": 0.6893873191571998,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.514359951019287,
      "eval_runtime": 15.8649,
      "eval_samples_per_second": 2.773,
      "eval_steps_per_second": 0.189,
      "step": 5325
    },
    {
      "epoch": 0.6895167815645532,
      "grad_norm": 1.6564306111926075,
      "learning_rate": 3.325201502544527e-07,
      "loss": 2.3599,
      "step": 5326
    },
    {
      "epoch": 0.6896462439719067,
      "grad_norm": 1.5945678877481886,
      "learning_rate": 3.3227054262913484e-07,
      "loss": 2.2998,
      "step": 5327
    },
    {
      "epoch": 0.6897757063792601,
      "grad_norm": 1.3550227880786858,
      "learning_rate": 3.320209947997418e-07,
      "loss": 2.3027,
      "step": 5328
    },
    {
      "epoch": 0.6899051687866136,
      "grad_norm": 1.3086287694967784,
      "learning_rate": 3.3177150681724944e-07,
      "loss": 2.499,
      "step": 5329
    },
    {
      "epoch": 0.690034631193967,
      "grad_norm": 1.650943473562167,
      "learning_rate": 3.3152207873261983e-07,
      "loss": 2.1572,
      "step": 5330
    },
    {
      "epoch": 0.690034631193967,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5136497020721436,
      "eval_runtime": 13.9794,
      "eval_samples_per_second": 3.147,
      "eval_steps_per_second": 0.215,
      "step": 5330
    },
    {
      "epoch": 0.6901640936013205,
      "grad_norm": 1.2915484217800368,
      "learning_rate": 3.312727105968042e-07,
      "loss": 2.3193,
      "step": 5331
    },
    {
      "epoch": 0.690293556008674,
      "grad_norm": 3.7580423761323236,
      "learning_rate": 3.3102340246074115e-07,
      "loss": 3.0388,
      "step": 5332
    },
    {
      "epoch": 0.6904230184160275,
      "grad_norm": 2.8934251653665,
      "learning_rate": 3.3077415437535664e-07,
      "loss": 2.313,
      "step": 5333
    },
    {
      "epoch": 0.6905524808233809,
      "grad_norm": 1.7966369994170994,
      "learning_rate": 3.3052496639156495e-07,
      "loss": 2.3757,
      "step": 5334
    },
    {
      "epoch": 0.6906819432307344,
      "grad_norm": 2.068137813876219,
      "learning_rate": 3.3027583856026767e-07,
      "loss": 2.4111,
      "step": 5335
    },
    {
      "epoch": 0.6906819432307344,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5142933130264282,
      "eval_runtime": 13.6879,
      "eval_samples_per_second": 3.215,
      "eval_steps_per_second": 0.219,
      "step": 5335
    },
    {
      "epoch": 0.6908114056380878,
      "grad_norm": 2.2781863072102597,
      "learning_rate": 3.3002677093235387e-07,
      "loss": 2.6138,
      "step": 5336
    },
    {
      "epoch": 0.6909408680454413,
      "grad_norm": 2.586761375232325,
      "learning_rate": 3.297777635587012e-07,
      "loss": 2.5593,
      "step": 5337
    },
    {
      "epoch": 0.6910703304527948,
      "grad_norm": 1.746465970557761,
      "learning_rate": 3.29528816490174e-07,
      "loss": 2.5435,
      "step": 5338
    },
    {
      "epoch": 0.6911997928601482,
      "grad_norm": 1.2455390194498344,
      "learning_rate": 3.2927992977762516e-07,
      "loss": 2.168,
      "step": 5339
    },
    {
      "epoch": 0.6913292552675017,
      "grad_norm": 2.9018345175822784,
      "learning_rate": 3.290311034718946e-07,
      "loss": 2.3441,
      "step": 5340
    },
    {
      "epoch": 0.6913292552675017,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5108309984207153,
      "eval_runtime": 15.1055,
      "eval_samples_per_second": 2.913,
      "eval_steps_per_second": 0.199,
      "step": 5340
    },
    {
      "epoch": 0.6914587176748551,
      "grad_norm": 1.3597390088524353,
      "learning_rate": 3.2878233762381003e-07,
      "loss": 2.4189,
      "step": 5341
    },
    {
      "epoch": 0.6915881800822087,
      "grad_norm": 1.8226028613013434,
      "learning_rate": 3.285336322841872e-07,
      "loss": 2.3486,
      "step": 5342
    },
    {
      "epoch": 0.6917176424895621,
      "grad_norm": 1.808497782035811,
      "learning_rate": 3.2828498750382887e-07,
      "loss": 2.4976,
      "step": 5343
    },
    {
      "epoch": 0.6918471048969156,
      "grad_norm": 3.7786067130303653,
      "learning_rate": 3.280364033335262e-07,
      "loss": 2.5557,
      "step": 5344
    },
    {
      "epoch": 0.691976567304269,
      "grad_norm": 1.7876273026010536,
      "learning_rate": 3.2778787982405737e-07,
      "loss": 2.5051,
      "step": 5345
    },
    {
      "epoch": 0.691976567304269,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.513627529144287,
      "eval_runtime": 14.1363,
      "eval_samples_per_second": 3.113,
      "eval_steps_per_second": 0.212,
      "step": 5345
    },
    {
      "epoch": 0.6921060297116225,
      "grad_norm": 1.6705348993732794,
      "learning_rate": 3.2753941702618807e-07,
      "loss": 2.2898,
      "step": 5346
    },
    {
      "epoch": 0.692235492118976,
      "grad_norm": 1.9443960265234832,
      "learning_rate": 3.272910149906723e-07,
      "loss": 2.3772,
      "step": 5347
    },
    {
      "epoch": 0.6923649545263294,
      "grad_norm": 1.9638856483260487,
      "learning_rate": 3.2704267376825073e-07,
      "loss": 2.3394,
      "step": 5348
    },
    {
      "epoch": 0.6924944169336829,
      "grad_norm": 2.13323576349333,
      "learning_rate": 3.267943934096525e-07,
      "loss": 2.623,
      "step": 5349
    },
    {
      "epoch": 0.6926238793410363,
      "grad_norm": 1.8045705918564887,
      "learning_rate": 3.265461739655938e-07,
      "loss": 2.4504,
      "step": 5350
    },
    {
      "epoch": 0.6926238793410363,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5150257349014282,
      "eval_runtime": 13.9846,
      "eval_samples_per_second": 3.146,
      "eval_steps_per_second": 0.215,
      "step": 5350
    },
    {
      "epoch": 0.6927533417483898,
      "grad_norm": 1.423514535967298,
      "learning_rate": 3.262980154867786e-07,
      "loss": 2.3643,
      "step": 5351
    },
    {
      "epoch": 0.6928828041557433,
      "grad_norm": 2.8997005798673525,
      "learning_rate": 3.260499180238981e-07,
      "loss": 2.5488,
      "step": 5352
    },
    {
      "epoch": 0.6930122665630968,
      "grad_norm": 1.958156059236238,
      "learning_rate": 3.258018816276311e-07,
      "loss": 2.3618,
      "step": 5353
    },
    {
      "epoch": 0.6931417289704502,
      "grad_norm": 1.2203410520597664,
      "learning_rate": 3.255539063486442e-07,
      "loss": 2.1858,
      "step": 5354
    },
    {
      "epoch": 0.6932711913778037,
      "grad_norm": 1.285710799241308,
      "learning_rate": 3.2530599223759175e-07,
      "loss": 2.4822,
      "step": 5355
    },
    {
      "epoch": 0.6932711913778037,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.513427734375,
      "eval_runtime": 14.9112,
      "eval_samples_per_second": 2.951,
      "eval_steps_per_second": 0.201,
      "step": 5355
    },
    {
      "epoch": 0.6934006537851571,
      "grad_norm": 1.1575368038575253,
      "learning_rate": 3.250581393451148e-07,
      "loss": 2.321,
      "step": 5356
    },
    {
      "epoch": 0.6935301161925106,
      "grad_norm": 2.2841893560977793,
      "learning_rate": 3.2481034772184216e-07,
      "loss": 2.3428,
      "step": 5357
    },
    {
      "epoch": 0.693659578599864,
      "grad_norm": 1.434021653021914,
      "learning_rate": 3.245626174183909e-07,
      "loss": 2.6001,
      "step": 5358
    },
    {
      "epoch": 0.6937890410072175,
      "grad_norm": 4.838179011618462,
      "learning_rate": 3.243149484853643e-07,
      "loss": 2.6089,
      "step": 5359
    },
    {
      "epoch": 0.693918503414571,
      "grad_norm": 1.7801335770591433,
      "learning_rate": 3.240673409733544e-07,
      "loss": 2.5592,
      "step": 5360
    },
    {
      "epoch": 0.693918503414571,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5164240598678589,
      "eval_runtime": 14.9471,
      "eval_samples_per_second": 2.944,
      "eval_steps_per_second": 0.201,
      "step": 5360
    },
    {
      "epoch": 0.6940479658219244,
      "grad_norm": 3.8792111881073925,
      "learning_rate": 3.238197949329395e-07,
      "loss": 2.9326,
      "step": 5361
    },
    {
      "epoch": 0.694177428229278,
      "grad_norm": 1.3674043055950762,
      "learning_rate": 3.235723104146863e-07,
      "loss": 2.3374,
      "step": 5362
    },
    {
      "epoch": 0.6943068906366314,
      "grad_norm": 2.169764738568285,
      "learning_rate": 3.233248874691484e-07,
      "loss": 2.5823,
      "step": 5363
    },
    {
      "epoch": 0.6944363530439849,
      "grad_norm": 2.28597995187258,
      "learning_rate": 3.2307752614686665e-07,
      "loss": 2.7053,
      "step": 5364
    },
    {
      "epoch": 0.6945658154513383,
      "grad_norm": 1.5132922490027412,
      "learning_rate": 3.2283022649837016e-07,
      "loss": 2.624,
      "step": 5365
    },
    {
      "epoch": 0.6945658154513383,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.513671875,
      "eval_runtime": 14.709,
      "eval_samples_per_second": 2.991,
      "eval_steps_per_second": 0.204,
      "step": 5365
    },
    {
      "epoch": 0.6946952778586918,
      "grad_norm": 2.139188625695551,
      "learning_rate": 3.225829885741743e-07,
      "loss": 2.168,
      "step": 5366
    },
    {
      "epoch": 0.6948247402660452,
      "grad_norm": 1.881337685064218,
      "learning_rate": 3.22335812424783e-07,
      "loss": 2.2883,
      "step": 5367
    },
    {
      "epoch": 0.6949542026733987,
      "grad_norm": 2.8039741866215344,
      "learning_rate": 3.2208869810068674e-07,
      "loss": 2.3616,
      "step": 5368
    },
    {
      "epoch": 0.6950836650807521,
      "grad_norm": 1.6342965507594556,
      "learning_rate": 3.218416456523634e-07,
      "loss": 2.3323,
      "step": 5369
    },
    {
      "epoch": 0.6952131274881056,
      "grad_norm": 4.108295650294959,
      "learning_rate": 3.215946551302789e-07,
      "loss": 2.6997,
      "step": 5370
    },
    {
      "epoch": 0.6952131274881056,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5164240598678589,
      "eval_runtime": 15.5815,
      "eval_samples_per_second": 2.824,
      "eval_steps_per_second": 0.193,
      "step": 5370
    },
    {
      "epoch": 0.695342589895459,
      "grad_norm": 1.4029174074335564,
      "learning_rate": 3.2134772658488563e-07,
      "loss": 2.417,
      "step": 5371
    },
    {
      "epoch": 0.6954720523028126,
      "grad_norm": 0.9979938909333347,
      "learning_rate": 3.2110086006662394e-07,
      "loss": 2.1899,
      "step": 5372
    },
    {
      "epoch": 0.6956015147101661,
      "grad_norm": 2.3507384812407146,
      "learning_rate": 3.2085405562592184e-07,
      "loss": 2.502,
      "step": 5373
    },
    {
      "epoch": 0.6957309771175195,
      "grad_norm": 1.9636587874866376,
      "learning_rate": 3.206073133131931e-07,
      "loss": 2.4915,
      "step": 5374
    },
    {
      "epoch": 0.695860439524873,
      "grad_norm": 1.7436191852879661,
      "learning_rate": 3.203606331788406e-07,
      "loss": 2.4773,
      "step": 5375
    },
    {
      "epoch": 0.695860439524873,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5128506422042847,
      "eval_runtime": 14.8286,
      "eval_samples_per_second": 2.967,
      "eval_steps_per_second": 0.202,
      "step": 5375
    },
    {
      "epoch": 0.6959899019322264,
      "grad_norm": 1.375526061002747,
      "learning_rate": 3.2011401527325334e-07,
      "loss": 2.3887,
      "step": 5376
    },
    {
      "epoch": 0.6961193643395799,
      "grad_norm": 1.6880793506449154,
      "learning_rate": 3.198674596468082e-07,
      "loss": 2.135,
      "step": 5377
    },
    {
      "epoch": 0.6962488267469333,
      "grad_norm": 1.9319906198231458,
      "learning_rate": 3.196209663498694e-07,
      "loss": 2.2468,
      "step": 5378
    },
    {
      "epoch": 0.6963782891542868,
      "grad_norm": 1.5176458868229556,
      "learning_rate": 3.1937453543278806e-07,
      "loss": 2.2559,
      "step": 5379
    },
    {
      "epoch": 0.6965077515616402,
      "grad_norm": 1.7327829117470102,
      "learning_rate": 3.1912816694590233e-07,
      "loss": 2.3621,
      "step": 5380
    },
    {
      "epoch": 0.6965077515616402,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5142933130264282,
      "eval_runtime": 14.4793,
      "eval_samples_per_second": 3.039,
      "eval_steps_per_second": 0.207,
      "step": 5380
    },
    {
      "epoch": 0.6966372139689938,
      "grad_norm": 1.7072638716547388,
      "learning_rate": 3.1888186093953845e-07,
      "loss": 2.3591,
      "step": 5381
    },
    {
      "epoch": 0.6967666763763473,
      "grad_norm": 1.648579315286244,
      "learning_rate": 3.186356174640089e-07,
      "loss": 2.4902,
      "step": 5382
    },
    {
      "epoch": 0.6968961387837007,
      "grad_norm": 1.3898330617933954,
      "learning_rate": 3.1838943656961457e-07,
      "loss": 2.438,
      "step": 5383
    },
    {
      "epoch": 0.6970256011910542,
      "grad_norm": 1.8591645372883243,
      "learning_rate": 3.181433183066425e-07,
      "loss": 2.3237,
      "step": 5384
    },
    {
      "epoch": 0.6971550635984076,
      "grad_norm": 3.156839388632553,
      "learning_rate": 3.1789726272536707e-07,
      "loss": 2.2163,
      "step": 5385
    },
    {
      "epoch": 0.6971550635984076,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5115412473678589,
      "eval_runtime": 14.545,
      "eval_samples_per_second": 3.025,
      "eval_steps_per_second": 0.206,
      "step": 5385
    },
    {
      "epoch": 0.6972845260057611,
      "grad_norm": 2.8515370536266373,
      "learning_rate": 3.1765126987605075e-07,
      "loss": 2.3923,
      "step": 5386
    },
    {
      "epoch": 0.6974139884131145,
      "grad_norm": 3.513373915602973,
      "learning_rate": 3.174053398089419e-07,
      "loss": 2.2219,
      "step": 5387
    },
    {
      "epoch": 0.697543450820468,
      "grad_norm": 1.5582575016486924,
      "learning_rate": 3.171594725742773e-07,
      "loss": 2.3403,
      "step": 5388
    },
    {
      "epoch": 0.6976729132278214,
      "grad_norm": 1.4521040399308904,
      "learning_rate": 3.1691366822227994e-07,
      "loss": 2.3269,
      "step": 5389
    },
    {
      "epoch": 0.6978023756351749,
      "grad_norm": 1.37440408682988,
      "learning_rate": 3.1666792680316054e-07,
      "loss": 2.4565,
      "step": 5390
    },
    {
      "epoch": 0.6978023756351749,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5149813890457153,
      "eval_runtime": 15.1876,
      "eval_samples_per_second": 2.897,
      "eval_steps_per_second": 0.198,
      "step": 5390
    },
    {
      "epoch": 0.6979318380425285,
      "grad_norm": 1.286224373879372,
      "learning_rate": 3.1642224836711686e-07,
      "loss": 2.3972,
      "step": 5391
    },
    {
      "epoch": 0.6980613004498819,
      "grad_norm": 2.1602040439474774,
      "learning_rate": 3.161766329643332e-07,
      "loss": 2.2573,
      "step": 5392
    },
    {
      "epoch": 0.6981907628572354,
      "grad_norm": 1.9647660843750425,
      "learning_rate": 3.159310806449821e-07,
      "loss": 2.2974,
      "step": 5393
    },
    {
      "epoch": 0.6983202252645888,
      "grad_norm": 1.5469109018669407,
      "learning_rate": 3.15685591459222e-07,
      "loss": 2.2053,
      "step": 5394
    },
    {
      "epoch": 0.6984496876719423,
      "grad_norm": 4.860701980807949,
      "learning_rate": 3.154401654571998e-07,
      "loss": 2.7876,
      "step": 5395
    },
    {
      "epoch": 0.6984496876719423,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.51416015625,
      "eval_runtime": 13.9488,
      "eval_samples_per_second": 3.154,
      "eval_steps_per_second": 0.215,
      "step": 5395
    },
    {
      "epoch": 0.6985791500792957,
      "grad_norm": 1.827738356764866,
      "learning_rate": 3.1519480268904813e-07,
      "loss": 2.4397,
      "step": 5396
    },
    {
      "epoch": 0.6987086124866492,
      "grad_norm": 1.866702650859283,
      "learning_rate": 3.149495032048873e-07,
      "loss": 2.3193,
      "step": 5397
    },
    {
      "epoch": 0.6988380748940026,
      "grad_norm": 1.8753172545572196,
      "learning_rate": 3.1470426705482505e-07,
      "loss": 2.2368,
      "step": 5398
    },
    {
      "epoch": 0.6989675373013561,
      "grad_norm": 1.6360525619899013,
      "learning_rate": 3.1445909428895546e-07,
      "loss": 2.2825,
      "step": 5399
    },
    {
      "epoch": 0.6990969997087095,
      "grad_norm": 1.4791927266508855,
      "learning_rate": 3.1421398495736026e-07,
      "loss": 2.4946,
      "step": 5400
    },
    {
      "epoch": 0.6990969997087095,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5142489671707153,
      "eval_runtime": 14.4992,
      "eval_samples_per_second": 3.035,
      "eval_steps_per_second": 0.207,
      "step": 5400
    },
    {
      "epoch": 0.6992264621160631,
      "grad_norm": 1.186635029016169,
      "learning_rate": 3.139689391101085e-07,
      "loss": 2.2834,
      "step": 5401
    },
    {
      "epoch": 0.6993559245234166,
      "grad_norm": 1.8995638273612991,
      "learning_rate": 3.1372395679725466e-07,
      "loss": 2.4224,
      "step": 5402
    },
    {
      "epoch": 0.69948538693077,
      "grad_norm": 1.6429652709146856,
      "learning_rate": 3.13479038068842e-07,
      "loss": 2.5115,
      "step": 5403
    },
    {
      "epoch": 0.6996148493381235,
      "grad_norm": 1.9328421132076632,
      "learning_rate": 3.132341829749002e-07,
      "loss": 2.3618,
      "step": 5404
    },
    {
      "epoch": 0.6997443117454769,
      "grad_norm": 1.6013927953761928,
      "learning_rate": 3.1298939156544554e-07,
      "loss": 2.2534,
      "step": 5405
    },
    {
      "epoch": 0.6997443117454769,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5170010328292847,
      "eval_runtime": 14.4779,
      "eval_samples_per_second": 3.039,
      "eval_steps_per_second": 0.207,
      "step": 5405
    },
    {
      "epoch": 0.6998737741528304,
      "grad_norm": 1.734191857629449,
      "learning_rate": 3.1274466389048214e-07,
      "loss": 2.5435,
      "step": 5406
    },
    {
      "epoch": 0.7000032365601838,
      "grad_norm": 1.8077362625993985,
      "learning_rate": 3.125000000000002e-07,
      "loss": 2.4829,
      "step": 5407
    },
    {
      "epoch": 0.7001326989675373,
      "grad_norm": 1.1982756877128877,
      "learning_rate": 3.1225539994397704e-07,
      "loss": 2.0586,
      "step": 5408
    },
    {
      "epoch": 0.7002621613748907,
      "grad_norm": 1.4034583092961634,
      "learning_rate": 3.120108637723778e-07,
      "loss": 2.2112,
      "step": 5409
    },
    {
      "epoch": 0.7003916237822442,
      "grad_norm": 1.8325979482049013,
      "learning_rate": 3.1176639153515343e-07,
      "loss": 2.4893,
      "step": 5410
    },
    {
      "epoch": 0.7003916237822442,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5114967823028564,
      "eval_runtime": 14.0872,
      "eval_samples_per_second": 3.123,
      "eval_steps_per_second": 0.213,
      "step": 5410
    },
    {
      "epoch": 0.7005210861895977,
      "grad_norm": 1.031947826763034,
      "learning_rate": 3.115219832822427e-07,
      "loss": 2.335,
      "step": 5411
    },
    {
      "epoch": 0.7006505485969512,
      "grad_norm": 1.2032347547295184,
      "learning_rate": 3.1127763906357086e-07,
      "loss": 2.3384,
      "step": 5412
    },
    {
      "epoch": 0.7007800110043046,
      "grad_norm": 1.40643082208662,
      "learning_rate": 3.1103335892904983e-07,
      "loss": 2.3037,
      "step": 5413
    },
    {
      "epoch": 0.7009094734116581,
      "grad_norm": 2.602429291181452,
      "learning_rate": 3.1078914292857925e-07,
      "loss": 2.3455,
      "step": 5414
    },
    {
      "epoch": 0.7010389358190116,
      "grad_norm": 3.1560069744271364,
      "learning_rate": 3.1054499111204486e-07,
      "loss": 3.1663,
      "step": 5415
    },
    {
      "epoch": 0.7010389358190116,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5157581567764282,
      "eval_runtime": 14.8477,
      "eval_samples_per_second": 2.963,
      "eval_steps_per_second": 0.202,
      "step": 5415
    },
    {
      "epoch": 0.701168398226365,
      "grad_norm": 2.1503959903003667,
      "learning_rate": 3.1030090352931995e-07,
      "loss": 2.3535,
      "step": 5416
    },
    {
      "epoch": 0.7012978606337185,
      "grad_norm": 1.1744324577985705,
      "learning_rate": 3.1005688023026387e-07,
      "loss": 2.2505,
      "step": 5417
    },
    {
      "epoch": 0.7014273230410719,
      "grad_norm": 2.1089863170437857,
      "learning_rate": 3.098129212647239e-07,
      "loss": 2.5706,
      "step": 5418
    },
    {
      "epoch": 0.7015567854484254,
      "grad_norm": 1.419840139890607,
      "learning_rate": 3.0956902668253327e-07,
      "loss": 2.27,
      "step": 5419
    },
    {
      "epoch": 0.7016862478557789,
      "grad_norm": 2.0435726528843596,
      "learning_rate": 3.093251965335122e-07,
      "loss": 2.6697,
      "step": 5420
    },
    {
      "epoch": 0.7016862478557789,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.515625,
      "eval_runtime": 14.0954,
      "eval_samples_per_second": 3.122,
      "eval_steps_per_second": 0.213,
      "step": 5420
    },
    {
      "epoch": 0.7018157102631324,
      "grad_norm": 1.4000930091500776,
      "learning_rate": 3.090814308674682e-07,
      "loss": 2.4956,
      "step": 5421
    },
    {
      "epoch": 0.7019451726704858,
      "grad_norm": 1.4296183949185863,
      "learning_rate": 3.0883772973419553e-07,
      "loss": 2.2876,
      "step": 5422
    },
    {
      "epoch": 0.7020746350778393,
      "grad_norm": 1.7022564155071238,
      "learning_rate": 3.085940931834749e-07,
      "loss": 2.4507,
      "step": 5423
    },
    {
      "epoch": 0.7022040974851927,
      "grad_norm": 1.609684352223778,
      "learning_rate": 3.08350521265074e-07,
      "loss": 2.4968,
      "step": 5424
    },
    {
      "epoch": 0.7023335598925462,
      "grad_norm": 1.6220631821775575,
      "learning_rate": 3.081070140287471e-07,
      "loss": 2.5127,
      "step": 5425
    },
    {
      "epoch": 0.7023335598925462,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5156471729278564,
      "eval_runtime": 13.9005,
      "eval_samples_per_second": 3.165,
      "eval_steps_per_second": 0.216,
      "step": 5425
    },
    {
      "epoch": 0.7024630222998997,
      "grad_norm": 1.33830444157398,
      "learning_rate": 3.078635715242357e-07,
      "loss": 2.29,
      "step": 5426
    },
    {
      "epoch": 0.7025924847072531,
      "grad_norm": 1.674054403047687,
      "learning_rate": 3.076201938012681e-07,
      "loss": 2.406,
      "step": 5427
    },
    {
      "epoch": 0.7027219471146066,
      "grad_norm": 1.7850975837638183,
      "learning_rate": 3.073768809095587e-07,
      "loss": 2.2429,
      "step": 5428
    },
    {
      "epoch": 0.70285140952196,
      "grad_norm": 2.0379084555803524,
      "learning_rate": 3.071336328988095e-07,
      "loss": 2.519,
      "step": 5429
    },
    {
      "epoch": 0.7029808719293136,
      "grad_norm": 2.0983829443582382,
      "learning_rate": 3.0689044981870865e-07,
      "loss": 2.4026,
      "step": 5430
    },
    {
      "epoch": 0.7029808719293136,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5149813890457153,
      "eval_runtime": 14.4294,
      "eval_samples_per_second": 3.049,
      "eval_steps_per_second": 0.208,
      "step": 5430
    },
    {
      "epoch": 0.703110334336667,
      "grad_norm": 1.4576542430247819,
      "learning_rate": 3.06647331718931e-07,
      "loss": 2.5098,
      "step": 5431
    },
    {
      "epoch": 0.7032397967440205,
      "grad_norm": 1.4999026407300369,
      "learning_rate": 3.064042786491388e-07,
      "loss": 2.3525,
      "step": 5432
    },
    {
      "epoch": 0.7033692591513739,
      "grad_norm": 2.3911531733057116,
      "learning_rate": 3.0616129065898004e-07,
      "loss": 2.447,
      "step": 5433
    },
    {
      "epoch": 0.7034987215587274,
      "grad_norm": 1.691530032429575,
      "learning_rate": 3.059183677980905e-07,
      "loss": 2.3381,
      "step": 5434
    },
    {
      "epoch": 0.7036281839660808,
      "grad_norm": 2.1022824376195604,
      "learning_rate": 3.056755101160918e-07,
      "loss": 2.4399,
      "step": 5435
    },
    {
      "epoch": 0.7036281839660808,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.510786533355713,
      "eval_runtime": 15.5736,
      "eval_samples_per_second": 2.825,
      "eval_steps_per_second": 0.193,
      "step": 5435
    },
    {
      "epoch": 0.7037576463734343,
      "grad_norm": 2.0826754739211832,
      "learning_rate": 3.054327176625924e-07,
      "loss": 2.3694,
      "step": 5436
    },
    {
      "epoch": 0.7038871087807878,
      "grad_norm": 1.552936864115532,
      "learning_rate": 3.0518999048718795e-07,
      "loss": 2.2969,
      "step": 5437
    },
    {
      "epoch": 0.7040165711881412,
      "grad_norm": 2.3615324849454415,
      "learning_rate": 3.049473286394599e-07,
      "loss": 2.2988,
      "step": 5438
    },
    {
      "epoch": 0.7041460335954947,
      "grad_norm": 1.9947144084490287,
      "learning_rate": 3.047047321689775e-07,
      "loss": 2.4597,
      "step": 5439
    },
    {
      "epoch": 0.7042754960028482,
      "grad_norm": 2.484209298485715,
      "learning_rate": 3.0446220112529555e-07,
      "loss": 2.3804,
      "step": 5440
    },
    {
      "epoch": 0.7042754960028482,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.515625,
      "eval_runtime": 14.9414,
      "eval_samples_per_second": 2.945,
      "eval_steps_per_second": 0.201,
      "step": 5440
    },
    {
      "epoch": 0.7044049584102017,
      "grad_norm": 1.5148496222853771,
      "learning_rate": 3.0421973555795585e-07,
      "loss": 2.2559,
      "step": 5441
    },
    {
      "epoch": 0.7045344208175551,
      "grad_norm": 2.050627791832364,
      "learning_rate": 3.0397733551648725e-07,
      "loss": 2.2622,
      "step": 5442
    },
    {
      "epoch": 0.7046638832249086,
      "grad_norm": 1.556620680990198,
      "learning_rate": 3.0373500105040456e-07,
      "loss": 2.5139,
      "step": 5443
    },
    {
      "epoch": 0.704793345632262,
      "grad_norm": 2.105901211082287,
      "learning_rate": 3.0349273220920955e-07,
      "loss": 2.8171,
      "step": 5444
    },
    {
      "epoch": 0.7049228080396155,
      "grad_norm": 1.91482161355928,
      "learning_rate": 3.0325052904239096e-07,
      "loss": 2.4438,
      "step": 5445
    },
    {
      "epoch": 0.7049228080396155,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5136052370071411,
      "eval_runtime": 14.4262,
      "eval_samples_per_second": 3.05,
      "eval_steps_per_second": 0.208,
      "step": 5445
    },
    {
      "epoch": 0.7050522704469689,
      "grad_norm": 2.0759088329303936,
      "learning_rate": 3.0300839159942335e-07,
      "loss": 2.3867,
      "step": 5446
    },
    {
      "epoch": 0.7051817328543224,
      "grad_norm": 2.5915277587769,
      "learning_rate": 3.027663199297682e-07,
      "loss": 2.2825,
      "step": 5447
    },
    {
      "epoch": 0.7053111952616758,
      "grad_norm": 1.4249593657417663,
      "learning_rate": 3.025243140828734e-07,
      "loss": 2.4436,
      "step": 5448
    },
    {
      "epoch": 0.7054406576690293,
      "grad_norm": 1.8149636589990787,
      "learning_rate": 3.022823741081737e-07,
      "loss": 2.6514,
      "step": 5449
    },
    {
      "epoch": 0.7055701200763829,
      "grad_norm": 2.6025866686770653,
      "learning_rate": 3.020405000550905e-07,
      "loss": 2.3667,
      "step": 5450
    },
    {
      "epoch": 0.7055701200763829,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5157581567764282,
      "eval_runtime": 14.2301,
      "eval_samples_per_second": 3.092,
      "eval_steps_per_second": 0.211,
      "step": 5450
    },
    {
      "epoch": 0.7056995824837363,
      "grad_norm": 1.574276657392704,
      "learning_rate": 3.0179869197303125e-07,
      "loss": 2.4951,
      "step": 5451
    },
    {
      "epoch": 0.7058290448910898,
      "grad_norm": 1.8247724014621267,
      "learning_rate": 3.0155694991139e-07,
      "loss": 2.2301,
      "step": 5452
    },
    {
      "epoch": 0.7059585072984432,
      "grad_norm": 1.2739726890028502,
      "learning_rate": 3.0131527391954776e-07,
      "loss": 2.4631,
      "step": 5453
    },
    {
      "epoch": 0.7060879697057967,
      "grad_norm": 1.4741302234201548,
      "learning_rate": 3.0107366404687145e-07,
      "loss": 2.4326,
      "step": 5454
    },
    {
      "epoch": 0.7062174321131501,
      "grad_norm": 1.7257062152662175,
      "learning_rate": 3.008321203427151e-07,
      "loss": 2.5139,
      "step": 5455
    },
    {
      "epoch": 0.7062174321131501,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.512895107269287,
      "eval_runtime": 14.3123,
      "eval_samples_per_second": 3.074,
      "eval_steps_per_second": 0.21,
      "step": 5455
    },
    {
      "epoch": 0.7063468945205036,
      "grad_norm": 1.4845271698246092,
      "learning_rate": 3.0059064285641857e-07,
      "loss": 2.3633,
      "step": 5456
    },
    {
      "epoch": 0.706476356927857,
      "grad_norm": 1.8092886527834913,
      "learning_rate": 3.00349231637309e-07,
      "loss": 2.4194,
      "step": 5457
    },
    {
      "epoch": 0.7066058193352105,
      "grad_norm": 1.7936588912183051,
      "learning_rate": 3.001078867346992e-07,
      "loss": 2.5867,
      "step": 5458
    },
    {
      "epoch": 0.706735281742564,
      "grad_norm": 2.0701410817647896,
      "learning_rate": 2.9986660819788847e-07,
      "loss": 2.2192,
      "step": 5459
    },
    {
      "epoch": 0.7068647441499175,
      "grad_norm": 1.3732517347453355,
      "learning_rate": 2.9962539607616353e-07,
      "loss": 2.332,
      "step": 5460
    },
    {
      "epoch": 0.7068647441499175,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5136052370071411,
      "eval_runtime": 13.7143,
      "eval_samples_per_second": 3.208,
      "eval_steps_per_second": 0.219,
      "step": 5460
    },
    {
      "epoch": 0.706994206557271,
      "grad_norm": 1.3969195107471228,
      "learning_rate": 2.9938425041879615e-07,
      "loss": 2.2974,
      "step": 5461
    },
    {
      "epoch": 0.7071236689646244,
      "grad_norm": 4.364035985519196,
      "learning_rate": 2.9914317127504587e-07,
      "loss": 2.5571,
      "step": 5462
    },
    {
      "epoch": 0.7072531313719779,
      "grad_norm": 1.6150948091088442,
      "learning_rate": 2.9890215869415764e-07,
      "loss": 2.2448,
      "step": 5463
    },
    {
      "epoch": 0.7073825937793313,
      "grad_norm": 1.5417784519487534,
      "learning_rate": 2.9866121272536297e-07,
      "loss": 2.5642,
      "step": 5464
    },
    {
      "epoch": 0.7075120561866848,
      "grad_norm": 1.266865102094396,
      "learning_rate": 2.984203334178804e-07,
      "loss": 2.4429,
      "step": 5465
    },
    {
      "epoch": 0.7075120561866848,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5128728151321411,
      "eval_runtime": 12.8483,
      "eval_samples_per_second": 3.425,
      "eval_steps_per_second": 0.233,
      "step": 5465
    },
    {
      "epoch": 0.7076415185940382,
      "grad_norm": 1.7082974928030255,
      "learning_rate": 2.98179520820914e-07,
      "loss": 2.5574,
      "step": 5466
    },
    {
      "epoch": 0.7077709810013917,
      "grad_norm": 1.4952069787003692,
      "learning_rate": 2.979387749836548e-07,
      "loss": 2.292,
      "step": 5467
    },
    {
      "epoch": 0.7079004434087451,
      "grad_norm": 1.6882378452340079,
      "learning_rate": 2.976980959552805e-07,
      "loss": 2.3267,
      "step": 5468
    },
    {
      "epoch": 0.7080299058160987,
      "grad_norm": 1.4705000244831505,
      "learning_rate": 2.974574837849538e-07,
      "loss": 2.6318,
      "step": 5469
    },
    {
      "epoch": 0.7081593682234522,
      "grad_norm": 2.4697891945795516,
      "learning_rate": 2.972169385218252e-07,
      "loss": 2.6724,
      "step": 5470
    },
    {
      "epoch": 0.7081593682234522,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5128062963485718,
      "eval_runtime": 14.6489,
      "eval_samples_per_second": 3.004,
      "eval_steps_per_second": 0.205,
      "step": 5470
    },
    {
      "epoch": 0.7082888306308056,
      "grad_norm": 2.5365204004259683,
      "learning_rate": 2.969764602150305e-07,
      "loss": 2.3005,
      "step": 5471
    },
    {
      "epoch": 0.7084182930381591,
      "grad_norm": 2.2054646464841112,
      "learning_rate": 2.967360489136926e-07,
      "loss": 2.4115,
      "step": 5472
    },
    {
      "epoch": 0.7085477554455125,
      "grad_norm": 1.2705979197329882,
      "learning_rate": 2.964957046669208e-07,
      "loss": 2.3721,
      "step": 5473
    },
    {
      "epoch": 0.708677217852866,
      "grad_norm": 2.7935239563139307,
      "learning_rate": 2.9625542752380927e-07,
      "loss": 2.4829,
      "step": 5474
    },
    {
      "epoch": 0.7088066802602194,
      "grad_norm": 2.333047871088322,
      "learning_rate": 2.9601521753344e-07,
      "loss": 2.4336,
      "step": 5475
    },
    {
      "epoch": 0.7088066802602194,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5121182203292847,
      "eval_runtime": 14.6633,
      "eval_samples_per_second": 3.001,
      "eval_steps_per_second": 0.205,
      "step": 5475
    },
    {
      "epoch": 0.7089361426675729,
      "grad_norm": 1.9329422209847855,
      "learning_rate": 2.957750747448811e-07,
      "loss": 2.2339,
      "step": 5476
    },
    {
      "epoch": 0.7090656050749263,
      "grad_norm": 2.3940518724987983,
      "learning_rate": 2.9553499920718593e-07,
      "loss": 2.2429,
      "step": 5477
    },
    {
      "epoch": 0.7091950674822798,
      "grad_norm": 1.9642118631365526,
      "learning_rate": 2.952949909693954e-07,
      "loss": 2.7961,
      "step": 5478
    },
    {
      "epoch": 0.7093245298896333,
      "grad_norm": 1.062591419130222,
      "learning_rate": 2.9505505008053583e-07,
      "loss": 2.1123,
      "step": 5479
    },
    {
      "epoch": 0.7094539922969868,
      "grad_norm": 2.5608689953969916,
      "learning_rate": 2.9481517658961965e-07,
      "loss": 2.1968,
      "step": 5480
    },
    {
      "epoch": 0.7094539922969868,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5107200145721436,
      "eval_runtime": 14.1735,
      "eval_samples_per_second": 3.104,
      "eval_steps_per_second": 0.212,
      "step": 5480
    },
    {
      "epoch": 0.7095834547043403,
      "grad_norm": 1.337703409697096,
      "learning_rate": 2.9457537054564647e-07,
      "loss": 2.4028,
      "step": 5481
    },
    {
      "epoch": 0.7097129171116937,
      "grad_norm": 1.9010823142558655,
      "learning_rate": 2.9433563199760095e-07,
      "loss": 2.3994,
      "step": 5482
    },
    {
      "epoch": 0.7098423795190472,
      "grad_norm": 1.7257209889372729,
      "learning_rate": 2.940959609944552e-07,
      "loss": 2.3069,
      "step": 5483
    },
    {
      "epoch": 0.7099718419264006,
      "grad_norm": 2.0761339000769263,
      "learning_rate": 2.938563575851665e-07,
      "loss": 2.2153,
      "step": 5484
    },
    {
      "epoch": 0.7101013043337541,
      "grad_norm": 3.026756961222471,
      "learning_rate": 2.936168218186785e-07,
      "loss": 2.6416,
      "step": 5485
    },
    {
      "epoch": 0.7101013043337541,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5107200145721436,
      "eval_runtime": 13.9592,
      "eval_samples_per_second": 3.152,
      "eval_steps_per_second": 0.215,
      "step": 5485
    },
    {
      "epoch": 0.7102307667411075,
      "grad_norm": 1.4180293171233147,
      "learning_rate": 2.933773537439217e-07,
      "loss": 2.4192,
      "step": 5486
    },
    {
      "epoch": 0.710360229148461,
      "grad_norm": 1.3914393537554786,
      "learning_rate": 2.931379534098118e-07,
      "loss": 2.303,
      "step": 5487
    },
    {
      "epoch": 0.7104896915558144,
      "grad_norm": 1.7370934374122329,
      "learning_rate": 2.928986208652518e-07,
      "loss": 2.3853,
      "step": 5488
    },
    {
      "epoch": 0.710619153963168,
      "grad_norm": 1.585344876482885,
      "learning_rate": 2.9265935615912956e-07,
      "loss": 2.3574,
      "step": 5489
    },
    {
      "epoch": 0.7107486163705214,
      "grad_norm": 1.4600865931055707,
      "learning_rate": 2.9242015934032033e-07,
      "loss": 2.3567,
      "step": 5490
    },
    {
      "epoch": 0.7107486163705214,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.511518955230713,
      "eval_runtime": 14.7041,
      "eval_samples_per_second": 2.992,
      "eval_steps_per_second": 0.204,
      "step": 5490
    },
    {
      "epoch": 0.7108780787778749,
      "grad_norm": 1.5533301614533888,
      "learning_rate": 2.9218103045768456e-07,
      "loss": 2.21,
      "step": 5491
    },
    {
      "epoch": 0.7110075411852284,
      "grad_norm": 1.8222863543886278,
      "learning_rate": 2.9194196956006914e-07,
      "loss": 2.3228,
      "step": 5492
    },
    {
      "epoch": 0.7111370035925818,
      "grad_norm": 1.5259869863465987,
      "learning_rate": 2.9170297669630735e-07,
      "loss": 2.2871,
      "step": 5493
    },
    {
      "epoch": 0.7112664659999353,
      "grad_norm": 1.9441677128782537,
      "learning_rate": 2.91464051915218e-07,
      "loss": 2.3655,
      "step": 5494
    },
    {
      "epoch": 0.7113959284072887,
      "grad_norm": 2.031245198793413,
      "learning_rate": 2.912251952656065e-07,
      "loss": 2.2627,
      "step": 5495
    },
    {
      "epoch": 0.7113959284072887,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5135387182235718,
      "eval_runtime": 14.7326,
      "eval_samples_per_second": 2.987,
      "eval_steps_per_second": 0.204,
      "step": 5495
    },
    {
      "epoch": 0.7115253908146422,
      "grad_norm": 1.6557814810252556,
      "learning_rate": 2.9098640679626473e-07,
      "loss": 2.6448,
      "step": 5496
    },
    {
      "epoch": 0.7116548532219956,
      "grad_norm": 2.3030706547413096,
      "learning_rate": 2.907476865559689e-07,
      "loss": 2.6919,
      "step": 5497
    },
    {
      "epoch": 0.7117843156293491,
      "grad_norm": 1.454468789961194,
      "learning_rate": 2.9050903459348323e-07,
      "loss": 2.3105,
      "step": 5498
    },
    {
      "epoch": 0.7119137780367026,
      "grad_norm": 1.180216293314466,
      "learning_rate": 2.902704509575572e-07,
      "loss": 2.2432,
      "step": 5499
    },
    {
      "epoch": 0.7120432404440561,
      "grad_norm": 0.9700997666670762,
      "learning_rate": 2.90031935696926e-07,
      "loss": 2.1406,
      "step": 5500
    },
    {
      "epoch": 0.7120432404440561,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.510009765625,
      "eval_runtime": 13.9422,
      "eval_samples_per_second": 3.156,
      "eval_steps_per_second": 0.215,
      "step": 5500
    },
    {
      "epoch": 0.7121727028514095,
      "grad_norm": 1.6587128474393105,
      "learning_rate": 2.897934888603117e-07,
      "loss": 2.4314,
      "step": 5501
    },
    {
      "epoch": 0.712302165258763,
      "grad_norm": 1.4162562684840903,
      "learning_rate": 2.8955511049642156e-07,
      "loss": 2.166,
      "step": 5502
    },
    {
      "epoch": 0.7124316276661165,
      "grad_norm": 1.6084256458162793,
      "learning_rate": 2.8931680065394907e-07,
      "loss": 2.478,
      "step": 5503
    },
    {
      "epoch": 0.7125610900734699,
      "grad_norm": 2.464881038647762,
      "learning_rate": 2.8907855938157427e-07,
      "loss": 2.3914,
      "step": 5504
    },
    {
      "epoch": 0.7126905524808234,
      "grad_norm": 1.2356605888206103,
      "learning_rate": 2.888403867279623e-07,
      "loss": 2.3491,
      "step": 5505
    },
    {
      "epoch": 0.7126905524808234,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.511474609375,
      "eval_runtime": 13.4246,
      "eval_samples_per_second": 3.278,
      "eval_steps_per_second": 0.223,
      "step": 5505
    },
    {
      "epoch": 0.7128200148881768,
      "grad_norm": 1.608224140789146,
      "learning_rate": 2.8860228274176517e-07,
      "loss": 2.457,
      "step": 5506
    },
    {
      "epoch": 0.7129494772955303,
      "grad_norm": 1.2687054741524142,
      "learning_rate": 2.8836424747162034e-07,
      "loss": 2.1235,
      "step": 5507
    },
    {
      "epoch": 0.7130789397028838,
      "grad_norm": 2.2563542475474856,
      "learning_rate": 2.88126280966151e-07,
      "loss": 2.4487,
      "step": 5508
    },
    {
      "epoch": 0.7132084021102373,
      "grad_norm": 2.97185135449286,
      "learning_rate": 2.878883832739671e-07,
      "loss": 2.3892,
      "step": 5509
    },
    {
      "epoch": 0.7133378645175907,
      "grad_norm": 0.945093829392246,
      "learning_rate": 2.8765055444366364e-07,
      "loss": 2.1051,
      "step": 5510
    },
    {
      "epoch": 0.7133378645175907,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.510054111480713,
      "eval_runtime": 13.7777,
      "eval_samples_per_second": 3.194,
      "eval_steps_per_second": 0.218,
      "step": 5510
    },
    {
      "epoch": 0.7134673269249442,
      "grad_norm": 1.5028758044545727,
      "learning_rate": 2.8741279452382246e-07,
      "loss": 2.1548,
      "step": 5511
    },
    {
      "epoch": 0.7135967893322976,
      "grad_norm": 3.9063964122936476,
      "learning_rate": 2.8717510356301055e-07,
      "loss": 2.9185,
      "step": 5512
    },
    {
      "epoch": 0.7137262517396511,
      "grad_norm": 2.330277696128063,
      "learning_rate": 2.8693748160978095e-07,
      "loss": 2.3647,
      "step": 5513
    },
    {
      "epoch": 0.7138557141470045,
      "grad_norm": 2.0524267159751397,
      "learning_rate": 2.866999287126731e-07,
      "loss": 2.417,
      "step": 5514
    },
    {
      "epoch": 0.713985176554358,
      "grad_norm": 1.64944727637688,
      "learning_rate": 2.8646244492021163e-07,
      "loss": 2.3623,
      "step": 5515
    },
    {
      "epoch": 0.713985176554358,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.507768154144287,
      "eval_runtime": 13.7639,
      "eval_samples_per_second": 3.197,
      "eval_steps_per_second": 0.218,
      "step": 5515
    },
    {
      "epoch": 0.7141146389617115,
      "grad_norm": 4.338832030132758,
      "learning_rate": 2.8622503028090764e-07,
      "loss": 2.4463,
      "step": 5516
    },
    {
      "epoch": 0.7142441013690649,
      "grad_norm": 1.7430278071155016,
      "learning_rate": 2.859876848432581e-07,
      "loss": 2.4495,
      "step": 5517
    },
    {
      "epoch": 0.7143735637764185,
      "grad_norm": 1.907112513554861,
      "learning_rate": 2.857504086557454e-07,
      "loss": 2.1299,
      "step": 5518
    },
    {
      "epoch": 0.7145030261837719,
      "grad_norm": 2.708399660442652,
      "learning_rate": 2.85513201766838e-07,
      "loss": 2.5608,
      "step": 5519
    },
    {
      "epoch": 0.7146324885911254,
      "grad_norm": 1.296584160472222,
      "learning_rate": 2.8527606422499e-07,
      "loss": 2.4521,
      "step": 5520
    },
    {
      "epoch": 0.7146324885911254,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5079013109207153,
      "eval_runtime": 13.6967,
      "eval_samples_per_second": 3.212,
      "eval_steps_per_second": 0.219,
      "step": 5520
    },
    {
      "epoch": 0.7147619509984788,
      "grad_norm": 1.3739841531314605,
      "learning_rate": 2.850389960786418e-07,
      "loss": 2.3652,
      "step": 5521
    },
    {
      "epoch": 0.7148914134058323,
      "grad_norm": 1.623312421708892,
      "learning_rate": 2.848019973762195e-07,
      "loss": 2.4771,
      "step": 5522
    },
    {
      "epoch": 0.7150208758131857,
      "grad_norm": 1.958130293020498,
      "learning_rate": 2.845650681661348e-07,
      "loss": 2.5439,
      "step": 5523
    },
    {
      "epoch": 0.7151503382205392,
      "grad_norm": 1.6339741316942522,
      "learning_rate": 2.84328208496785e-07,
      "loss": 2.3599,
      "step": 5524
    },
    {
      "epoch": 0.7152798006278926,
      "grad_norm": 1.1990610677201585,
      "learning_rate": 2.840914184165539e-07,
      "loss": 2.4148,
      "step": 5525
    },
    {
      "epoch": 0.7152798006278926,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.512162685394287,
      "eval_runtime": 13.4574,
      "eval_samples_per_second": 3.27,
      "eval_steps_per_second": 0.223,
      "step": 5525
    },
    {
      "epoch": 0.7154092630352461,
      "grad_norm": 1.7739544361348252,
      "learning_rate": 2.8385469797381026e-07,
      "loss": 2.415,
      "step": 5526
    },
    {
      "epoch": 0.7155387254425996,
      "grad_norm": 3.45287956507049,
      "learning_rate": 2.836180472169096e-07,
      "loss": 2.6968,
      "step": 5527
    },
    {
      "epoch": 0.7156681878499531,
      "grad_norm": 1.2147623211617138,
      "learning_rate": 2.8338146619419195e-07,
      "loss": 2.1594,
      "step": 5528
    },
    {
      "epoch": 0.7157976502573066,
      "grad_norm": 1.3240594729424797,
      "learning_rate": 2.831449549539844e-07,
      "loss": 2.3896,
      "step": 5529
    },
    {
      "epoch": 0.71592711266466,
      "grad_norm": 1.476022294129693,
      "learning_rate": 2.8290851354459887e-07,
      "loss": 2.248,
      "step": 5530
    },
    {
      "epoch": 0.71592711266466,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5107421875,
      "eval_runtime": 14.0462,
      "eval_samples_per_second": 3.133,
      "eval_steps_per_second": 0.214,
      "step": 5530
    },
    {
      "epoch": 0.7160565750720135,
      "grad_norm": 1.824707853093141,
      "learning_rate": 2.826721420143331e-07,
      "loss": 2.4666,
      "step": 5531
    },
    {
      "epoch": 0.7161860374793669,
      "grad_norm": 1.4682217075595523,
      "learning_rate": 2.8243584041147117e-07,
      "loss": 2.3921,
      "step": 5532
    },
    {
      "epoch": 0.7163154998867204,
      "grad_norm": 2.168559203650329,
      "learning_rate": 2.821996087842822e-07,
      "loss": 2.5039,
      "step": 5533
    },
    {
      "epoch": 0.7164449622940738,
      "grad_norm": 1.5776966453526566,
      "learning_rate": 2.8196344718102145e-07,
      "loss": 2.4316,
      "step": 5534
    },
    {
      "epoch": 0.7165744247014273,
      "grad_norm": 1.5985645007345604,
      "learning_rate": 2.8172735564992977e-07,
      "loss": 2.2637,
      "step": 5535
    },
    {
      "epoch": 0.7165744247014273,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5105868577957153,
      "eval_runtime": 14.3985,
      "eval_samples_per_second": 3.056,
      "eval_steps_per_second": 0.208,
      "step": 5535
    },
    {
      "epoch": 0.7167038871087807,
      "grad_norm": 2.963478626299164,
      "learning_rate": 2.814913342392332e-07,
      "loss": 2.6663,
      "step": 5536
    },
    {
      "epoch": 0.7168333495161342,
      "grad_norm": 1.0946133801481273,
      "learning_rate": 2.8125538299714455e-07,
      "loss": 2.3401,
      "step": 5537
    },
    {
      "epoch": 0.7169628119234878,
      "grad_norm": 1.8524375943717375,
      "learning_rate": 2.81019501971861e-07,
      "loss": 2.2546,
      "step": 5538
    },
    {
      "epoch": 0.7170922743308412,
      "grad_norm": 1.7007590761978473,
      "learning_rate": 2.8078369121156633e-07,
      "loss": 2.1255,
      "step": 5539
    },
    {
      "epoch": 0.7172217367381947,
      "grad_norm": 1.7471511623985778,
      "learning_rate": 2.805479507644302e-07,
      "loss": 2.2898,
      "step": 5540
    },
    {
      "epoch": 0.7172217367381947,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.508589267730713,
      "eval_runtime": 14.602,
      "eval_samples_per_second": 3.013,
      "eval_steps_per_second": 0.205,
      "step": 5540
    },
    {
      "epoch": 0.7173511991455481,
      "grad_norm": 2.65354892426599,
      "learning_rate": 2.8031228067860633e-07,
      "loss": 2.4062,
      "step": 5541
    },
    {
      "epoch": 0.7174806615529016,
      "grad_norm": 1.7348926744322892,
      "learning_rate": 2.800766810022359e-07,
      "loss": 2.3301,
      "step": 5542
    },
    {
      "epoch": 0.717610123960255,
      "grad_norm": 2.4364528588960126,
      "learning_rate": 2.798411517834444e-07,
      "loss": 2.7969,
      "step": 5543
    },
    {
      "epoch": 0.7177395863676085,
      "grad_norm": 2.3385069966979697,
      "learning_rate": 2.7960569307034375e-07,
      "loss": 2.8467,
      "step": 5544
    },
    {
      "epoch": 0.7178690487749619,
      "grad_norm": 1.3253198369031496,
      "learning_rate": 2.7937030491103125e-07,
      "loss": 2.3762,
      "step": 5545
    },
    {
      "epoch": 0.7178690487749619,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5106533765792847,
      "eval_runtime": 14.6764,
      "eval_samples_per_second": 2.998,
      "eval_steps_per_second": 0.204,
      "step": 5545
    },
    {
      "epoch": 0.7179985111823154,
      "grad_norm": 1.864054273156971,
      "learning_rate": 2.7913498735358964e-07,
      "loss": 2.3914,
      "step": 5546
    },
    {
      "epoch": 0.718127973589669,
      "grad_norm": 1.145738931308922,
      "learning_rate": 2.78899740446087e-07,
      "loss": 2.2905,
      "step": 5547
    },
    {
      "epoch": 0.7182574359970224,
      "grad_norm": 2.223617748065961,
      "learning_rate": 2.7866456423657775e-07,
      "loss": 2.4192,
      "step": 5548
    },
    {
      "epoch": 0.7183868984043759,
      "grad_norm": 1.927960560650502,
      "learning_rate": 2.7842945877310087e-07,
      "loss": 2.3726,
      "step": 5549
    },
    {
      "epoch": 0.7185163608117293,
      "grad_norm": 1.280723933023443,
      "learning_rate": 2.7819442410368195e-07,
      "loss": 2.2849,
      "step": 5550
    },
    {
      "epoch": 0.7185163608117293,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5106533765792847,
      "eval_runtime": 13.8223,
      "eval_samples_per_second": 3.183,
      "eval_steps_per_second": 0.217,
      "step": 5550
    },
    {
      "epoch": 0.7186458232190828,
      "grad_norm": 2.69337669376921,
      "learning_rate": 2.7795946027633135e-07,
      "loss": 2.6292,
      "step": 5551
    },
    {
      "epoch": 0.7187752856264362,
      "grad_norm": 1.1707863558550546,
      "learning_rate": 2.7772456733904487e-07,
      "loss": 2.3149,
      "step": 5552
    },
    {
      "epoch": 0.7189047480337897,
      "grad_norm": 2.3302889611217963,
      "learning_rate": 2.7748974533980485e-07,
      "loss": 2.3171,
      "step": 5553
    },
    {
      "epoch": 0.7190342104411431,
      "grad_norm": 2.858880558289602,
      "learning_rate": 2.7725499432657766e-07,
      "loss": 2.3002,
      "step": 5554
    },
    {
      "epoch": 0.7191636728484966,
      "grad_norm": 1.7256317431863324,
      "learning_rate": 2.770203143473167e-07,
      "loss": 2.3389,
      "step": 5555
    },
    {
      "epoch": 0.7191636728484966,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5114967823028564,
      "eval_runtime": 14.9309,
      "eval_samples_per_second": 2.947,
      "eval_steps_per_second": 0.201,
      "step": 5555
    },
    {
      "epoch": 0.71929313525585,
      "grad_norm": 1.5143480210948022,
      "learning_rate": 2.767857054499595e-07,
      "loss": 2.2592,
      "step": 5556
    },
    {
      "epoch": 0.7194225976632036,
      "grad_norm": 1.295316019296011,
      "learning_rate": 2.765511676824301e-07,
      "loss": 2.2485,
      "step": 5557
    },
    {
      "epoch": 0.719552060070557,
      "grad_norm": 2.2133328807686854,
      "learning_rate": 2.763167010926376e-07,
      "loss": 2.334,
      "step": 5558
    },
    {
      "epoch": 0.7196815224779105,
      "grad_norm": 1.173305100737012,
      "learning_rate": 2.7608230572847604e-07,
      "loss": 2.1846,
      "step": 5559
    },
    {
      "epoch": 0.719810984885264,
      "grad_norm": 1.8169553166426808,
      "learning_rate": 2.7584798163782603e-07,
      "loss": 2.4248,
      "step": 5560
    },
    {
      "epoch": 0.719810984885264,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5099875926971436,
      "eval_runtime": 14.6133,
      "eval_samples_per_second": 3.011,
      "eval_steps_per_second": 0.205,
      "step": 5560
    },
    {
      "epoch": 0.7199404472926174,
      "grad_norm": 2.3093845859188873,
      "learning_rate": 2.7561372886855244e-07,
      "loss": 2.5293,
      "step": 5561
    },
    {
      "epoch": 0.7200699096999709,
      "grad_norm": 1.435623281551701,
      "learning_rate": 2.7537954746850686e-07,
      "loss": 2.3755,
      "step": 5562
    },
    {
      "epoch": 0.7201993721073243,
      "grad_norm": 1.1289074039455078,
      "learning_rate": 2.751454374855251e-07,
      "loss": 2.1689,
      "step": 5563
    },
    {
      "epoch": 0.7203288345146778,
      "grad_norm": 2.1193161850721127,
      "learning_rate": 2.7491139896742873e-07,
      "loss": 2.3691,
      "step": 5564
    },
    {
      "epoch": 0.7204582969220312,
      "grad_norm": 1.6980072454782136,
      "learning_rate": 2.746774319620253e-07,
      "loss": 2.4399,
      "step": 5565
    },
    {
      "epoch": 0.7204582969220312,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.50927734375,
      "eval_runtime": 13.0915,
      "eval_samples_per_second": 3.361,
      "eval_steps_per_second": 0.229,
      "step": 5565
    },
    {
      "epoch": 0.7205877593293847,
      "grad_norm": 1.5075953458803324,
      "learning_rate": 2.744435365171069e-07,
      "loss": 2.3982,
      "step": 5566
    },
    {
      "epoch": 0.7207172217367382,
      "grad_norm": 1.6918291383960713,
      "learning_rate": 2.7420971268045167e-07,
      "loss": 2.4092,
      "step": 5567
    },
    {
      "epoch": 0.7208466841440917,
      "grad_norm": 1.2506097152652387,
      "learning_rate": 2.739759604998232e-07,
      "loss": 2.2695,
      "step": 5568
    },
    {
      "epoch": 0.7209761465514452,
      "grad_norm": 2.954445096213623,
      "learning_rate": 2.7374228002296934e-07,
      "loss": 2.6201,
      "step": 5569
    },
    {
      "epoch": 0.7211056089587986,
      "grad_norm": 1.3859044433106302,
      "learning_rate": 2.7350867129762447e-07,
      "loss": 2.356,
      "step": 5570
    },
    {
      "epoch": 0.7211056089587986,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5093661546707153,
      "eval_runtime": 14.3231,
      "eval_samples_per_second": 3.072,
      "eval_steps_per_second": 0.209,
      "step": 5570
    },
    {
      "epoch": 0.7212350713661521,
      "grad_norm": 1.5674237481977822,
      "learning_rate": 2.7327513437150807e-07,
      "loss": 2.3964,
      "step": 5571
    },
    {
      "epoch": 0.7213645337735055,
      "grad_norm": 2.379028384116191,
      "learning_rate": 2.730416692923244e-07,
      "loss": 2.4958,
      "step": 5572
    },
    {
      "epoch": 0.721493996180859,
      "grad_norm": 1.5455502487512371,
      "learning_rate": 2.728082761077639e-07,
      "loss": 2.5,
      "step": 5573
    },
    {
      "epoch": 0.7216234585882124,
      "grad_norm": 1.7377203006836992,
      "learning_rate": 2.725749548655016e-07,
      "loss": 2.3423,
      "step": 5574
    },
    {
      "epoch": 0.7217529209955659,
      "grad_norm": 1.6312216160984048,
      "learning_rate": 2.723417056131978e-07,
      "loss": 2.4758,
      "step": 5575
    },
    {
      "epoch": 0.7217529209955659,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.508589267730713,
      "eval_runtime": 14.4312,
      "eval_samples_per_second": 3.049,
      "eval_steps_per_second": 0.208,
      "step": 5575
    },
    {
      "epoch": 0.7218823834029193,
      "grad_norm": 1.6272893864648796,
      "learning_rate": 2.7210852839849896e-07,
      "loss": 2.4421,
      "step": 5576
    },
    {
      "epoch": 0.7220118458102729,
      "grad_norm": 1.6615919727143313,
      "learning_rate": 2.7187542326903576e-07,
      "loss": 2.3574,
      "step": 5577
    },
    {
      "epoch": 0.7221413082176263,
      "grad_norm": 2.6841609605575374,
      "learning_rate": 2.7164239027242507e-07,
      "loss": 2.5779,
      "step": 5578
    },
    {
      "epoch": 0.7222707706249798,
      "grad_norm": 2.1363165319634607,
      "learning_rate": 2.7140942945626844e-07,
      "loss": 2.3572,
      "step": 5579
    },
    {
      "epoch": 0.7224002330323332,
      "grad_norm": 2.1343990824011305,
      "learning_rate": 2.7117654086815257e-07,
      "loss": 2.2312,
      "step": 5580
    },
    {
      "epoch": 0.7224002330323332,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5079013109207153,
      "eval_runtime": 14.0868,
      "eval_samples_per_second": 3.123,
      "eval_steps_per_second": 0.213,
      "step": 5580
    },
    {
      "epoch": 0.7225296954396867,
      "grad_norm": 1.2388845462084674,
      "learning_rate": 2.709437245556501e-07,
      "loss": 2.2566,
      "step": 5581
    },
    {
      "epoch": 0.7226591578470402,
      "grad_norm": 1.5019589883987903,
      "learning_rate": 2.7071098056631805e-07,
      "loss": 2.4832,
      "step": 5582
    },
    {
      "epoch": 0.7227886202543936,
      "grad_norm": 1.3062292130862347,
      "learning_rate": 2.7047830894769956e-07,
      "loss": 2.3442,
      "step": 5583
    },
    {
      "epoch": 0.7229180826617471,
      "grad_norm": 1.979726227766332,
      "learning_rate": 2.702457097473221e-07,
      "loss": 2.3982,
      "step": 5584
    },
    {
      "epoch": 0.7230475450691005,
      "grad_norm": 2.37744644501221,
      "learning_rate": 2.7001318301269917e-07,
      "loss": 2.4646,
      "step": 5585
    },
    {
      "epoch": 0.7230475450691005,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5107200145721436,
      "eval_runtime": 14.277,
      "eval_samples_per_second": 3.082,
      "eval_steps_per_second": 0.21,
      "step": 5585
    },
    {
      "epoch": 0.723177007476454,
      "grad_norm": 1.3507179228392634,
      "learning_rate": 2.6978072879132885e-07,
      "loss": 2.2634,
      "step": 5586
    },
    {
      "epoch": 0.7233064698838075,
      "grad_norm": 1.840242546852928,
      "learning_rate": 2.695483471306945e-07,
      "loss": 2.519,
      "step": 5587
    },
    {
      "epoch": 0.723435932291161,
      "grad_norm": 1.5270676035336725,
      "learning_rate": 2.6931603807826506e-07,
      "loss": 2.3647,
      "step": 5588
    },
    {
      "epoch": 0.7235653946985144,
      "grad_norm": 1.1749397727549062,
      "learning_rate": 2.69083801681494e-07,
      "loss": 2.3071,
      "step": 5589
    },
    {
      "epoch": 0.7236948571058679,
      "grad_norm": 1.3322640689962522,
      "learning_rate": 2.688516379878208e-07,
      "loss": 2.3247,
      "step": 5590
    },
    {
      "epoch": 0.7236948571058679,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.508500576019287,
      "eval_runtime": 14.9033,
      "eval_samples_per_second": 2.952,
      "eval_steps_per_second": 0.201,
      "step": 5590
    },
    {
      "epoch": 0.7238243195132213,
      "grad_norm": 1.705210521195304,
      "learning_rate": 2.686195470446693e-07,
      "loss": 2.4436,
      "step": 5591
    },
    {
      "epoch": 0.7239537819205748,
      "grad_norm": 2.192928496883699,
      "learning_rate": 2.683875288994487e-07,
      "loss": 2.4543,
      "step": 5592
    },
    {
      "epoch": 0.7240832443279283,
      "grad_norm": 2.7342780910093536,
      "learning_rate": 2.6815558359955343e-07,
      "loss": 2.2744,
      "step": 5593
    },
    {
      "epoch": 0.7242127067352817,
      "grad_norm": 1.4334541547349253,
      "learning_rate": 2.6792371119236335e-07,
      "loss": 2.5435,
      "step": 5594
    },
    {
      "epoch": 0.7243421691426352,
      "grad_norm": 1.703577301806479,
      "learning_rate": 2.6769191172524267e-07,
      "loss": 2.3723,
      "step": 5595
    },
    {
      "epoch": 0.7243421691426352,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5086115598678589,
      "eval_runtime": 14.8315,
      "eval_samples_per_second": 2.967,
      "eval_steps_per_second": 0.202,
      "step": 5595
    },
    {
      "epoch": 0.7244716315499887,
      "grad_norm": 2.085591545761733,
      "learning_rate": 2.6746018524554153e-07,
      "loss": 2.564,
      "step": 5596
    },
    {
      "epoch": 0.7246010939573422,
      "grad_norm": 1.413747440699509,
      "learning_rate": 2.6722853180059453e-07,
      "loss": 2.459,
      "step": 5597
    },
    {
      "epoch": 0.7247305563646956,
      "grad_norm": 1.3260162174318653,
      "learning_rate": 2.6699695143772143e-07,
      "loss": 2.3755,
      "step": 5598
    },
    {
      "epoch": 0.7248600187720491,
      "grad_norm": 2.32683899665351,
      "learning_rate": 2.6676544420422763e-07,
      "loss": 2.407,
      "step": 5599
    },
    {
      "epoch": 0.7249894811794025,
      "grad_norm": 1.7434693139780224,
      "learning_rate": 2.6653401014740266e-07,
      "loss": 2.4805,
      "step": 5600
    },
    {
      "epoch": 0.7249894811794025,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5093661546707153,
      "eval_runtime": 13.5865,
      "eval_samples_per_second": 3.239,
      "eval_steps_per_second": 0.221,
      "step": 5600
    },
    {
      "epoch": 0.725118943586756,
      "grad_norm": 1.3972322914181314,
      "learning_rate": 2.6630264931452215e-07,
      "loss": 2.2969,
      "step": 5601
    },
    {
      "epoch": 0.7252484059941094,
      "grad_norm": 1.2590977917336583,
      "learning_rate": 2.6607136175284597e-07,
      "loss": 2.3389,
      "step": 5602
    },
    {
      "epoch": 0.7253778684014629,
      "grad_norm": 1.760524997475405,
      "learning_rate": 2.658401475096191e-07,
      "loss": 2.3555,
      "step": 5603
    },
    {
      "epoch": 0.7255073308088164,
      "grad_norm": 1.6595253890090977,
      "learning_rate": 2.6560900663207213e-07,
      "loss": 2.5142,
      "step": 5604
    },
    {
      "epoch": 0.7256367932161698,
      "grad_norm": 1.8617391917069044,
      "learning_rate": 2.6537793916741987e-07,
      "loss": 2.4392,
      "step": 5605
    },
    {
      "epoch": 0.7256367932161698,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5079900026321411,
      "eval_runtime": 14.0275,
      "eval_samples_per_second": 3.137,
      "eval_steps_per_second": 0.214,
      "step": 5605
    },
    {
      "epoch": 0.7257662556235234,
      "grad_norm": 1.4913957516562677,
      "learning_rate": 2.6514694516286294e-07,
      "loss": 2.2512,
      "step": 5606
    },
    {
      "epoch": 0.7258957180308768,
      "grad_norm": 2.006513836069113,
      "learning_rate": 2.6491602466558636e-07,
      "loss": 2.2847,
      "step": 5607
    },
    {
      "epoch": 0.7260251804382303,
      "grad_norm": 1.629313339312131,
      "learning_rate": 2.6468517772276014e-07,
      "loss": 2.4475,
      "step": 5608
    },
    {
      "epoch": 0.7261546428455837,
      "grad_norm": 1.5139261830156856,
      "learning_rate": 2.6445440438153975e-07,
      "loss": 2.4216,
      "step": 5609
    },
    {
      "epoch": 0.7262841052529372,
      "grad_norm": 1.5199421615926034,
      "learning_rate": 2.64223704689065e-07,
      "loss": 2.3335,
      "step": 5610
    },
    {
      "epoch": 0.7262841052529372,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.509321689605713,
      "eval_runtime": 14.6745,
      "eval_samples_per_second": 2.998,
      "eval_steps_per_second": 0.204,
      "step": 5610
    },
    {
      "epoch": 0.7264135676602906,
      "grad_norm": 1.7194164107061336,
      "learning_rate": 2.639930786924612e-07,
      "loss": 2.3953,
      "step": 5611
    },
    {
      "epoch": 0.7265430300676441,
      "grad_norm": 1.9642260942677225,
      "learning_rate": 2.637625264388388e-07,
      "loss": 2.4902,
      "step": 5612
    },
    {
      "epoch": 0.7266724924749975,
      "grad_norm": 2.1264741538545775,
      "learning_rate": 2.635320479752919e-07,
      "loss": 2.5557,
      "step": 5613
    },
    {
      "epoch": 0.726801954882351,
      "grad_norm": 2.5791993356191707,
      "learning_rate": 2.63301643348901e-07,
      "loss": 2.5876,
      "step": 5614
    },
    {
      "epoch": 0.7269314172897045,
      "grad_norm": 2.5041629726106733,
      "learning_rate": 2.6307131260673053e-07,
      "loss": 2.5605,
      "step": 5615
    },
    {
      "epoch": 0.7269314172897045,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5086559057235718,
      "eval_runtime": 14.8714,
      "eval_samples_per_second": 2.959,
      "eval_steps_per_second": 0.202,
      "step": 5615
    },
    {
      "epoch": 0.727060879697058,
      "grad_norm": 1.2063006392780222,
      "learning_rate": 2.628410557958304e-07,
      "loss": 2.2524,
      "step": 5616
    },
    {
      "epoch": 0.7271903421044115,
      "grad_norm": 2.8536342445032203,
      "learning_rate": 2.626108729632355e-07,
      "loss": 2.2452,
      "step": 5617
    },
    {
      "epoch": 0.7273198045117649,
      "grad_norm": 1.367484206121569,
      "learning_rate": 2.623807641559651e-07,
      "loss": 2.4668,
      "step": 5618
    },
    {
      "epoch": 0.7274492669191184,
      "grad_norm": 2.4325491839380176,
      "learning_rate": 2.6215072942102334e-07,
      "loss": 2.6729,
      "step": 5619
    },
    {
      "epoch": 0.7275787293264718,
      "grad_norm": 1.9967059733797172,
      "learning_rate": 2.6192076880539994e-07,
      "loss": 2.4453,
      "step": 5620
    },
    {
      "epoch": 0.7275787293264718,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.505859375,
      "eval_runtime": 14.5143,
      "eval_samples_per_second": 3.031,
      "eval_steps_per_second": 0.207,
      "step": 5620
    },
    {
      "epoch": 0.7277081917338253,
      "grad_norm": 2.4584002698532785,
      "learning_rate": 2.6169088235606854e-07,
      "loss": 2.4597,
      "step": 5621
    },
    {
      "epoch": 0.7278376541411787,
      "grad_norm": 1.9104850311196249,
      "learning_rate": 2.6146107011998865e-07,
      "loss": 2.2139,
      "step": 5622
    },
    {
      "epoch": 0.7279671165485322,
      "grad_norm": 1.9890462841262417,
      "learning_rate": 2.6123133214410347e-07,
      "loss": 2.2876,
      "step": 5623
    },
    {
      "epoch": 0.7280965789558856,
      "grad_norm": 1.5827879729141614,
      "learning_rate": 2.610016684753423e-07,
      "loss": 2.4524,
      "step": 5624
    },
    {
      "epoch": 0.7282260413632391,
      "grad_norm": 1.488852333731658,
      "learning_rate": 2.607720791606181e-07,
      "loss": 2.4172,
      "step": 5625
    },
    {
      "epoch": 0.7282260413632391,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.507279872894287,
      "eval_runtime": 14.6041,
      "eval_samples_per_second": 3.013,
      "eval_steps_per_second": 0.205,
      "step": 5625
    },
    {
      "epoch": 0.7283555037705927,
      "grad_norm": 1.7973964371338502,
      "learning_rate": 2.605425642468292e-07,
      "loss": 2.657,
      "step": 5626
    },
    {
      "epoch": 0.7284849661779461,
      "grad_norm": 2.02635983372142,
      "learning_rate": 2.603131237808589e-07,
      "loss": 2.4758,
      "step": 5627
    },
    {
      "epoch": 0.7286144285852996,
      "grad_norm": 2.1844991879154714,
      "learning_rate": 2.6008375780957483e-07,
      "loss": 2.4307,
      "step": 5628
    },
    {
      "epoch": 0.728743890992653,
      "grad_norm": 2.090889483271363,
      "learning_rate": 2.5985446637982984e-07,
      "loss": 2.3623,
      "step": 5629
    },
    {
      "epoch": 0.7288733534000065,
      "grad_norm": 1.9441238857189067,
      "learning_rate": 2.596252495384613e-07,
      "loss": 2.5552,
      "step": 5630
    },
    {
      "epoch": 0.7288733534000065,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5057040452957153,
      "eval_runtime": 14.3437,
      "eval_samples_per_second": 3.068,
      "eval_steps_per_second": 0.209,
      "step": 5630
    },
    {
      "epoch": 0.7290028158073599,
      "grad_norm": 1.3746058017043443,
      "learning_rate": 2.5939610733229115e-07,
      "loss": 2.3572,
      "step": 5631
    },
    {
      "epoch": 0.7291322782147134,
      "grad_norm": 1.9874513263203206,
      "learning_rate": 2.591670398081268e-07,
      "loss": 2.3345,
      "step": 5632
    },
    {
      "epoch": 0.7292617406220668,
      "grad_norm": 1.7802389172942312,
      "learning_rate": 2.5893804701275936e-07,
      "loss": 2.3894,
      "step": 5633
    },
    {
      "epoch": 0.7293912030294203,
      "grad_norm": 1.6169981298725404,
      "learning_rate": 2.5870912899296553e-07,
      "loss": 2.2141,
      "step": 5634
    },
    {
      "epoch": 0.7295206654367739,
      "grad_norm": 1.3814381541868355,
      "learning_rate": 2.58480285795507e-07,
      "loss": 2.3362,
      "step": 5635
    },
    {
      "epoch": 0.7295206654367739,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.505815029144287,
      "eval_runtime": 13.6086,
      "eval_samples_per_second": 3.233,
      "eval_steps_per_second": 0.22,
      "step": 5635
    },
    {
      "epoch": 0.7296501278441273,
      "grad_norm": 1.179809472767984,
      "learning_rate": 2.5825151746712864e-07,
      "loss": 2.2295,
      "step": 5636
    },
    {
      "epoch": 0.7297795902514808,
      "grad_norm": 1.8873093220526136,
      "learning_rate": 2.5802282405456174e-07,
      "loss": 2.3521,
      "step": 5637
    },
    {
      "epoch": 0.7299090526588342,
      "grad_norm": 1.3724222177448588,
      "learning_rate": 2.57794205604521e-07,
      "loss": 2.3384,
      "step": 5638
    },
    {
      "epoch": 0.7300385150661877,
      "grad_norm": 1.6424070725179651,
      "learning_rate": 2.5756566216370673e-07,
      "loss": 2.438,
      "step": 5639
    },
    {
      "epoch": 0.7301679774735411,
      "grad_norm": 1.5747823883073737,
      "learning_rate": 2.5733719377880405e-07,
      "loss": 2.4412,
      "step": 5640
    },
    {
      "epoch": 0.7301679774735411,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5072354078292847,
      "eval_runtime": 15.4701,
      "eval_samples_per_second": 2.844,
      "eval_steps_per_second": 0.194,
      "step": 5640
    },
    {
      "epoch": 0.7302974398808946,
      "grad_norm": 1.35759617611263,
      "learning_rate": 2.571088004964812e-07,
      "loss": 2.4189,
      "step": 5641
    },
    {
      "epoch": 0.730426902288248,
      "grad_norm": 1.6387510116346904,
      "learning_rate": 2.5688048236339266e-07,
      "loss": 2.4739,
      "step": 5642
    },
    {
      "epoch": 0.7305563646956015,
      "grad_norm": 2.3020838929599043,
      "learning_rate": 2.566522394261773e-07,
      "loss": 2.3151,
      "step": 5643
    },
    {
      "epoch": 0.7306858271029549,
      "grad_norm": 1.5935686621172203,
      "learning_rate": 2.564240717314579e-07,
      "loss": 2.3442,
      "step": 5644
    },
    {
      "epoch": 0.7308152895103085,
      "grad_norm": 1.6830966785197252,
      "learning_rate": 2.5619597932584267e-07,
      "loss": 2.3645,
      "step": 5645
    },
    {
      "epoch": 0.7308152895103085,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5021750926971436,
      "eval_runtime": 14.3982,
      "eval_samples_per_second": 3.056,
      "eval_steps_per_second": 0.208,
      "step": 5645
    },
    {
      "epoch": 0.730944751917662,
      "grad_norm": 1.9970508536071547,
      "learning_rate": 2.5596796225592407e-07,
      "loss": 2.4727,
      "step": 5646
    },
    {
      "epoch": 0.7310742143250154,
      "grad_norm": 1.9898553690333884,
      "learning_rate": 2.557400205682789e-07,
      "loss": 2.4949,
      "step": 5647
    },
    {
      "epoch": 0.7312036767323689,
      "grad_norm": 1.9148069220140647,
      "learning_rate": 2.555121543094693e-07,
      "loss": 2.313,
      "step": 5648
    },
    {
      "epoch": 0.7313331391397223,
      "grad_norm": 1.2900324130894611,
      "learning_rate": 2.5528436352604107e-07,
      "loss": 2.3911,
      "step": 5649
    },
    {
      "epoch": 0.7314626015470758,
      "grad_norm": 1.284739442695453,
      "learning_rate": 2.5505664826452566e-07,
      "loss": 2.1541,
      "step": 5650
    },
    {
      "epoch": 0.7314626015470758,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.506547451019287,
      "eval_runtime": 14.1461,
      "eval_samples_per_second": 3.11,
      "eval_steps_per_second": 0.212,
      "step": 5650
    },
    {
      "epoch": 0.7315920639544292,
      "grad_norm": 1.5182764140197063,
      "learning_rate": 2.5482900857143825e-07,
      "loss": 2.3674,
      "step": 5651
    },
    {
      "epoch": 0.7317215263617827,
      "grad_norm": 1.4357677924074022,
      "learning_rate": 2.5460144449327865e-07,
      "loss": 2.2288,
      "step": 5652
    },
    {
      "epoch": 0.7318509887691361,
      "grad_norm": 1.378134534669576,
      "learning_rate": 2.5437395607653183e-07,
      "loss": 2.4429,
      "step": 5653
    },
    {
      "epoch": 0.7319804511764896,
      "grad_norm": 1.3953457573399068,
      "learning_rate": 2.5414654336766655e-07,
      "loss": 2.457,
      "step": 5654
    },
    {
      "epoch": 0.7321099135838431,
      "grad_norm": 1.418139383621082,
      "learning_rate": 2.5391920641313684e-07,
      "loss": 2.4565,
      "step": 5655
    },
    {
      "epoch": 0.7321099135838431,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.503662109375,
      "eval_runtime": 15.0866,
      "eval_samples_per_second": 2.916,
      "eval_steps_per_second": 0.199,
      "step": 5655
    },
    {
      "epoch": 0.7322393759911966,
      "grad_norm": 1.6689871638675509,
      "learning_rate": 2.536919452593805e-07,
      "loss": 2.3201,
      "step": 5656
    },
    {
      "epoch": 0.73236883839855,
      "grad_norm": 1.457188260016672,
      "learning_rate": 2.534647599528206e-07,
      "loss": 2.5325,
      "step": 5657
    },
    {
      "epoch": 0.7324983008059035,
      "grad_norm": 1.09003263851734,
      "learning_rate": 2.5323765053986413e-07,
      "loss": 2.1025,
      "step": 5658
    },
    {
      "epoch": 0.732627763213257,
      "grad_norm": 1.6983701917924374,
      "learning_rate": 2.5301061706690266e-07,
      "loss": 2.3052,
      "step": 5659
    },
    {
      "epoch": 0.7327572256206104,
      "grad_norm": 2.0924030152886584,
      "learning_rate": 2.527836595803127e-07,
      "loss": 2.584,
      "step": 5660
    },
    {
      "epoch": 0.7327572256206104,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5094548463821411,
      "eval_runtime": 19.5867,
      "eval_samples_per_second": 2.246,
      "eval_steps_per_second": 0.153,
      "step": 5660
    },
    {
      "epoch": 0.7328866880279639,
      "grad_norm": 1.5504150261635312,
      "learning_rate": 2.5255677812645455e-07,
      "loss": 2.1606,
      "step": 5661
    },
    {
      "epoch": 0.7330161504353173,
      "grad_norm": 2.4354298586938357,
      "learning_rate": 2.5232997275167354e-07,
      "loss": 2.4573,
      "step": 5662
    },
    {
      "epoch": 0.7331456128426708,
      "grad_norm": 2.971163976364456,
      "learning_rate": 2.521032435022998e-07,
      "loss": 2.2607,
      "step": 5663
    },
    {
      "epoch": 0.7332750752500242,
      "grad_norm": 1.9232236768873372,
      "learning_rate": 2.5187659042464633e-07,
      "loss": 2.6726,
      "step": 5664
    },
    {
      "epoch": 0.7334045376573778,
      "grad_norm": 3.5001323145717405,
      "learning_rate": 2.516500135650121e-07,
      "loss": 2.8154,
      "step": 5665
    },
    {
      "epoch": 0.7334045376573778,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5008655786514282,
      "eval_runtime": 14.8822,
      "eval_samples_per_second": 2.957,
      "eval_steps_per_second": 0.202,
      "step": 5665
    },
    {
      "epoch": 0.7335340000647312,
      "grad_norm": 1.7902291812424176,
      "learning_rate": 2.514235129696803e-07,
      "loss": 2.4338,
      "step": 5666
    },
    {
      "epoch": 0.7336634624720847,
      "grad_norm": 2.6247102366354045,
      "learning_rate": 2.5119708868491783e-07,
      "loss": 2.4045,
      "step": 5667
    },
    {
      "epoch": 0.7337929248794381,
      "grad_norm": 2.2188401294436524,
      "learning_rate": 2.509707407569771e-07,
      "loss": 2.3306,
      "step": 5668
    },
    {
      "epoch": 0.7339223872867916,
      "grad_norm": 1.615794595158225,
      "learning_rate": 2.507444692320933e-07,
      "loss": 2.4666,
      "step": 5669
    },
    {
      "epoch": 0.734051849694145,
      "grad_norm": 1.9520027225816075,
      "learning_rate": 2.505182741564875e-07,
      "loss": 2.5312,
      "step": 5670
    },
    {
      "epoch": 0.734051849694145,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5079456567764282,
      "eval_runtime": 14.6782,
      "eval_samples_per_second": 2.998,
      "eval_steps_per_second": 0.204,
      "step": 5670
    },
    {
      "epoch": 0.7341813121014985,
      "grad_norm": 1.4703596299837818,
      "learning_rate": 2.502921555763648e-07,
      "loss": 2.5527,
      "step": 5671
    },
    {
      "epoch": 0.734310774508852,
      "grad_norm": 1.3541719162793477,
      "learning_rate": 2.5006611353791415e-07,
      "loss": 2.4666,
      "step": 5672
    },
    {
      "epoch": 0.7344402369162054,
      "grad_norm": 2.0854522050832447,
      "learning_rate": 2.4984014808730954e-07,
      "loss": 2.354,
      "step": 5673
    },
    {
      "epoch": 0.7345696993235589,
      "grad_norm": 2.2035598476389886,
      "learning_rate": 2.4961425927070883e-07,
      "loss": 2.5095,
      "step": 5674
    },
    {
      "epoch": 0.7346991617309124,
      "grad_norm": 1.2488568471900237,
      "learning_rate": 2.493884471342542e-07,
      "loss": 2.3457,
      "step": 5675
    },
    {
      "epoch": 0.7346991617309124,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5079678297042847,
      "eval_runtime": 15.4707,
      "eval_samples_per_second": 2.844,
      "eval_steps_per_second": 0.194,
      "step": 5675
    },
    {
      "epoch": 0.7348286241382659,
      "grad_norm": 2.1526248325363726,
      "learning_rate": 2.4916271172407275e-07,
      "loss": 2.6167,
      "step": 5676
    },
    {
      "epoch": 0.7349580865456193,
      "grad_norm": 1.7487231587904244,
      "learning_rate": 2.48937053086275e-07,
      "loss": 2.4727,
      "step": 5677
    },
    {
      "epoch": 0.7350875489529728,
      "grad_norm": 2.059881484230393,
      "learning_rate": 2.4871147126695694e-07,
      "loss": 2.4202,
      "step": 5678
    },
    {
      "epoch": 0.7352170113603262,
      "grad_norm": 2.0333970419315985,
      "learning_rate": 2.4848596631219787e-07,
      "loss": 2.2969,
      "step": 5679
    },
    {
      "epoch": 0.7353464737676797,
      "grad_norm": 1.7041612371808403,
      "learning_rate": 2.4826053826806154e-07,
      "loss": 2.5162,
      "step": 5680
    },
    {
      "epoch": 0.7353464737676797,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.504438877105713,
      "eval_runtime": 14.6295,
      "eval_samples_per_second": 3.008,
      "eval_steps_per_second": 0.205,
      "step": 5680
    },
    {
      "epoch": 0.7354759361750332,
      "grad_norm": 2.3093911467602593,
      "learning_rate": 2.480351871805966e-07,
      "loss": 2.3008,
      "step": 5681
    },
    {
      "epoch": 0.7356053985823866,
      "grad_norm": 2.3891438360485924,
      "learning_rate": 2.478099130958352e-07,
      "loss": 2.4646,
      "step": 5682
    },
    {
      "epoch": 0.7357348609897401,
      "grad_norm": 2.5342601961660143,
      "learning_rate": 2.4758471605979466e-07,
      "loss": 2.5247,
      "step": 5683
    },
    {
      "epoch": 0.7358643233970936,
      "grad_norm": 2.686969533358233,
      "learning_rate": 2.4735959611847543e-07,
      "loss": 2.593,
      "step": 5684
    },
    {
      "epoch": 0.7359937858044471,
      "grad_norm": 2.404409337208305,
      "learning_rate": 2.471345533178633e-07,
      "loss": 2.7876,
      "step": 5685
    },
    {
      "epoch": 0.7359937858044471,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.505815029144287,
      "eval_runtime": 14.3501,
      "eval_samples_per_second": 3.066,
      "eval_steps_per_second": 0.209,
      "step": 5685
    },
    {
      "epoch": 0.7361232482118005,
      "grad_norm": 1.8538153353750857,
      "learning_rate": 2.469095877039277e-07,
      "loss": 2.1504,
      "step": 5686
    },
    {
      "epoch": 0.736252710619154,
      "grad_norm": 3.4703900312985594,
      "learning_rate": 2.466846993226221e-07,
      "loss": 2.3835,
      "step": 5687
    },
    {
      "epoch": 0.7363821730265074,
      "grad_norm": 1.512334269338688,
      "learning_rate": 2.4645988821988484e-07,
      "loss": 2.437,
      "step": 5688
    },
    {
      "epoch": 0.7365116354338609,
      "grad_norm": 1.8137655117001923,
      "learning_rate": 2.462351544416382e-07,
      "loss": 2.1697,
      "step": 5689
    },
    {
      "epoch": 0.7366410978412143,
      "grad_norm": 1.8610060580286094,
      "learning_rate": 2.4601049803378854e-07,
      "loss": 2.3376,
      "step": 5690
    },
    {
      "epoch": 0.7366410978412143,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5065251588821411,
      "eval_runtime": 14.5886,
      "eval_samples_per_second": 3.016,
      "eval_steps_per_second": 0.206,
      "step": 5690
    },
    {
      "epoch": 0.7367705602485678,
      "grad_norm": 3.0868821327146914,
      "learning_rate": 2.457859190422262e-07,
      "loss": 2.3054,
      "step": 5691
    },
    {
      "epoch": 0.7369000226559212,
      "grad_norm": 2.446197983909844,
      "learning_rate": 2.455614175128264e-07,
      "loss": 2.53,
      "step": 5692
    },
    {
      "epoch": 0.7370294850632747,
      "grad_norm": 1.5213681017230851,
      "learning_rate": 2.453369934914477e-07,
      "loss": 2.1814,
      "step": 5693
    },
    {
      "epoch": 0.7371589474706283,
      "grad_norm": 2.303615540528132,
      "learning_rate": 2.451126470239338e-07,
      "loss": 2.4673,
      "step": 5694
    },
    {
      "epoch": 0.7372884098779817,
      "grad_norm": 1.4862575777019944,
      "learning_rate": 2.448883781561114e-07,
      "loss": 2.3198,
      "step": 5695
    },
    {
      "epoch": 0.7372884098779817,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5035732984542847,
      "eval_runtime": 13.9781,
      "eval_samples_per_second": 3.148,
      "eval_steps_per_second": 0.215,
      "step": 5695
    },
    {
      "epoch": 0.7374178722853352,
      "grad_norm": 2.4251094151406334,
      "learning_rate": 2.4466418693379253e-07,
      "loss": 2.2124,
      "step": 5696
    },
    {
      "epoch": 0.7375473346926886,
      "grad_norm": 2.588251885086974,
      "learning_rate": 2.4444007340277257e-07,
      "loss": 2.6187,
      "step": 5697
    },
    {
      "epoch": 0.7376767971000421,
      "grad_norm": 1.6953634322425963,
      "learning_rate": 2.4421603760883095e-07,
      "loss": 2.5024,
      "step": 5698
    },
    {
      "epoch": 0.7378062595073955,
      "grad_norm": 2.390150542524861,
      "learning_rate": 2.4399207959773207e-07,
      "loss": 2.2717,
      "step": 5699
    },
    {
      "epoch": 0.737935721914749,
      "grad_norm": 3.1099283204917243,
      "learning_rate": 2.4376819941522333e-07,
      "loss": 2.6899,
      "step": 5700
    },
    {
      "epoch": 0.737935721914749,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5071688890457153,
      "eval_runtime": 13.9585,
      "eval_samples_per_second": 3.152,
      "eval_steps_per_second": 0.215,
      "step": 5700
    },
    {
      "epoch": 0.7380651843221024,
      "grad_norm": 1.636977704177955,
      "learning_rate": 2.435443971070373e-07,
      "loss": 2.2634,
      "step": 5701
    },
    {
      "epoch": 0.7381946467294559,
      "grad_norm": 3.7066921817979797,
      "learning_rate": 2.433206727188899e-07,
      "loss": 2.7783,
      "step": 5702
    },
    {
      "epoch": 0.7383241091368093,
      "grad_norm": 1.200859903233997,
      "learning_rate": 2.4309702629648124e-07,
      "loss": 2.2935,
      "step": 5703
    },
    {
      "epoch": 0.7384535715441629,
      "grad_norm": 1.3806974133034717,
      "learning_rate": 2.428734578854959e-07,
      "loss": 2.3254,
      "step": 5704
    },
    {
      "epoch": 0.7385830339515164,
      "grad_norm": 1.4230174637008688,
      "learning_rate": 2.42649967531602e-07,
      "loss": 2.3352,
      "step": 5705
    },
    {
      "epoch": 0.7385830339515164,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5049716234207153,
      "eval_runtime": 14.5211,
      "eval_samples_per_second": 3.03,
      "eval_steps_per_second": 0.207,
      "step": 5705
    },
    {
      "epoch": 0.7387124963588698,
      "grad_norm": 1.3485431936495211,
      "learning_rate": 2.424265552804521e-07,
      "loss": 2.5105,
      "step": 5706
    },
    {
      "epoch": 0.7388419587662233,
      "grad_norm": 1.250070075372513,
      "learning_rate": 2.4220322117768313e-07,
      "loss": 2.2883,
      "step": 5707
    },
    {
      "epoch": 0.7389714211735767,
      "grad_norm": 1.835187632190345,
      "learning_rate": 2.419799652689147e-07,
      "loss": 2.5251,
      "step": 5708
    },
    {
      "epoch": 0.7391008835809302,
      "grad_norm": 2.011807836136752,
      "learning_rate": 2.417567875997521e-07,
      "loss": 2.46,
      "step": 5709
    },
    {
      "epoch": 0.7392303459882836,
      "grad_norm": 2.07268958026009,
      "learning_rate": 2.415336882157833e-07,
      "loss": 2.46,
      "step": 5710
    },
    {
      "epoch": 0.7392303459882836,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5065251588821411,
      "eval_runtime": 14.6732,
      "eval_samples_per_second": 2.999,
      "eval_steps_per_second": 0.204,
      "step": 5710
    },
    {
      "epoch": 0.7393598083956371,
      "grad_norm": 1.294458580347615,
      "learning_rate": 2.413106671625813e-07,
      "loss": 2.4053,
      "step": 5711
    },
    {
      "epoch": 0.7394892708029905,
      "grad_norm": 2.7758333889794105,
      "learning_rate": 2.410877244857027e-07,
      "loss": 2.3513,
      "step": 5712
    },
    {
      "epoch": 0.739618733210344,
      "grad_norm": 1.6304735267055515,
      "learning_rate": 2.4086486023068794e-07,
      "loss": 2.2817,
      "step": 5713
    },
    {
      "epoch": 0.7397481956176976,
      "grad_norm": 1.6245164929996876,
      "learning_rate": 2.406420744430614e-07,
      "loss": 2.343,
      "step": 5714
    },
    {
      "epoch": 0.739877658025051,
      "grad_norm": 1.8144417692251245,
      "learning_rate": 2.404193671683319e-07,
      "loss": 2.3185,
      "step": 5715
    },
    {
      "epoch": 0.739877658025051,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5079013109207153,
      "eval_runtime": 15.9937,
      "eval_samples_per_second": 2.751,
      "eval_steps_per_second": 0.188,
      "step": 5715
    },
    {
      "epoch": 0.7400071204324045,
      "grad_norm": 1.5295296340283542,
      "learning_rate": 2.4019673845199146e-07,
      "loss": 2.4756,
      "step": 5716
    },
    {
      "epoch": 0.7401365828397579,
      "grad_norm": 2.1146407143179546,
      "learning_rate": 2.3997418833951706e-07,
      "loss": 2.3105,
      "step": 5717
    },
    {
      "epoch": 0.7402660452471114,
      "grad_norm": 1.3504357321603342,
      "learning_rate": 2.397517168763689e-07,
      "loss": 2.2764,
      "step": 5718
    },
    {
      "epoch": 0.7403955076544648,
      "grad_norm": 1.6580937796607456,
      "learning_rate": 2.395293241079909e-07,
      "loss": 2.377,
      "step": 5719
    },
    {
      "epoch": 0.7405249700618183,
      "grad_norm": 1.832292336307156,
      "learning_rate": 2.393070100798118e-07,
      "loss": 2.2739,
      "step": 5720
    },
    {
      "epoch": 0.7405249700618183,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5051047801971436,
      "eval_runtime": 14.2528,
      "eval_samples_per_second": 3.087,
      "eval_steps_per_second": 0.21,
      "step": 5720
    },
    {
      "epoch": 0.7406544324691717,
      "grad_norm": 1.1558103016170562,
      "learning_rate": 2.390847748372433e-07,
      "loss": 2.2605,
      "step": 5721
    },
    {
      "epoch": 0.7407838948765252,
      "grad_norm": 2.2592009863859515,
      "learning_rate": 2.3886261842568194e-07,
      "loss": 2.5537,
      "step": 5722
    },
    {
      "epoch": 0.7409133572838787,
      "grad_norm": 1.1882650440240674,
      "learning_rate": 2.386405408905072e-07,
      "loss": 2.3228,
      "step": 5723
    },
    {
      "epoch": 0.7410428196912322,
      "grad_norm": 3.134434526580684,
      "learning_rate": 2.3841854227708337e-07,
      "loss": 2.6421,
      "step": 5724
    },
    {
      "epoch": 0.7411722820985857,
      "grad_norm": 2.5262998603920255,
      "learning_rate": 2.3819662263075783e-07,
      "loss": 2.4707,
      "step": 5725
    },
    {
      "epoch": 0.7411722820985857,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5014870166778564,
      "eval_runtime": 14.3866,
      "eval_samples_per_second": 3.058,
      "eval_steps_per_second": 0.209,
      "step": 5725
    },
    {
      "epoch": 0.7413017445059391,
      "grad_norm": 1.9346772369502783,
      "learning_rate": 2.379747819968621e-07,
      "loss": 2.3423,
      "step": 5726
    },
    {
      "epoch": 0.7414312069132926,
      "grad_norm": 1.5656660722330134,
      "learning_rate": 2.3775302042071207e-07,
      "loss": 2.4119,
      "step": 5727
    },
    {
      "epoch": 0.741560669320646,
      "grad_norm": 2.888111920311494,
      "learning_rate": 2.375313379476065e-07,
      "loss": 2.3584,
      "step": 5728
    },
    {
      "epoch": 0.7416901317279995,
      "grad_norm": 2.0321396189871703,
      "learning_rate": 2.3730973462282898e-07,
      "loss": 2.7117,
      "step": 5729
    },
    {
      "epoch": 0.7418195941353529,
      "grad_norm": 2.09991671564426,
      "learning_rate": 2.370882104916463e-07,
      "loss": 2.2905,
      "step": 5730
    },
    {
      "epoch": 0.7418195941353529,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5042835474014282,
      "eval_runtime": 14.1895,
      "eval_samples_per_second": 3.101,
      "eval_steps_per_second": 0.211,
      "step": 5730
    },
    {
      "epoch": 0.7419490565427064,
      "grad_norm": 1.149303201344051,
      "learning_rate": 2.36866765599309e-07,
      "loss": 2.3013,
      "step": 5731
    },
    {
      "epoch": 0.7420785189500598,
      "grad_norm": 1.4679625395047817,
      "learning_rate": 2.3664539999105216e-07,
      "loss": 2.1853,
      "step": 5732
    },
    {
      "epoch": 0.7422079813574134,
      "grad_norm": 2.24926289680569,
      "learning_rate": 2.364241137120938e-07,
      "loss": 2.3717,
      "step": 5733
    },
    {
      "epoch": 0.7423374437647668,
      "grad_norm": 1.9690265007605825,
      "learning_rate": 2.3620290680763624e-07,
      "loss": 2.5654,
      "step": 5734
    },
    {
      "epoch": 0.7424669061721203,
      "grad_norm": 2.437057319726628,
      "learning_rate": 2.3598177932286595e-07,
      "loss": 2.3335,
      "step": 5735
    },
    {
      "epoch": 0.7424669061721203,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.502974033355713,
      "eval_runtime": 15.1362,
      "eval_samples_per_second": 2.907,
      "eval_steps_per_second": 0.198,
      "step": 5735
    },
    {
      "epoch": 0.7425963685794738,
      "grad_norm": 2.2327673280717155,
      "learning_rate": 2.3576073130295184e-07,
      "loss": 2.4385,
      "step": 5736
    },
    {
      "epoch": 0.7427258309868272,
      "grad_norm": 1.9250282974682684,
      "learning_rate": 2.3553976279304787e-07,
      "loss": 2.4343,
      "step": 5737
    },
    {
      "epoch": 0.7428552933941807,
      "grad_norm": 1.1809759202796595,
      "learning_rate": 2.353188738382916e-07,
      "loss": 2.3423,
      "step": 5738
    },
    {
      "epoch": 0.7429847558015341,
      "grad_norm": 1.6868649266849252,
      "learning_rate": 2.3509806448380357e-07,
      "loss": 2.4597,
      "step": 5739
    },
    {
      "epoch": 0.7431142182088876,
      "grad_norm": 1.9366055483828428,
      "learning_rate": 2.348773347746893e-07,
      "loss": 2.2764,
      "step": 5740
    },
    {
      "epoch": 0.7431142182088876,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.505859375,
      "eval_runtime": 14.3837,
      "eval_samples_per_second": 3.059,
      "eval_steps_per_second": 0.209,
      "step": 5740
    },
    {
      "epoch": 0.743243680616241,
      "grad_norm": 2.4785060988254806,
      "learning_rate": 2.3465668475603636e-07,
      "loss": 2.3579,
      "step": 5741
    },
    {
      "epoch": 0.7433731430235945,
      "grad_norm": 2.721171272482728,
      "learning_rate": 2.344361144729175e-07,
      "loss": 2.6748,
      "step": 5742
    },
    {
      "epoch": 0.743502605430948,
      "grad_norm": 1.8367135148130516,
      "learning_rate": 2.3421562397038884e-07,
      "loss": 2.2817,
      "step": 5743
    },
    {
      "epoch": 0.7436320678383015,
      "grad_norm": 2.9016193425260175,
      "learning_rate": 2.3399521329348953e-07,
      "loss": 2.5791,
      "step": 5744
    },
    {
      "epoch": 0.743761530245655,
      "grad_norm": 1.5504410210589437,
      "learning_rate": 2.337748824872435e-07,
      "loss": 2.3179,
      "step": 5745
    },
    {
      "epoch": 0.743761530245655,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5057483911514282,
      "eval_runtime": 14.0921,
      "eval_samples_per_second": 3.122,
      "eval_steps_per_second": 0.213,
      "step": 5745
    },
    {
      "epoch": 0.7438909926530084,
      "grad_norm": 1.885305214076558,
      "learning_rate": 2.3355463159665749e-07,
      "loss": 2.49,
      "step": 5746
    },
    {
      "epoch": 0.7440204550603619,
      "grad_norm": 1.6665165486333617,
      "learning_rate": 2.33334460666722e-07,
      "loss": 2.5271,
      "step": 5747
    },
    {
      "epoch": 0.7441499174677153,
      "grad_norm": 1.7026733029610266,
      "learning_rate": 2.331143697424118e-07,
      "loss": 2.3281,
      "step": 5748
    },
    {
      "epoch": 0.7442793798750688,
      "grad_norm": 1.3933721389200384,
      "learning_rate": 2.3289435886868453e-07,
      "loss": 2.335,
      "step": 5749
    },
    {
      "epoch": 0.7444088422824222,
      "grad_norm": 1.116011354477912,
      "learning_rate": 2.3267442809048234e-07,
      "loss": 2.2881,
      "step": 5750
    },
    {
      "epoch": 0.7444088422824222,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5023082494735718,
      "eval_runtime": 17.0839,
      "eval_samples_per_second": 2.576,
      "eval_steps_per_second": 0.176,
      "step": 5750
    },
    {
      "epoch": 0.7445383046897757,
      "grad_norm": 1.4148398740897135,
      "learning_rate": 2.3245457745273003e-07,
      "loss": 2.3015,
      "step": 5751
    },
    {
      "epoch": 0.7446677670971291,
      "grad_norm": 2.117948750725282,
      "learning_rate": 2.3223480700033696e-07,
      "loss": 2.2646,
      "step": 5752
    },
    {
      "epoch": 0.7447972295044827,
      "grad_norm": 1.4051410784765883,
      "learning_rate": 2.3201511677819567e-07,
      "loss": 2.2461,
      "step": 5753
    },
    {
      "epoch": 0.7449266919118361,
      "grad_norm": 2.6867551357303148,
      "learning_rate": 2.3179550683118197e-07,
      "loss": 2.5649,
      "step": 5754
    },
    {
      "epoch": 0.7450561543191896,
      "grad_norm": 1.7292971611742591,
      "learning_rate": 2.3157597720415607e-07,
      "loss": 2.198,
      "step": 5755
    },
    {
      "epoch": 0.7450561543191896,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5043723583221436,
      "eval_runtime": 19.067,
      "eval_samples_per_second": 2.308,
      "eval_steps_per_second": 0.157,
      "step": 5755
    },
    {
      "epoch": 0.745185616726543,
      "grad_norm": 2.147978560830328,
      "learning_rate": 2.31356527941961e-07,
      "loss": 2.4733,
      "step": 5756
    },
    {
      "epoch": 0.7453150791338965,
      "grad_norm": 2.022867481141554,
      "learning_rate": 2.3113715908942403e-07,
      "loss": 2.3513,
      "step": 5757
    },
    {
      "epoch": 0.74544454154125,
      "grad_norm": 5.079058456847697,
      "learning_rate": 2.3091787069135563e-07,
      "loss": 2.824,
      "step": 5758
    },
    {
      "epoch": 0.7455740039486034,
      "grad_norm": 1.6758215541070285,
      "learning_rate": 2.3069866279254955e-07,
      "loss": 2.4097,
      "step": 5759
    },
    {
      "epoch": 0.7457034663559569,
      "grad_norm": 1.7644838942104952,
      "learning_rate": 2.3047953543778375e-07,
      "loss": 2.3442,
      "step": 5760
    },
    {
      "epoch": 0.7457034663559569,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4994229078292847,
      "eval_runtime": 13.7953,
      "eval_samples_per_second": 3.189,
      "eval_steps_per_second": 0.217,
      "step": 5760
    },
    {
      "epoch": 0.7458329287633103,
      "grad_norm": 1.6870904459979024,
      "learning_rate": 2.302604886718196e-07,
      "loss": 2.374,
      "step": 5761
    },
    {
      "epoch": 0.7459623911706638,
      "grad_norm": 1.1533928909331745,
      "learning_rate": 2.300415225394014e-07,
      "loss": 2.3413,
      "step": 5762
    },
    {
      "epoch": 0.7460918535780173,
      "grad_norm": 3.1282882856696035,
      "learning_rate": 2.298226370852582e-07,
      "loss": 2.4636,
      "step": 5763
    },
    {
      "epoch": 0.7462213159853708,
      "grad_norm": 1.563479304384251,
      "learning_rate": 2.2960383235410077e-07,
      "loss": 2.2686,
      "step": 5764
    },
    {
      "epoch": 0.7463507783927242,
      "grad_norm": 1.5971294582503306,
      "learning_rate": 2.2938510839062484e-07,
      "loss": 2.4404,
      "step": 5765
    },
    {
      "epoch": 0.7463507783927242,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5030628442764282,
      "eval_runtime": 14.8761,
      "eval_samples_per_second": 2.958,
      "eval_steps_per_second": 0.202,
      "step": 5765
    },
    {
      "epoch": 0.7464802408000777,
      "grad_norm": 3.6200775909620138,
      "learning_rate": 2.2916646523950952e-07,
      "loss": 2.6165,
      "step": 5766
    },
    {
      "epoch": 0.7466097032074311,
      "grad_norm": 1.502943780133985,
      "learning_rate": 2.2894790294541668e-07,
      "loss": 2.3403,
      "step": 5767
    },
    {
      "epoch": 0.7467391656147846,
      "grad_norm": 1.9252283045414476,
      "learning_rate": 2.2872942155299248e-07,
      "loss": 2.1272,
      "step": 5768
    },
    {
      "epoch": 0.746868628022138,
      "grad_norm": 1.8357036001414795,
      "learning_rate": 2.285110211068659e-07,
      "loss": 2.2083,
      "step": 5769
    },
    {
      "epoch": 0.7469980904294915,
      "grad_norm": 1.402098070812646,
      "learning_rate": 2.282927016516496e-07,
      "loss": 2.4163,
      "step": 5770
    },
    {
      "epoch": 0.7469980904294915,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.501420497894287,
      "eval_runtime": 15.5172,
      "eval_samples_per_second": 2.836,
      "eval_steps_per_second": 0.193,
      "step": 5770
    },
    {
      "epoch": 0.747127552836845,
      "grad_norm": 1.874322282527136,
      "learning_rate": 2.2807446323194003e-07,
      "loss": 2.4714,
      "step": 5771
    },
    {
      "epoch": 0.7472570152441985,
      "grad_norm": 1.714467438649989,
      "learning_rate": 2.2785630589231646e-07,
      "loss": 2.2888,
      "step": 5772
    },
    {
      "epoch": 0.747386477651552,
      "grad_norm": 1.3200850637029224,
      "learning_rate": 2.276382296773424e-07,
      "loss": 2.3416,
      "step": 5773
    },
    {
      "epoch": 0.7475159400589054,
      "grad_norm": 2.815673748855942,
      "learning_rate": 2.2742023463156405e-07,
      "loss": 2.4072,
      "step": 5774
    },
    {
      "epoch": 0.7476454024662589,
      "grad_norm": 1.3913837637007695,
      "learning_rate": 2.272023207995111e-07,
      "loss": 2.3845,
      "step": 5775
    },
    {
      "epoch": 0.7476454024662589,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5022194385528564,
      "eval_runtime": 15.2171,
      "eval_samples_per_second": 2.891,
      "eval_steps_per_second": 0.197,
      "step": 5775
    },
    {
      "epoch": 0.7477748648736123,
      "grad_norm": 2.1597954337338123,
      "learning_rate": 2.269844882256974e-07,
      "loss": 2.8374,
      "step": 5776
    },
    {
      "epoch": 0.7479043272809658,
      "grad_norm": 1.5402660364256493,
      "learning_rate": 2.2676673695461913e-07,
      "loss": 2.4219,
      "step": 5777
    },
    {
      "epoch": 0.7480337896883192,
      "grad_norm": 1.8661053171492508,
      "learning_rate": 2.2654906703075664e-07,
      "loss": 2.7046,
      "step": 5778
    },
    {
      "epoch": 0.7481632520956727,
      "grad_norm": 1.645395126294986,
      "learning_rate": 2.2633147849857395e-07,
      "loss": 2.2996,
      "step": 5779
    },
    {
      "epoch": 0.7482927145030261,
      "grad_norm": 2.2561182987384667,
      "learning_rate": 2.2611397140251692e-07,
      "loss": 2.4625,
      "step": 5780
    },
    {
      "epoch": 0.7482927145030261,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979802370071411,
      "eval_runtime": 13.7667,
      "eval_samples_per_second": 3.196,
      "eval_steps_per_second": 0.218,
      "step": 5780
    },
    {
      "epoch": 0.7484221769103796,
      "grad_norm": 1.3371922326166457,
      "learning_rate": 2.2589654578701653e-07,
      "loss": 2.2361,
      "step": 5781
    },
    {
      "epoch": 0.7485516393177332,
      "grad_norm": 1.9407566309567759,
      "learning_rate": 2.2567920169648587e-07,
      "loss": 2.3838,
      "step": 5782
    },
    {
      "epoch": 0.7486811017250866,
      "grad_norm": 1.83692195687607,
      "learning_rate": 2.254619391753222e-07,
      "loss": 2.4863,
      "step": 5783
    },
    {
      "epoch": 0.7488105641324401,
      "grad_norm": 4.340319107503327,
      "learning_rate": 2.2524475826790585e-07,
      "loss": 2.6577,
      "step": 5784
    },
    {
      "epoch": 0.7489400265397935,
      "grad_norm": 1.498664464152015,
      "learning_rate": 2.2502765901860034e-07,
      "loss": 2.3208,
      "step": 5785
    },
    {
      "epoch": 0.7489400265397935,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5020862817764282,
      "eval_runtime": 13.4527,
      "eval_samples_per_second": 3.271,
      "eval_steps_per_second": 0.223,
      "step": 5785
    },
    {
      "epoch": 0.749069488947147,
      "grad_norm": 1.669561872005248,
      "learning_rate": 2.2481064147175234e-07,
      "loss": 2.3477,
      "step": 5786
    },
    {
      "epoch": 0.7491989513545004,
      "grad_norm": 2.7942096472646614,
      "learning_rate": 2.2459370567169244e-07,
      "loss": 2.573,
      "step": 5787
    },
    {
      "epoch": 0.7493284137618539,
      "grad_norm": 1.8587148438346652,
      "learning_rate": 2.2437685166273384e-07,
      "loss": 2.3375,
      "step": 5788
    },
    {
      "epoch": 0.7494578761692073,
      "grad_norm": 1.4509462729135019,
      "learning_rate": 2.241600794891738e-07,
      "loss": 2.343,
      "step": 5789
    },
    {
      "epoch": 0.7495873385765608,
      "grad_norm": 1.6681557174623465,
      "learning_rate": 2.2394338919529199e-07,
      "loss": 2.3665,
      "step": 5790
    },
    {
      "epoch": 0.7495873385765608,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5028630495071411,
      "eval_runtime": 14.0959,
      "eval_samples_per_second": 3.121,
      "eval_steps_per_second": 0.213,
      "step": 5790
    },
    {
      "epoch": 0.7497168009839142,
      "grad_norm": 2.1816736864515254,
      "learning_rate": 2.237267808253521e-07,
      "loss": 2.3633,
      "step": 5791
    },
    {
      "epoch": 0.7498462633912678,
      "grad_norm": 1.6088210292697507,
      "learning_rate": 2.2351025442360072e-07,
      "loss": 2.2981,
      "step": 5792
    },
    {
      "epoch": 0.7499757257986213,
      "grad_norm": 2.842564527386317,
      "learning_rate": 2.232938100342675e-07,
      "loss": 2.5859,
      "step": 5793
    },
    {
      "epoch": 0.7501051882059747,
      "grad_norm": 1.6004002835668634,
      "learning_rate": 2.2307744770156598e-07,
      "loss": 2.2358,
      "step": 5794
    },
    {
      "epoch": 0.7502346506133282,
      "grad_norm": 2.332282601343465,
      "learning_rate": 2.2286116746969218e-07,
      "loss": 2.4663,
      "step": 5795
    },
    {
      "epoch": 0.7502346506133282,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4994007349014282,
      "eval_runtime": 13.8871,
      "eval_samples_per_second": 3.168,
      "eval_steps_per_second": 0.216,
      "step": 5795
    },
    {
      "epoch": 0.7503641130206816,
      "grad_norm": 1.563404749958465,
      "learning_rate": 2.2264496938282608e-07,
      "loss": 2.2307,
      "step": 5796
    },
    {
      "epoch": 0.7504935754280351,
      "grad_norm": 1.7218288638978925,
      "learning_rate": 2.2242885348513037e-07,
      "loss": 2.376,
      "step": 5797
    },
    {
      "epoch": 0.7506230378353885,
      "grad_norm": 4.1520240968988364,
      "learning_rate": 2.2221281982075085e-07,
      "loss": 2.6885,
      "step": 5798
    },
    {
      "epoch": 0.750752500242742,
      "grad_norm": 1.6439855781455703,
      "learning_rate": 2.2199686843381724e-07,
      "loss": 2.3845,
      "step": 5799
    },
    {
      "epoch": 0.7508819626500954,
      "grad_norm": 1.2343710915623454,
      "learning_rate": 2.2178099936844153e-07,
      "loss": 2.3928,
      "step": 5800
    },
    {
      "epoch": 0.7508819626500954,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.499467372894287,
      "eval_runtime": 13.9521,
      "eval_samples_per_second": 3.154,
      "eval_steps_per_second": 0.215,
      "step": 5800
    },
    {
      "epoch": 0.7510114250574489,
      "grad_norm": 2.4198345859304196,
      "learning_rate": 2.2156521266871963e-07,
      "loss": 2.4072,
      "step": 5801
    },
    {
      "epoch": 0.7511408874648025,
      "grad_norm": 1.5102624631034194,
      "learning_rate": 2.2134950837873066e-07,
      "loss": 2.4109,
      "step": 5802
    },
    {
      "epoch": 0.7512703498721559,
      "grad_norm": 1.4704684748471535,
      "learning_rate": 2.211338865425358e-07,
      "loss": 2.354,
      "step": 5803
    },
    {
      "epoch": 0.7513998122795094,
      "grad_norm": 1.3634893051480144,
      "learning_rate": 2.2091834720418085e-07,
      "loss": 2.4023,
      "step": 5804
    },
    {
      "epoch": 0.7515292746868628,
      "grad_norm": 1.7659414099753081,
      "learning_rate": 2.2070289040769362e-07,
      "loss": 2.4258,
      "step": 5805
    },
    {
      "epoch": 0.7515292746868628,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5008434057235718,
      "eval_runtime": 13.7516,
      "eval_samples_per_second": 3.2,
      "eval_steps_per_second": 0.218,
      "step": 5805
    },
    {
      "epoch": 0.7516587370942163,
      "grad_norm": 1.8938426151378294,
      "learning_rate": 2.204875161970857e-07,
      "loss": 2.532,
      "step": 5806
    },
    {
      "epoch": 0.7517881995015697,
      "grad_norm": 2.3418997079422073,
      "learning_rate": 2.2027222461635217e-07,
      "loss": 2.3567,
      "step": 5807
    },
    {
      "epoch": 0.7519176619089232,
      "grad_norm": 1.5563997643011263,
      "learning_rate": 2.2005701570946982e-07,
      "loss": 2.314,
      "step": 5808
    },
    {
      "epoch": 0.7520471243162766,
      "grad_norm": 1.7858463301155232,
      "learning_rate": 2.1984188952039977e-07,
      "loss": 2.5083,
      "step": 5809
    },
    {
      "epoch": 0.7521765867236301,
      "grad_norm": 1.1839448817651672,
      "learning_rate": 2.1962684609308623e-07,
      "loss": 2.2014,
      "step": 5810
    },
    {
      "epoch": 0.7521765867236301,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5029963254928589,
      "eval_runtime": 14.3925,
      "eval_samples_per_second": 3.057,
      "eval_steps_per_second": 0.208,
      "step": 5810
    },
    {
      "epoch": 0.7523060491309836,
      "grad_norm": 1.77223257260354,
      "learning_rate": 2.1941188547145567e-07,
      "loss": 2.448,
      "step": 5811
    },
    {
      "epoch": 0.7524355115383371,
      "grad_norm": 1.7572487436439015,
      "learning_rate": 2.1919700769941864e-07,
      "loss": 2.4866,
      "step": 5812
    },
    {
      "epoch": 0.7525649739456906,
      "grad_norm": 1.4882126206999426,
      "learning_rate": 2.1898221282086802e-07,
      "loss": 2.459,
      "step": 5813
    },
    {
      "epoch": 0.752694436353044,
      "grad_norm": 2.0967141929893303,
      "learning_rate": 2.1876750087967982e-07,
      "loss": 2.3584,
      "step": 5814
    },
    {
      "epoch": 0.7528238987603975,
      "grad_norm": 2.580482483314053,
      "learning_rate": 2.185528719197138e-07,
      "loss": 2.5562,
      "step": 5815
    },
    {
      "epoch": 0.7528238987603975,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500199794769287,
      "eval_runtime": 14.1446,
      "eval_samples_per_second": 3.111,
      "eval_steps_per_second": 0.212,
      "step": 5815
    },
    {
      "epoch": 0.7529533611677509,
      "grad_norm": 1.959816140512178,
      "learning_rate": 2.1833832598481166e-07,
      "loss": 2.3865,
      "step": 5816
    },
    {
      "epoch": 0.7530828235751044,
      "grad_norm": 1.9862527337467335,
      "learning_rate": 2.1812386311879942e-07,
      "loss": 2.3065,
      "step": 5817
    },
    {
      "epoch": 0.7532122859824578,
      "grad_norm": 2.4568435787390857,
      "learning_rate": 2.179094833654851e-07,
      "loss": 2.5728,
      "step": 5818
    },
    {
      "epoch": 0.7533417483898113,
      "grad_norm": 2.9824604578507317,
      "learning_rate": 2.176951867686599e-07,
      "loss": 2.3652,
      "step": 5819
    },
    {
      "epoch": 0.7534712107971647,
      "grad_norm": 2.854202729042815,
      "learning_rate": 2.1748097337209874e-07,
      "loss": 2.5706,
      "step": 5820
    },
    {
      "epoch": 0.7534712107971647,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5008655786514282,
      "eval_runtime": 14.4718,
      "eval_samples_per_second": 3.04,
      "eval_steps_per_second": 0.207,
      "step": 5820
    },
    {
      "epoch": 0.7536006732045183,
      "grad_norm": 1.5490050823786856,
      "learning_rate": 2.1726684321955864e-07,
      "loss": 2.4033,
      "step": 5821
    },
    {
      "epoch": 0.7537301356118717,
      "grad_norm": 1.7718613400586933,
      "learning_rate": 2.1705279635478034e-07,
      "loss": 2.6299,
      "step": 5822
    },
    {
      "epoch": 0.7538595980192252,
      "grad_norm": 2.6252808455848338,
      "learning_rate": 2.1683883282148688e-07,
      "loss": 2.4365,
      "step": 5823
    },
    {
      "epoch": 0.7539890604265787,
      "grad_norm": 1.5580361297154999,
      "learning_rate": 2.1662495266338512e-07,
      "loss": 2.3257,
      "step": 5824
    },
    {
      "epoch": 0.7541185228339321,
      "grad_norm": 2.0954953948086246,
      "learning_rate": 2.164111559241641e-07,
      "loss": 2.2993,
      "step": 5825
    },
    {
      "epoch": 0.7541185228339321,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5009765625,
      "eval_runtime": 13.8884,
      "eval_samples_per_second": 3.168,
      "eval_steps_per_second": 0.216,
      "step": 5825
    },
    {
      "epoch": 0.7542479852412856,
      "grad_norm": 1.8732852881099835,
      "learning_rate": 2.161974426474961e-07,
      "loss": 2.3396,
      "step": 5826
    },
    {
      "epoch": 0.754377447648639,
      "grad_norm": 1.6494606523117687,
      "learning_rate": 2.1598381287703662e-07,
      "loss": 2.4102,
      "step": 5827
    },
    {
      "epoch": 0.7545069100559925,
      "grad_norm": 2.183039698870735,
      "learning_rate": 2.1577026665642355e-07,
      "loss": 2.322,
      "step": 5828
    },
    {
      "epoch": 0.7546363724633459,
      "grad_norm": 1.3704214102627932,
      "learning_rate": 2.1555680402927818e-07,
      "loss": 2.3706,
      "step": 5829
    },
    {
      "epoch": 0.7547658348706994,
      "grad_norm": 1.5718198086108035,
      "learning_rate": 2.1534342503920508e-07,
      "loss": 2.3579,
      "step": 5830
    },
    {
      "epoch": 0.7547658348706994,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5030184984207153,
      "eval_runtime": 13.3784,
      "eval_samples_per_second": 3.289,
      "eval_steps_per_second": 0.224,
      "step": 5830
    },
    {
      "epoch": 0.7548952972780529,
      "grad_norm": 1.2953292597988508,
      "learning_rate": 2.1513012972979037e-07,
      "loss": 2.2964,
      "step": 5831
    },
    {
      "epoch": 0.7550247596854064,
      "grad_norm": 2.8143790455089253,
      "learning_rate": 2.149169181446043e-07,
      "loss": 2.5725,
      "step": 5832
    },
    {
      "epoch": 0.7551542220927598,
      "grad_norm": 1.561761962705944,
      "learning_rate": 2.147037903272e-07,
      "loss": 2.4038,
      "step": 5833
    },
    {
      "epoch": 0.7552836845001133,
      "grad_norm": 1.5135770535139144,
      "learning_rate": 2.1449074632111254e-07,
      "loss": 2.4106,
      "step": 5834
    },
    {
      "epoch": 0.7554131469074667,
      "grad_norm": 1.9510422997681827,
      "learning_rate": 2.1427778616986126e-07,
      "loss": 2.3423,
      "step": 5835
    },
    {
      "epoch": 0.7554131469074667,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498002529144287,
      "eval_runtime": 14.3633,
      "eval_samples_per_second": 3.063,
      "eval_steps_per_second": 0.209,
      "step": 5835
    },
    {
      "epoch": 0.7555426093148202,
      "grad_norm": 2.8136067249278183,
      "learning_rate": 2.1406490991694672e-07,
      "loss": 2.5547,
      "step": 5836
    },
    {
      "epoch": 0.7556720717221737,
      "grad_norm": 2.2518835814522515,
      "learning_rate": 2.1385211760585355e-07,
      "loss": 2.3887,
      "step": 5837
    },
    {
      "epoch": 0.7558015341295271,
      "grad_norm": 3.1990213138561368,
      "learning_rate": 2.1363940928004916e-07,
      "loss": 2.5039,
      "step": 5838
    },
    {
      "epoch": 0.7559309965368806,
      "grad_norm": 1.9339674321589873,
      "learning_rate": 2.1342678498298317e-07,
      "loss": 2.3853,
      "step": 5839
    },
    {
      "epoch": 0.756060458944234,
      "grad_norm": 3.0297434106965477,
      "learning_rate": 2.1321424475808872e-07,
      "loss": 2.5613,
      "step": 5840
    },
    {
      "epoch": 0.756060458944234,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986239671707153,
      "eval_runtime": 13.5902,
      "eval_samples_per_second": 3.238,
      "eval_steps_per_second": 0.221,
      "step": 5840
    },
    {
      "epoch": 0.7561899213515876,
      "grad_norm": 1.4276704769685569,
      "learning_rate": 2.1300178864878133e-07,
      "loss": 2.4243,
      "step": 5841
    },
    {
      "epoch": 0.756319383758941,
      "grad_norm": 1.9799590651593946,
      "learning_rate": 2.127894166984592e-07,
      "loss": 2.3115,
      "step": 5842
    },
    {
      "epoch": 0.7564488461662945,
      "grad_norm": 1.7754285516238946,
      "learning_rate": 2.12577128950504e-07,
      "loss": 2.5503,
      "step": 5843
    },
    {
      "epoch": 0.7565783085736479,
      "grad_norm": 2.239658834594748,
      "learning_rate": 2.123649254482795e-07,
      "loss": 2.3618,
      "step": 5844
    },
    {
      "epoch": 0.7567077709810014,
      "grad_norm": 1.836751532298756,
      "learning_rate": 2.1215280623513287e-07,
      "loss": 2.3984,
      "step": 5845
    },
    {
      "epoch": 0.7567077709810014,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5001331567764282,
      "eval_runtime": 14.1511,
      "eval_samples_per_second": 3.109,
      "eval_steps_per_second": 0.212,
      "step": 5845
    },
    {
      "epoch": 0.7568372333883548,
      "grad_norm": 1.5455753907625613,
      "learning_rate": 2.119407713543936e-07,
      "loss": 2.4397,
      "step": 5846
    },
    {
      "epoch": 0.7569666957957083,
      "grad_norm": 1.9368709358290113,
      "learning_rate": 2.1172882084937397e-07,
      "loss": 2.2344,
      "step": 5847
    },
    {
      "epoch": 0.7570961582030618,
      "grad_norm": 2.436297641108228,
      "learning_rate": 2.1151695476336948e-07,
      "loss": 2.3818,
      "step": 5848
    },
    {
      "epoch": 0.7572256206104152,
      "grad_norm": 2.5561262580328545,
      "learning_rate": 2.1130517313965767e-07,
      "loss": 2.7991,
      "step": 5849
    },
    {
      "epoch": 0.7573550830177687,
      "grad_norm": 1.810400595822349,
      "learning_rate": 2.110934760214997e-07,
      "loss": 2.4092,
      "step": 5850
    },
    {
      "epoch": 0.7573550830177687,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4994450807571411,
      "eval_runtime": 16.0322,
      "eval_samples_per_second": 2.744,
      "eval_steps_per_second": 0.187,
      "step": 5850
    },
    {
      "epoch": 0.7574845454251222,
      "grad_norm": 1.4244126918376885,
      "learning_rate": 2.1088186345213846e-07,
      "loss": 2.123,
      "step": 5851
    },
    {
      "epoch": 0.7576140078324757,
      "grad_norm": 1.862321548250587,
      "learning_rate": 2.1067033547480068e-07,
      "loss": 2.5386,
      "step": 5852
    },
    {
      "epoch": 0.7577434702398291,
      "grad_norm": 1.269591250880515,
      "learning_rate": 2.10458892132695e-07,
      "loss": 2.1313,
      "step": 5853
    },
    {
      "epoch": 0.7578729326471826,
      "grad_norm": 1.8516166311990983,
      "learning_rate": 2.1024753346901272e-07,
      "loss": 2.3911,
      "step": 5854
    },
    {
      "epoch": 0.758002395054536,
      "grad_norm": 1.8588250908961133,
      "learning_rate": 2.1003625952692835e-07,
      "loss": 2.4712,
      "step": 5855
    },
    {
      "epoch": 0.758002395054536,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500044345855713,
      "eval_runtime": 15.077,
      "eval_samples_per_second": 2.918,
      "eval_steps_per_second": 0.199,
      "step": 5855
    },
    {
      "epoch": 0.7581318574618895,
      "grad_norm": 1.9009920870396213,
      "learning_rate": 2.0982507034959917e-07,
      "loss": 2.5229,
      "step": 5856
    },
    {
      "epoch": 0.758261319869243,
      "grad_norm": 1.1134535998530464,
      "learning_rate": 2.096139659801645e-07,
      "loss": 2.1763,
      "step": 5857
    },
    {
      "epoch": 0.7583907822765964,
      "grad_norm": 2.1086126757773753,
      "learning_rate": 2.0940294646174683e-07,
      "loss": 2.3394,
      "step": 5858
    },
    {
      "epoch": 0.7585202446839499,
      "grad_norm": 1.486306292850415,
      "learning_rate": 2.0919201183745089e-07,
      "loss": 2.3103,
      "step": 5859
    },
    {
      "epoch": 0.7586497070913034,
      "grad_norm": 1.2114077432887764,
      "learning_rate": 2.0898116215036456e-07,
      "loss": 2.2319,
      "step": 5860
    },
    {
      "epoch": 0.7586497070913034,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4993563890457153,
      "eval_runtime": 14.2575,
      "eval_samples_per_second": 3.086,
      "eval_steps_per_second": 0.21,
      "step": 5860
    },
    {
      "epoch": 0.7587791694986569,
      "grad_norm": 1.2559508015931904,
      "learning_rate": 2.0877039744355843e-07,
      "loss": 2.1653,
      "step": 5861
    },
    {
      "epoch": 0.7589086319060103,
      "grad_norm": 2.1189008374382827,
      "learning_rate": 2.0855971776008495e-07,
      "loss": 2.4387,
      "step": 5862
    },
    {
      "epoch": 0.7590380943133638,
      "grad_norm": 2.3990482743751795,
      "learning_rate": 2.0834912314298025e-07,
      "loss": 2.4204,
      "step": 5863
    },
    {
      "epoch": 0.7591675567207172,
      "grad_norm": 2.327636679954212,
      "learning_rate": 2.081386136352622e-07,
      "loss": 2.4438,
      "step": 5864
    },
    {
      "epoch": 0.7592970191280707,
      "grad_norm": 1.391628372715472,
      "learning_rate": 2.079281892799316e-07,
      "loss": 2.3311,
      "step": 5865
    },
    {
      "epoch": 0.7592970191280707,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.496537685394287,
      "eval_runtime": 13.8112,
      "eval_samples_per_second": 3.186,
      "eval_steps_per_second": 0.217,
      "step": 5865
    },
    {
      "epoch": 0.7594264815354241,
      "grad_norm": 1.3662541680668019,
      "learning_rate": 2.0771785011997212e-07,
      "loss": 2.1653,
      "step": 5866
    },
    {
      "epoch": 0.7595559439427776,
      "grad_norm": 1.70273889789075,
      "learning_rate": 2.075075961983495e-07,
      "loss": 2.4429,
      "step": 5867
    },
    {
      "epoch": 0.759685406350131,
      "grad_norm": 2.3417248361813003,
      "learning_rate": 2.072974275580128e-07,
      "loss": 2.3477,
      "step": 5868
    },
    {
      "epoch": 0.7598148687574845,
      "grad_norm": 1.4082598294894912,
      "learning_rate": 2.070873442418929e-07,
      "loss": 2.3438,
      "step": 5869
    },
    {
      "epoch": 0.7599443311648381,
      "grad_norm": 1.4615806857028715,
      "learning_rate": 2.0687734629290343e-07,
      "loss": 2.2566,
      "step": 5870
    },
    {
      "epoch": 0.7599443311648381,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4974254369735718,
      "eval_runtime": 14.9405,
      "eval_samples_per_second": 2.945,
      "eval_steps_per_second": 0.201,
      "step": 5870
    },
    {
      "epoch": 0.7600737935721915,
      "grad_norm": 1.2953449127370287,
      "learning_rate": 2.0666743375394116e-07,
      "loss": 2.4097,
      "step": 5871
    },
    {
      "epoch": 0.760203255979545,
      "grad_norm": 1.3745094758828758,
      "learning_rate": 2.064576066678846e-07,
      "loss": 2.2834,
      "step": 5872
    },
    {
      "epoch": 0.7603327183868984,
      "grad_norm": 1.9308242666352429,
      "learning_rate": 2.062478650775953e-07,
      "loss": 2.3555,
      "step": 5873
    },
    {
      "epoch": 0.7604621807942519,
      "grad_norm": 3.9295688564422178,
      "learning_rate": 2.060382090259177e-07,
      "loss": 2.3154,
      "step": 5874
    },
    {
      "epoch": 0.7605916432016053,
      "grad_norm": 1.414342744272351,
      "learning_rate": 2.0582863855567743e-07,
      "loss": 2.4702,
      "step": 5875
    },
    {
      "epoch": 0.7605916432016053,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5001331567764282,
      "eval_runtime": 13.5483,
      "eval_samples_per_second": 3.248,
      "eval_steps_per_second": 0.221,
      "step": 5875
    },
    {
      "epoch": 0.7607211056089588,
      "grad_norm": 1.7593476608674339,
      "learning_rate": 2.0561915370968423e-07,
      "loss": 2.3765,
      "step": 5876
    },
    {
      "epoch": 0.7608505680163122,
      "grad_norm": 1.3573945972191568,
      "learning_rate": 2.054097545307291e-07,
      "loss": 2.3594,
      "step": 5877
    },
    {
      "epoch": 0.7609800304236657,
      "grad_norm": 1.4600431466486712,
      "learning_rate": 2.0520044106158625e-07,
      "loss": 2.4426,
      "step": 5878
    },
    {
      "epoch": 0.7611094928310191,
      "grad_norm": 2.2378712758784816,
      "learning_rate": 2.0499121334501258e-07,
      "loss": 2.4822,
      "step": 5879
    },
    {
      "epoch": 0.7612389552383727,
      "grad_norm": 2.4570825836503136,
      "learning_rate": 2.0478207142374674e-07,
      "loss": 2.3691,
      "step": 5880
    },
    {
      "epoch": 0.7612389552383727,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4966042041778564,
      "eval_runtime": 14.0528,
      "eval_samples_per_second": 3.131,
      "eval_steps_per_second": 0.213,
      "step": 5880
    },
    {
      "epoch": 0.7613684176457262,
      "grad_norm": 1.7874177609679103,
      "learning_rate": 2.0457301534051004e-07,
      "loss": 2.416,
      "step": 5881
    },
    {
      "epoch": 0.7614978800530796,
      "grad_norm": 1.5142164972034926,
      "learning_rate": 2.0436404513800675e-07,
      "loss": 2.6133,
      "step": 5882
    },
    {
      "epoch": 0.7616273424604331,
      "grad_norm": 1.8254061667671044,
      "learning_rate": 2.0415516085892303e-07,
      "loss": 2.5039,
      "step": 5883
    },
    {
      "epoch": 0.7617568048677865,
      "grad_norm": 1.572149682501575,
      "learning_rate": 2.0394636254592802e-07,
      "loss": 2.4492,
      "step": 5884
    },
    {
      "epoch": 0.76188626727514,
      "grad_norm": 1.5063564419177435,
      "learning_rate": 2.0373765024167284e-07,
      "loss": 2.6472,
      "step": 5885
    },
    {
      "epoch": 0.76188626727514,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5008877515792847,
      "eval_runtime": 13.8895,
      "eval_samples_per_second": 3.168,
      "eval_steps_per_second": 0.216,
      "step": 5885
    },
    {
      "epoch": 0.7620157296824934,
      "grad_norm": 2.1321630772537,
      "learning_rate": 2.0352902398879098e-07,
      "loss": 2.4194,
      "step": 5886
    },
    {
      "epoch": 0.7621451920898469,
      "grad_norm": 1.846813322395439,
      "learning_rate": 2.0332048382989895e-07,
      "loss": 2.2397,
      "step": 5887
    },
    {
      "epoch": 0.7622746544972003,
      "grad_norm": 4.134731739638366,
      "learning_rate": 2.0311202980759504e-07,
      "loss": 2.6807,
      "step": 5888
    },
    {
      "epoch": 0.7624041169045538,
      "grad_norm": 3.5200971486495347,
      "learning_rate": 2.0290366196446045e-07,
      "loss": 2.8784,
      "step": 5889
    },
    {
      "epoch": 0.7625335793119074,
      "grad_norm": 1.393468781608451,
      "learning_rate": 2.0269538034305827e-07,
      "loss": 2.3853,
      "step": 5890
    },
    {
      "epoch": 0.7625335793119074,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500199794769287,
      "eval_runtime": 14.4333,
      "eval_samples_per_second": 3.048,
      "eval_steps_per_second": 0.208,
      "step": 5890
    },
    {
      "epoch": 0.7626630417192608,
      "grad_norm": 4.058911303706695,
      "learning_rate": 2.0248718498593453e-07,
      "loss": 2.9038,
      "step": 5891
    },
    {
      "epoch": 0.7627925041266143,
      "grad_norm": 1.9489576287446886,
      "learning_rate": 2.0227907593561718e-07,
      "loss": 2.6946,
      "step": 5892
    },
    {
      "epoch": 0.7629219665339677,
      "grad_norm": 1.2233543433416605,
      "learning_rate": 2.0207105323461658e-07,
      "loss": 2.3494,
      "step": 5893
    },
    {
      "epoch": 0.7630514289413212,
      "grad_norm": 2.7917420980757366,
      "learning_rate": 2.0186311692542594e-07,
      "loss": 2.3591,
      "step": 5894
    },
    {
      "epoch": 0.7631808913486746,
      "grad_norm": 4.952772839611828,
      "learning_rate": 2.0165526705052007e-07,
      "loss": 2.3567,
      "step": 5895
    },
    {
      "epoch": 0.7631808913486746,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5015536546707153,
      "eval_runtime": 14.1098,
      "eval_samples_per_second": 3.118,
      "eval_steps_per_second": 0.213,
      "step": 5895
    },
    {
      "epoch": 0.7633103537560281,
      "grad_norm": 2.5469536401721076,
      "learning_rate": 2.0144750365235687e-07,
      "loss": 2.5813,
      "step": 5896
    },
    {
      "epoch": 0.7634398161633815,
      "grad_norm": 1.4416734309857384,
      "learning_rate": 2.012398267733761e-07,
      "loss": 2.3984,
      "step": 5897
    },
    {
      "epoch": 0.763569278570735,
      "grad_norm": 3.5685517982616104,
      "learning_rate": 2.0103223645599974e-07,
      "loss": 2.9966,
      "step": 5898
    },
    {
      "epoch": 0.7636987409780885,
      "grad_norm": 2.833785840312419,
      "learning_rate": 2.008247327426327e-07,
      "loss": 2.3867,
      "step": 5899
    },
    {
      "epoch": 0.763828203385442,
      "grad_norm": 2.6422615350344825,
      "learning_rate": 2.006173156756614e-07,
      "loss": 2.6487,
      "step": 5900
    },
    {
      "epoch": 0.763828203385442,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500044345855713,
      "eval_runtime": 13.8866,
      "eval_samples_per_second": 3.169,
      "eval_steps_per_second": 0.216,
      "step": 5900
    },
    {
      "epoch": 0.7639576657927954,
      "grad_norm": 3.167058360094262,
      "learning_rate": 2.0040998529745532e-07,
      "loss": 2.4365,
      "step": 5901
    },
    {
      "epoch": 0.7640871282001489,
      "grad_norm": 1.5622768032380048,
      "learning_rate": 2.0020274165036608e-07,
      "loss": 2.3027,
      "step": 5902
    },
    {
      "epoch": 0.7642165906075024,
      "grad_norm": 1.5589453517866538,
      "learning_rate": 1.9999558477672674e-07,
      "loss": 2.1697,
      "step": 5903
    },
    {
      "epoch": 0.7643460530148558,
      "grad_norm": 1.5635110718355771,
      "learning_rate": 1.9978851471885373e-07,
      "loss": 2.4434,
      "step": 5904
    },
    {
      "epoch": 0.7644755154222093,
      "grad_norm": 3.0729043453070823,
      "learning_rate": 1.995815315190454e-07,
      "loss": 3.3315,
      "step": 5905
    },
    {
      "epoch": 0.7644755154222093,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986904859542847,
      "eval_runtime": 14.7505,
      "eval_samples_per_second": 2.983,
      "eval_steps_per_second": 0.203,
      "step": 5905
    },
    {
      "epoch": 0.7646049778295627,
      "grad_norm": 1.8095601699611552,
      "learning_rate": 1.9937463521958183e-07,
      "loss": 2.2881,
      "step": 5906
    },
    {
      "epoch": 0.7647344402369162,
      "grad_norm": 1.8593368639132177,
      "learning_rate": 1.991678258627265e-07,
      "loss": 2.3794,
      "step": 5907
    },
    {
      "epoch": 0.7648639026442696,
      "grad_norm": 2.770245655129175,
      "learning_rate": 1.9896110349072357e-07,
      "loss": 2.4385,
      "step": 5908
    },
    {
      "epoch": 0.7649933650516232,
      "grad_norm": 1.3787170669825033,
      "learning_rate": 1.9875446814580064e-07,
      "loss": 2.2878,
      "step": 5909
    },
    {
      "epoch": 0.7651228274589766,
      "grad_norm": 2.369476370607867,
      "learning_rate": 1.9854791987016739e-07,
      "loss": 2.459,
      "step": 5910
    },
    {
      "epoch": 0.7651228274589766,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498734951019287,
      "eval_runtime": 13.8554,
      "eval_samples_per_second": 3.176,
      "eval_steps_per_second": 0.217,
      "step": 5910
    },
    {
      "epoch": 0.7652522898663301,
      "grad_norm": 1.2830844503907153,
      "learning_rate": 1.9834145870601513e-07,
      "loss": 2.2778,
      "step": 5911
    },
    {
      "epoch": 0.7653817522736835,
      "grad_norm": 2.5302739216629244,
      "learning_rate": 1.9813508469551808e-07,
      "loss": 2.5352,
      "step": 5912
    },
    {
      "epoch": 0.765511214681037,
      "grad_norm": 1.5276455432435188,
      "learning_rate": 1.9792879788083219e-07,
      "loss": 2.2871,
      "step": 5913
    },
    {
      "epoch": 0.7656406770883905,
      "grad_norm": 3.3943950562170246,
      "learning_rate": 1.9772259830409542e-07,
      "loss": 2.9097,
      "step": 5914
    },
    {
      "epoch": 0.7657701394957439,
      "grad_norm": 1.4990112841006813,
      "learning_rate": 1.9751648600742863e-07,
      "loss": 2.5159,
      "step": 5915
    },
    {
      "epoch": 0.7657701394957439,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.502241611480713,
      "eval_runtime": 14.3836,
      "eval_samples_per_second": 3.059,
      "eval_steps_per_second": 0.209,
      "step": 5915
    },
    {
      "epoch": 0.7658996019030974,
      "grad_norm": 1.0252065395231953,
      "learning_rate": 1.973104610329341e-07,
      "loss": 2.2339,
      "step": 5916
    },
    {
      "epoch": 0.7660290643104508,
      "grad_norm": 1.561934321130518,
      "learning_rate": 1.9710452342269703e-07,
      "loss": 2.271,
      "step": 5917
    },
    {
      "epoch": 0.7661585267178043,
      "grad_norm": 2.3183876245973787,
      "learning_rate": 1.9689867321878387e-07,
      "loss": 2.3923,
      "step": 5918
    },
    {
      "epoch": 0.7662879891251578,
      "grad_norm": 1.4481714440717672,
      "learning_rate": 1.966929104632441e-07,
      "loss": 2.3518,
      "step": 5919
    },
    {
      "epoch": 0.7664174515325113,
      "grad_norm": 1.7657229319959258,
      "learning_rate": 1.9648723519810876e-07,
      "loss": 2.3081,
      "step": 5920
    },
    {
      "epoch": 0.7664174515325113,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5001775026321411,
      "eval_runtime": 14.8734,
      "eval_samples_per_second": 2.958,
      "eval_steps_per_second": 0.202,
      "step": 5920
    },
    {
      "epoch": 0.7665469139398647,
      "grad_norm": 3.0911986408405423,
      "learning_rate": 1.96281647465391e-07,
      "loss": 2.2639,
      "step": 5921
    },
    {
      "epoch": 0.7666763763472182,
      "grad_norm": 2.1304295172481202,
      "learning_rate": 1.9607614730708665e-07,
      "loss": 2.3624,
      "step": 5922
    },
    {
      "epoch": 0.7668058387545716,
      "grad_norm": 1.3855225283903845,
      "learning_rate": 1.9587073476517289e-07,
      "loss": 2.3735,
      "step": 5923
    },
    {
      "epoch": 0.7669353011619251,
      "grad_norm": 2.984396702597027,
      "learning_rate": 1.9566540988160986e-07,
      "loss": 2.2583,
      "step": 5924
    },
    {
      "epoch": 0.7670647635692786,
      "grad_norm": 1.7741862050664121,
      "learning_rate": 1.9546017269833906e-07,
      "loss": 2.478,
      "step": 5925
    },
    {
      "epoch": 0.7670647635692786,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.502152919769287,
      "eval_runtime": 13.805,
      "eval_samples_per_second": 3.187,
      "eval_steps_per_second": 0.217,
      "step": 5925
    },
    {
      "epoch": 0.767194225976632,
      "grad_norm": 2.101543329687786,
      "learning_rate": 1.9525502325728418e-07,
      "loss": 2.3989,
      "step": 5926
    },
    {
      "epoch": 0.7673236883839855,
      "grad_norm": 1.4956955927310565,
      "learning_rate": 1.9504996160035128e-07,
      "loss": 2.3948,
      "step": 5927
    },
    {
      "epoch": 0.7674531507913389,
      "grad_norm": 1.9052839059065405,
      "learning_rate": 1.948449877694286e-07,
      "loss": 2.5186,
      "step": 5928
    },
    {
      "epoch": 0.7675826131986925,
      "grad_norm": 2.6304563370898584,
      "learning_rate": 1.9464010180638592e-07,
      "loss": 2.2791,
      "step": 5929
    },
    {
      "epoch": 0.7677120756060459,
      "grad_norm": 1.1883510455228177,
      "learning_rate": 1.9443530375307579e-07,
      "loss": 2.2644,
      "step": 5930
    },
    {
      "epoch": 0.7677120756060459,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5002219676971436,
      "eval_runtime": 14.4103,
      "eval_samples_per_second": 3.053,
      "eval_steps_per_second": 0.208,
      "step": 5930
    },
    {
      "epoch": 0.7678415380133994,
      "grad_norm": 1.6880308486218178,
      "learning_rate": 1.9423059365133162e-07,
      "loss": 2.4407,
      "step": 5931
    },
    {
      "epoch": 0.7679710004207528,
      "grad_norm": 2.766901150411268,
      "learning_rate": 1.9402597154296995e-07,
      "loss": 2.5522,
      "step": 5932
    },
    {
      "epoch": 0.7681004628281063,
      "grad_norm": 1.5873555048956454,
      "learning_rate": 1.938214374697893e-07,
      "loss": 2.415,
      "step": 5933
    },
    {
      "epoch": 0.7682299252354597,
      "grad_norm": 2.7049780084981565,
      "learning_rate": 1.9361699147356937e-07,
      "loss": 2.2012,
      "step": 5934
    },
    {
      "epoch": 0.7683593876428132,
      "grad_norm": 1.8552129165576046,
      "learning_rate": 1.9341263359607295e-07,
      "loss": 2.4868,
      "step": 5935
    },
    {
      "epoch": 0.7683593876428132,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500932216644287,
      "eval_runtime": 14.2395,
      "eval_samples_per_second": 3.09,
      "eval_steps_per_second": 0.211,
      "step": 5935
    },
    {
      "epoch": 0.7684888500501666,
      "grad_norm": 1.6354437818331267,
      "learning_rate": 1.9320836387904392e-07,
      "loss": 2.3818,
      "step": 5936
    },
    {
      "epoch": 0.7686183124575201,
      "grad_norm": 2.380509089967578,
      "learning_rate": 1.9300418236420842e-07,
      "loss": 2.4316,
      "step": 5937
    },
    {
      "epoch": 0.7687477748648737,
      "grad_norm": 1.6673693466640416,
      "learning_rate": 1.92800089093275e-07,
      "loss": 2.5042,
      "step": 5938
    },
    {
      "epoch": 0.7688772372722271,
      "grad_norm": 1.354707275436024,
      "learning_rate": 1.925960841079335e-07,
      "loss": 2.4111,
      "step": 5939
    },
    {
      "epoch": 0.7690066996795806,
      "grad_norm": 1.884803974676278,
      "learning_rate": 1.923921674498565e-07,
      "loss": 2.5173,
      "step": 5940
    },
    {
      "epoch": 0.7690066996795806,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4966042041778564,
      "eval_runtime": 16.4031,
      "eval_samples_per_second": 2.682,
      "eval_steps_per_second": 0.183,
      "step": 5940
    },
    {
      "epoch": 0.769136162086934,
      "grad_norm": 1.4961730243630353,
      "learning_rate": 1.9218833916069775e-07,
      "loss": 2.2083,
      "step": 5941
    },
    {
      "epoch": 0.7692656244942875,
      "grad_norm": 2.0095951494477857,
      "learning_rate": 1.9198459928209332e-07,
      "loss": 2.1587,
      "step": 5942
    },
    {
      "epoch": 0.7693950869016409,
      "grad_norm": 1.6736478695801495,
      "learning_rate": 1.9178094785566142e-07,
      "loss": 2.3335,
      "step": 5943
    },
    {
      "epoch": 0.7695245493089944,
      "grad_norm": 2.192720257106963,
      "learning_rate": 1.9157738492300176e-07,
      "loss": 2.2224,
      "step": 5944
    },
    {
      "epoch": 0.7696540117163478,
      "grad_norm": 2.36502245282972,
      "learning_rate": 1.913739105256964e-07,
      "loss": 2.3296,
      "step": 5945
    },
    {
      "epoch": 0.7696540117163478,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.503706455230713,
      "eval_runtime": 13.9897,
      "eval_samples_per_second": 3.145,
      "eval_steps_per_second": 0.214,
      "step": 5945
    },
    {
      "epoch": 0.7697834741237013,
      "grad_norm": 1.4459858726170736,
      "learning_rate": 1.911705247053091e-07,
      "loss": 2.4153,
      "step": 5946
    },
    {
      "epoch": 0.7699129365310547,
      "grad_norm": 1.9920734193437843,
      "learning_rate": 1.9096722750338524e-07,
      "loss": 2.5237,
      "step": 5947
    },
    {
      "epoch": 0.7700423989384083,
      "grad_norm": 1.6284989520443458,
      "learning_rate": 1.9076401896145284e-07,
      "loss": 2.3054,
      "step": 5948
    },
    {
      "epoch": 0.7701718613457618,
      "grad_norm": 2.3742044914808735,
      "learning_rate": 1.9056089912102088e-07,
      "loss": 2.6118,
      "step": 5949
    },
    {
      "epoch": 0.7703013237531152,
      "grad_norm": 2.4977493282187706,
      "learning_rate": 1.903578680235809e-07,
      "loss": 2.3774,
      "step": 5950
    },
    {
      "epoch": 0.7703013237531152,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5000221729278564,
      "eval_runtime": 14.7863,
      "eval_samples_per_second": 2.976,
      "eval_steps_per_second": 0.203,
      "step": 5950
    },
    {
      "epoch": 0.7704307861604687,
      "grad_norm": 1.4355870507424202,
      "learning_rate": 1.9015492571060648e-07,
      "loss": 2.2729,
      "step": 5951
    },
    {
      "epoch": 0.7705602485678221,
      "grad_norm": 2.0544240421967004,
      "learning_rate": 1.8995207222355233e-07,
      "loss": 2.5181,
      "step": 5952
    },
    {
      "epoch": 0.7706897109751756,
      "grad_norm": 1.9853028041490168,
      "learning_rate": 1.897493076038555e-07,
      "loss": 2.438,
      "step": 5953
    },
    {
      "epoch": 0.770819173382529,
      "grad_norm": 1.523101933486617,
      "learning_rate": 1.8954663189293446e-07,
      "loss": 2.1951,
      "step": 5954
    },
    {
      "epoch": 0.7709486357898825,
      "grad_norm": 1.585523135004017,
      "learning_rate": 1.8934404513219011e-07,
      "loss": 2.4333,
      "step": 5955
    },
    {
      "epoch": 0.7709486357898825,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978915452957153,
      "eval_runtime": 14.4892,
      "eval_samples_per_second": 3.037,
      "eval_steps_per_second": 0.207,
      "step": 5955
    },
    {
      "epoch": 0.7710780981972359,
      "grad_norm": 1.5669963664772997,
      "learning_rate": 1.8914154736300511e-07,
      "loss": 2.5242,
      "step": 5956
    },
    {
      "epoch": 0.7712075606045894,
      "grad_norm": 1.8538204540122294,
      "learning_rate": 1.8893913862674328e-07,
      "loss": 2.4797,
      "step": 5957
    },
    {
      "epoch": 0.771337023011943,
      "grad_norm": 2.4622943772964843,
      "learning_rate": 1.8873681896475113e-07,
      "loss": 2.2961,
      "step": 5958
    },
    {
      "epoch": 0.7714664854192964,
      "grad_norm": 1.3910046817559374,
      "learning_rate": 1.8853458841835635e-07,
      "loss": 2.4067,
      "step": 5959
    },
    {
      "epoch": 0.7715959478266499,
      "grad_norm": 1.6375854849941913,
      "learning_rate": 1.8833244702886847e-07,
      "loss": 2.3096,
      "step": 5960
    },
    {
      "epoch": 0.7715959478266499,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500932216644287,
      "eval_runtime": 14.433,
      "eval_samples_per_second": 3.049,
      "eval_steps_per_second": 0.208,
      "step": 5960
    },
    {
      "epoch": 0.7717254102340033,
      "grad_norm": 2.6916282684969253,
      "learning_rate": 1.8813039483757935e-07,
      "loss": 2.5405,
      "step": 5961
    },
    {
      "epoch": 0.7718548726413568,
      "grad_norm": 1.5912020564879144,
      "learning_rate": 1.879284318857618e-07,
      "loss": 2.4434,
      "step": 5962
    },
    {
      "epoch": 0.7719843350487102,
      "grad_norm": 2.4860316455501517,
      "learning_rate": 1.8772655821467124e-07,
      "loss": 2.5452,
      "step": 5963
    },
    {
      "epoch": 0.7721137974560637,
      "grad_norm": 1.8619342726099855,
      "learning_rate": 1.8752477386554433e-07,
      "loss": 2.635,
      "step": 5964
    },
    {
      "epoch": 0.7722432598634171,
      "grad_norm": 2.174473733903905,
      "learning_rate": 1.8732307887959939e-07,
      "loss": 2.4175,
      "step": 5965
    },
    {
      "epoch": 0.7722432598634171,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986461400985718,
      "eval_runtime": 14.5514,
      "eval_samples_per_second": 3.024,
      "eval_steps_per_second": 0.206,
      "step": 5965
    },
    {
      "epoch": 0.7723727222707706,
      "grad_norm": 2.2701006308701284,
      "learning_rate": 1.871214732980371e-07,
      "loss": 2.4016,
      "step": 5966
    },
    {
      "epoch": 0.772502184678124,
      "grad_norm": 1.727138334459291,
      "learning_rate": 1.8691995716203906e-07,
      "loss": 2.2354,
      "step": 5967
    },
    {
      "epoch": 0.7726316470854776,
      "grad_norm": 1.4155427550076716,
      "learning_rate": 1.8671853051276924e-07,
      "loss": 2.2651,
      "step": 5968
    },
    {
      "epoch": 0.772761109492831,
      "grad_norm": 1.9437943650458096,
      "learning_rate": 1.8651719339137355e-07,
      "loss": 2.2378,
      "step": 5969
    },
    {
      "epoch": 0.7728905719001845,
      "grad_norm": 1.3922818829841908,
      "learning_rate": 1.8631594583897835e-07,
      "loss": 2.3579,
      "step": 5970
    },
    {
      "epoch": 0.7728905719001845,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5015980005264282,
      "eval_runtime": 14.6426,
      "eval_samples_per_second": 3.005,
      "eval_steps_per_second": 0.205,
      "step": 5970
    },
    {
      "epoch": 0.773020034307538,
      "grad_norm": 1.690810646949823,
      "learning_rate": 1.861147878966931e-07,
      "loss": 2.3079,
      "step": 5971
    },
    {
      "epoch": 0.7731494967148914,
      "grad_norm": 2.0594655012712964,
      "learning_rate": 1.859137196056081e-07,
      "loss": 2.6616,
      "step": 5972
    },
    {
      "epoch": 0.7732789591222449,
      "grad_norm": 1.735278684515017,
      "learning_rate": 1.8571274100679562e-07,
      "loss": 2.4963,
      "step": 5973
    },
    {
      "epoch": 0.7734084215295983,
      "grad_norm": 1.2496522003017534,
      "learning_rate": 1.8551185214131023e-07,
      "loss": 2.4058,
      "step": 5974
    },
    {
      "epoch": 0.7735378839369518,
      "grad_norm": 2.0783386478791175,
      "learning_rate": 1.8531105305018664e-07,
      "loss": 2.3503,
      "step": 5975
    },
    {
      "epoch": 0.7735378839369518,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5000888109207153,
      "eval_runtime": 14.801,
      "eval_samples_per_second": 2.973,
      "eval_steps_per_second": 0.203,
      "step": 5975
    },
    {
      "epoch": 0.7736673463443052,
      "grad_norm": 2.107689541065852,
      "learning_rate": 1.8511034377444254e-07,
      "loss": 2.4397,
      "step": 5976
    },
    {
      "epoch": 0.7737968087516587,
      "grad_norm": 2.0955878298144386,
      "learning_rate": 1.849097243550771e-07,
      "loss": 2.4419,
      "step": 5977
    },
    {
      "epoch": 0.7739262711590122,
      "grad_norm": 1.8049458331472423,
      "learning_rate": 1.8470919483307044e-07,
      "loss": 2.6097,
      "step": 5978
    },
    {
      "epoch": 0.7740557335663657,
      "grad_norm": 1.9295600186183828,
      "learning_rate": 1.8450875524938523e-07,
      "loss": 2.4265,
      "step": 5979
    },
    {
      "epoch": 0.7741851959737192,
      "grad_norm": 1.314442472325798,
      "learning_rate": 1.8430840564496516e-07,
      "loss": 2.3301,
      "step": 5980
    },
    {
      "epoch": 0.7741851959737192,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5035954713821411,
      "eval_runtime": 14.0332,
      "eval_samples_per_second": 3.135,
      "eval_steps_per_second": 0.214,
      "step": 5980
    },
    {
      "epoch": 0.7743146583810726,
      "grad_norm": 3.4863758435551997,
      "learning_rate": 1.8410814606073544e-07,
      "loss": 2.7817,
      "step": 5981
    },
    {
      "epoch": 0.7744441207884261,
      "grad_norm": 1.2463084969887293,
      "learning_rate": 1.8390797653760353e-07,
      "loss": 2.4587,
      "step": 5982
    },
    {
      "epoch": 0.7745735831957795,
      "grad_norm": 1.6988732542503902,
      "learning_rate": 1.8370789711645778e-07,
      "loss": 2.4956,
      "step": 5983
    },
    {
      "epoch": 0.774703045603133,
      "grad_norm": 1.4724544971075704,
      "learning_rate": 1.8350790783816873e-07,
      "loss": 2.356,
      "step": 5984
    },
    {
      "epoch": 0.7748325080104864,
      "grad_norm": 1.386543429629265,
      "learning_rate": 1.8330800874358825e-07,
      "loss": 2.2573,
      "step": 5985
    },
    {
      "epoch": 0.7748325080104864,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5000221729278564,
      "eval_runtime": 14.5141,
      "eval_samples_per_second": 3.032,
      "eval_steps_per_second": 0.207,
      "step": 5985
    },
    {
      "epoch": 0.7749619704178399,
      "grad_norm": 2.7942180539683985,
      "learning_rate": 1.831081998735494e-07,
      "loss": 2.625,
      "step": 5986
    },
    {
      "epoch": 0.7750914328251934,
      "grad_norm": 4.192735906577279,
      "learning_rate": 1.8290848126886772e-07,
      "loss": 2.7822,
      "step": 5987
    },
    {
      "epoch": 0.7752208952325469,
      "grad_norm": 1.5906056842781113,
      "learning_rate": 1.8270885297033928e-07,
      "loss": 2.5168,
      "step": 5988
    },
    {
      "epoch": 0.7753503576399003,
      "grad_norm": 1.4682108535111058,
      "learning_rate": 1.825093150187427e-07,
      "loss": 2.4207,
      "step": 5989
    },
    {
      "epoch": 0.7754798200472538,
      "grad_norm": 1.42088272727761,
      "learning_rate": 1.8230986745483715e-07,
      "loss": 2.2549,
      "step": 5990
    },
    {
      "epoch": 0.7754798200472538,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5007545948028564,
      "eval_runtime": 14.0179,
      "eval_samples_per_second": 3.139,
      "eval_steps_per_second": 0.214,
      "step": 5990
    },
    {
      "epoch": 0.7756092824546073,
      "grad_norm": 2.104141643266121,
      "learning_rate": 1.8211051031936433e-07,
      "loss": 2.7314,
      "step": 5991
    },
    {
      "epoch": 0.7757387448619607,
      "grad_norm": 1.4343831673431897,
      "learning_rate": 1.8191124365304673e-07,
      "loss": 2.3268,
      "step": 5992
    },
    {
      "epoch": 0.7758682072693142,
      "grad_norm": 4.005288644184857,
      "learning_rate": 1.8171206749658846e-07,
      "loss": 2.8022,
      "step": 5993
    },
    {
      "epoch": 0.7759976696766676,
      "grad_norm": 1.8738858536186669,
      "learning_rate": 1.8151298189067558e-07,
      "loss": 2.3717,
      "step": 5994
    },
    {
      "epoch": 0.7761271320840211,
      "grad_norm": 1.7372780979044173,
      "learning_rate": 1.813139868759751e-07,
      "loss": 2.334,
      "step": 5995
    },
    {
      "epoch": 0.7761271320840211,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5007102489471436,
      "eval_runtime": 13.5662,
      "eval_samples_per_second": 3.243,
      "eval_steps_per_second": 0.221,
      "step": 5995
    },
    {
      "epoch": 0.7762565944913745,
      "grad_norm": 2.0536557945484213,
      "learning_rate": 1.8111508249313596e-07,
      "loss": 2.334,
      "step": 5996
    },
    {
      "epoch": 0.7763860568987281,
      "grad_norm": 2.9657504393827807,
      "learning_rate": 1.8091626878278846e-07,
      "loss": 2.4902,
      "step": 5997
    },
    {
      "epoch": 0.7765155193060815,
      "grad_norm": 3.226283522133443,
      "learning_rate": 1.8071754578554403e-07,
      "loss": 2.8318,
      "step": 5998
    },
    {
      "epoch": 0.776644981713435,
      "grad_norm": 1.669540781358496,
      "learning_rate": 1.8051891354199609e-07,
      "loss": 2.3899,
      "step": 5999
    },
    {
      "epoch": 0.7767744441207884,
      "grad_norm": 1.2407471452519037,
      "learning_rate": 1.8032037209271942e-07,
      "loss": 2.3147,
      "step": 6000
    },
    {
      "epoch": 0.7767744441207884,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986239671707153,
      "eval_runtime": 13.5279,
      "eval_samples_per_second": 3.253,
      "eval_steps_per_second": 0.222,
      "step": 6000
    },
    {
      "epoch": 0.7769039065281419,
      "grad_norm": 5.404269384830357,
      "learning_rate": 1.8012192147826978e-07,
      "loss": 3.0322,
      "step": 6001
    },
    {
      "epoch": 0.7770333689354953,
      "grad_norm": 1.485329955279955,
      "learning_rate": 1.7992356173918542e-07,
      "loss": 2.3342,
      "step": 6002
    },
    {
      "epoch": 0.7771628313428488,
      "grad_norm": 1.4623511786783625,
      "learning_rate": 1.7972529291598442e-07,
      "loss": 2.2488,
      "step": 6003
    },
    {
      "epoch": 0.7772922937502023,
      "grad_norm": 1.6740134219863172,
      "learning_rate": 1.7952711504916765e-07,
      "loss": 2.2202,
      "step": 6004
    },
    {
      "epoch": 0.7774217561575557,
      "grad_norm": 2.2524652678886503,
      "learning_rate": 1.793290281792172e-07,
      "loss": 2.7532,
      "step": 6005
    },
    {
      "epoch": 0.7774217561575557,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.505082607269287,
      "eval_runtime": 13.288,
      "eval_samples_per_second": 3.311,
      "eval_steps_per_second": 0.226,
      "step": 6005
    },
    {
      "epoch": 0.7775512185649092,
      "grad_norm": 1.962717132276596,
      "learning_rate": 1.7913103234659577e-07,
      "loss": 2.6069,
      "step": 6006
    },
    {
      "epoch": 0.7776806809722627,
      "grad_norm": 2.100006605008063,
      "learning_rate": 1.7893312759174864e-07,
      "loss": 2.2217,
      "step": 6007
    },
    {
      "epoch": 0.7778101433796162,
      "grad_norm": 2.489081244834181,
      "learning_rate": 1.7873531395510153e-07,
      "loss": 2.2036,
      "step": 6008
    },
    {
      "epoch": 0.7779396057869696,
      "grad_norm": 3.2805075022238737,
      "learning_rate": 1.7853759147706172e-07,
      "loss": 2.4844,
      "step": 6009
    },
    {
      "epoch": 0.7780690681943231,
      "grad_norm": 1.5028489009570547,
      "learning_rate": 1.7833996019801847e-07,
      "loss": 2.3665,
      "step": 6010
    },
    {
      "epoch": 0.7780690681943231,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.503617763519287,
      "eval_runtime": 15.0366,
      "eval_samples_per_second": 2.926,
      "eval_steps_per_second": 0.2,
      "step": 6010
    },
    {
      "epoch": 0.7781985306016765,
      "grad_norm": 2.922475990757813,
      "learning_rate": 1.7814242015834148e-07,
      "loss": 2.4111,
      "step": 6011
    },
    {
      "epoch": 0.77832799300903,
      "grad_norm": 1.9206538040083312,
      "learning_rate": 1.7794497139838273e-07,
      "loss": 2.5093,
      "step": 6012
    },
    {
      "epoch": 0.7784574554163834,
      "grad_norm": 4.446330398825083,
      "learning_rate": 1.7774761395847502e-07,
      "loss": 2.5483,
      "step": 6013
    },
    {
      "epoch": 0.7785869178237369,
      "grad_norm": 2.4702999114034525,
      "learning_rate": 1.775503478789324e-07,
      "loss": 2.4204,
      "step": 6014
    },
    {
      "epoch": 0.7787163802310904,
      "grad_norm": 1.5528279845555708,
      "learning_rate": 1.7735317320005074e-07,
      "loss": 2.3706,
      "step": 6015
    },
    {
      "epoch": 0.7787163802310904,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5014426708221436,
      "eval_runtime": 13.4565,
      "eval_samples_per_second": 3.27,
      "eval_steps_per_second": 0.223,
      "step": 6015
    },
    {
      "epoch": 0.7788458426384438,
      "grad_norm": 1.3725469383103805,
      "learning_rate": 1.7715608996210664e-07,
      "loss": 2.3918,
      "step": 6016
    },
    {
      "epoch": 0.7789753050457974,
      "grad_norm": 2.2257848269725606,
      "learning_rate": 1.769590982053588e-07,
      "loss": 2.5085,
      "step": 6017
    },
    {
      "epoch": 0.7791047674531508,
      "grad_norm": 1.8245421349674038,
      "learning_rate": 1.767621979700463e-07,
      "loss": 2.6221,
      "step": 6018
    },
    {
      "epoch": 0.7792342298605043,
      "grad_norm": 1.7483464784352891,
      "learning_rate": 1.765653892963904e-07,
      "loss": 2.3556,
      "step": 6019
    },
    {
      "epoch": 0.7793636922678577,
      "grad_norm": 1.1651156549114077,
      "learning_rate": 1.7636867222459318e-07,
      "loss": 2.3182,
      "step": 6020
    },
    {
      "epoch": 0.7793636922678577,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5020862817764282,
      "eval_runtime": 13.9999,
      "eval_samples_per_second": 3.143,
      "eval_steps_per_second": 0.214,
      "step": 6020
    },
    {
      "epoch": 0.7794931546752112,
      "grad_norm": 1.5953001045011856,
      "learning_rate": 1.7617204679483772e-07,
      "loss": 2.3447,
      "step": 6021
    },
    {
      "epoch": 0.7796226170825646,
      "grad_norm": 1.9302400920823097,
      "learning_rate": 1.7597551304728905e-07,
      "loss": 2.4697,
      "step": 6022
    },
    {
      "epoch": 0.7797520794899181,
      "grad_norm": 1.8846460464272907,
      "learning_rate": 1.7577907102209337e-07,
      "loss": 2.4314,
      "step": 6023
    },
    {
      "epoch": 0.7798815418972715,
      "grad_norm": 1.526217732623638,
      "learning_rate": 1.7558272075937772e-07,
      "loss": 2.5356,
      "step": 6024
    },
    {
      "epoch": 0.780011004304625,
      "grad_norm": 1.618006403978587,
      "learning_rate": 1.753864622992506e-07,
      "loss": 2.4326,
      "step": 6025
    },
    {
      "epoch": 0.780011004304625,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5020862817764282,
      "eval_runtime": 13.7856,
      "eval_samples_per_second": 3.192,
      "eval_steps_per_second": 0.218,
      "step": 6025
    },
    {
      "epoch": 0.7801404667119786,
      "grad_norm": 2.026234510379186,
      "learning_rate": 1.7519029568180163e-07,
      "loss": 2.5347,
      "step": 6026
    },
    {
      "epoch": 0.780269929119332,
      "grad_norm": 2.1192244142877965,
      "learning_rate": 1.74994220947102e-07,
      "loss": 2.2634,
      "step": 6027
    },
    {
      "epoch": 0.7803993915266855,
      "grad_norm": 2.033496732176806,
      "learning_rate": 1.747982381352041e-07,
      "loss": 2.3887,
      "step": 6028
    },
    {
      "epoch": 0.7805288539340389,
      "grad_norm": 1.8620097104164741,
      "learning_rate": 1.7460234728614113e-07,
      "loss": 2.2507,
      "step": 6029
    },
    {
      "epoch": 0.7806583163413924,
      "grad_norm": 1.7779340059390638,
      "learning_rate": 1.74406548439928e-07,
      "loss": 2.4829,
      "step": 6030
    },
    {
      "epoch": 0.7806583163413924,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5020419359207153,
      "eval_runtime": 17.1196,
      "eval_samples_per_second": 2.57,
      "eval_steps_per_second": 0.175,
      "step": 6030
    },
    {
      "epoch": 0.7807877787487458,
      "grad_norm": 2.038422175176203,
      "learning_rate": 1.7421084163656046e-07,
      "loss": 2.5088,
      "step": 6031
    },
    {
      "epoch": 0.7809172411560993,
      "grad_norm": 1.6833494822807096,
      "learning_rate": 1.7401522691601543e-07,
      "loss": 2.1967,
      "step": 6032
    },
    {
      "epoch": 0.7810467035634527,
      "grad_norm": 1.4285830997090943,
      "learning_rate": 1.738197043182515e-07,
      "loss": 2.3337,
      "step": 6033
    },
    {
      "epoch": 0.7811761659708062,
      "grad_norm": 2.4258863029688333,
      "learning_rate": 1.736242738832078e-07,
      "loss": 2.3711,
      "step": 6034
    },
    {
      "epoch": 0.7813056283781596,
      "grad_norm": 2.093567375033744,
      "learning_rate": 1.7342893565080536e-07,
      "loss": 2.4604,
      "step": 6035
    },
    {
      "epoch": 0.7813056283781596,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5007545948028564,
      "eval_runtime": 18.8237,
      "eval_samples_per_second": 2.337,
      "eval_steps_per_second": 0.159,
      "step": 6035
    },
    {
      "epoch": 0.7814350907855132,
      "grad_norm": 1.1240397853850335,
      "learning_rate": 1.7323368966094572e-07,
      "loss": 2.2632,
      "step": 6036
    },
    {
      "epoch": 0.7815645531928667,
      "grad_norm": 1.3347350391466952,
      "learning_rate": 1.730385359535116e-07,
      "loss": 2.3159,
      "step": 6037
    },
    {
      "epoch": 0.7816940156002201,
      "grad_norm": 2.9446166274539816,
      "learning_rate": 1.7284347456836764e-07,
      "loss": 2.4211,
      "step": 6038
    },
    {
      "epoch": 0.7818234780075736,
      "grad_norm": 2.6360210185149797,
      "learning_rate": 1.7264850554535854e-07,
      "loss": 3.1143,
      "step": 6039
    },
    {
      "epoch": 0.781952940414927,
      "grad_norm": 2.2356221365487157,
      "learning_rate": 1.7245362892431117e-07,
      "loss": 2.531,
      "step": 6040
    },
    {
      "epoch": 0.781952940414927,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5029296875,
      "eval_runtime": 14.4423,
      "eval_samples_per_second": 3.047,
      "eval_steps_per_second": 0.208,
      "step": 6040
    },
    {
      "epoch": 0.7820824028222805,
      "grad_norm": 2.2135576940179207,
      "learning_rate": 1.7225884474503276e-07,
      "loss": 2.6123,
      "step": 6041
    },
    {
      "epoch": 0.7822118652296339,
      "grad_norm": 1.3036475983430482,
      "learning_rate": 1.7206415304731174e-07,
      "loss": 2.2976,
      "step": 6042
    },
    {
      "epoch": 0.7823413276369874,
      "grad_norm": 1.4777565691449221,
      "learning_rate": 1.7186955387091823e-07,
      "loss": 2.3896,
      "step": 6043
    },
    {
      "epoch": 0.7824707900443408,
      "grad_norm": 2.2304746610198753,
      "learning_rate": 1.7167504725560272e-07,
      "loss": 2.3325,
      "step": 6044
    },
    {
      "epoch": 0.7826002524516943,
      "grad_norm": 1.3812777900337168,
      "learning_rate": 1.714806332410973e-07,
      "loss": 2.4429,
      "step": 6045
    },
    {
      "epoch": 0.7826002524516943,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4994229078292847,
      "eval_runtime": 13.8247,
      "eval_samples_per_second": 3.183,
      "eval_steps_per_second": 0.217,
      "step": 6045
    },
    {
      "epoch": 0.7827297148590479,
      "grad_norm": 1.746772466519468,
      "learning_rate": 1.7128631186711515e-07,
      "loss": 2.3796,
      "step": 6046
    },
    {
      "epoch": 0.7828591772664013,
      "grad_norm": 2.6313624599130048,
      "learning_rate": 1.7109208317335015e-07,
      "loss": 2.2812,
      "step": 6047
    },
    {
      "epoch": 0.7829886396737548,
      "grad_norm": 1.3415880733774872,
      "learning_rate": 1.7089794719947747e-07,
      "loss": 2.4106,
      "step": 6048
    },
    {
      "epoch": 0.7831181020811082,
      "grad_norm": 2.0164834403545115,
      "learning_rate": 1.7070390398515303e-07,
      "loss": 2.7104,
      "step": 6049
    },
    {
      "epoch": 0.7832475644884617,
      "grad_norm": 2.3076471201319064,
      "learning_rate": 1.7050995357001442e-07,
      "loss": 2.3557,
      "step": 6050
    },
    {
      "epoch": 0.7832475644884617,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5035954713821411,
      "eval_runtime": 13.4685,
      "eval_samples_per_second": 3.267,
      "eval_steps_per_second": 0.223,
      "step": 6050
    },
    {
      "epoch": 0.7833770268958151,
      "grad_norm": 1.5333860193828228,
      "learning_rate": 1.703160959936801e-07,
      "loss": 2.5913,
      "step": 6051
    },
    {
      "epoch": 0.7835064893031686,
      "grad_norm": 1.6555111364802721,
      "learning_rate": 1.7012233129574923e-07,
      "loss": 2.469,
      "step": 6052
    },
    {
      "epoch": 0.783635951710522,
      "grad_norm": 2.380052559698246,
      "learning_rate": 1.6992865951580187e-07,
      "loss": 2.29,
      "step": 6053
    },
    {
      "epoch": 0.7837654141178755,
      "grad_norm": 1.4981023070718518,
      "learning_rate": 1.6973508069339991e-07,
      "loss": 2.2139,
      "step": 6054
    },
    {
      "epoch": 0.7838948765252289,
      "grad_norm": 1.229957188560505,
      "learning_rate": 1.6954159486808535e-07,
      "loss": 2.1428,
      "step": 6055
    },
    {
      "epoch": 0.7838948765252289,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5022194385528564,
      "eval_runtime": 14.1945,
      "eval_samples_per_second": 3.1,
      "eval_steps_per_second": 0.211,
      "step": 6055
    },
    {
      "epoch": 0.7840243389325825,
      "grad_norm": 1.5247551864195683,
      "learning_rate": 1.6934820207938192e-07,
      "loss": 2.2375,
      "step": 6056
    },
    {
      "epoch": 0.784153801339936,
      "grad_norm": 1.9851496314498456,
      "learning_rate": 1.6915490236679364e-07,
      "loss": 2.4702,
      "step": 6057
    },
    {
      "epoch": 0.7842832637472894,
      "grad_norm": 1.2452083464703811,
      "learning_rate": 1.6896169576980634e-07,
      "loss": 2.2529,
      "step": 6058
    },
    {
      "epoch": 0.7844127261546429,
      "grad_norm": 1.9030383543744958,
      "learning_rate": 1.687685823278861e-07,
      "loss": 2.3208,
      "step": 6059
    },
    {
      "epoch": 0.7845421885619963,
      "grad_norm": 1.4702490374142587,
      "learning_rate": 1.685755620804802e-07,
      "loss": 2.4272,
      "step": 6060
    },
    {
      "epoch": 0.7845421885619963,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5014870166778564,
      "eval_runtime": 14.6061,
      "eval_samples_per_second": 3.012,
      "eval_steps_per_second": 0.205,
      "step": 6060
    },
    {
      "epoch": 0.7846716509693498,
      "grad_norm": 1.6072682710766513,
      "learning_rate": 1.6838263506701723e-07,
      "loss": 2.5132,
      "step": 6061
    },
    {
      "epoch": 0.7848011133767032,
      "grad_norm": 1.445757114460307,
      "learning_rate": 1.68189801326906e-07,
      "loss": 2.3555,
      "step": 6062
    },
    {
      "epoch": 0.7849305757840567,
      "grad_norm": 1.1522232300616841,
      "learning_rate": 1.6799706089953724e-07,
      "loss": 2.2075,
      "step": 6063
    },
    {
      "epoch": 0.7850600381914101,
      "grad_norm": 1.236587720915618,
      "learning_rate": 1.6780441382428175e-07,
      "loss": 2.25,
      "step": 6064
    },
    {
      "epoch": 0.7851895005987636,
      "grad_norm": 2.3212270236514723,
      "learning_rate": 1.676118601404915e-07,
      "loss": 2.1948,
      "step": 6065
    },
    {
      "epoch": 0.7851895005987636,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.501420497894287,
      "eval_runtime": 14.199,
      "eval_samples_per_second": 3.099,
      "eval_steps_per_second": 0.211,
      "step": 6065
    },
    {
      "epoch": 0.7853189630061171,
      "grad_norm": 1.4149513489043382,
      "learning_rate": 1.6741939988749983e-07,
      "loss": 2.3635,
      "step": 6066
    },
    {
      "epoch": 0.7854484254134706,
      "grad_norm": 2.2352431498710525,
      "learning_rate": 1.672270331046203e-07,
      "loss": 2.4707,
      "step": 6067
    },
    {
      "epoch": 0.785577887820824,
      "grad_norm": 1.8757982586603268,
      "learning_rate": 1.6703475983114794e-07,
      "loss": 2.5403,
      "step": 6068
    },
    {
      "epoch": 0.7857073502281775,
      "grad_norm": 1.4586302955174717,
      "learning_rate": 1.668425801063587e-07,
      "loss": 2.2856,
      "step": 6069
    },
    {
      "epoch": 0.785836812635531,
      "grad_norm": 1.9999095283170056,
      "learning_rate": 1.6665049396950858e-07,
      "loss": 2.5237,
      "step": 6070
    },
    {
      "epoch": 0.785836812635531,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5028408765792847,
      "eval_runtime": 14.8205,
      "eval_samples_per_second": 2.969,
      "eval_steps_per_second": 0.202,
      "step": 6070
    },
    {
      "epoch": 0.7859662750428844,
      "grad_norm": 1.9046671137401614,
      "learning_rate": 1.6645850145983535e-07,
      "loss": 2.4021,
      "step": 6071
    },
    {
      "epoch": 0.7860957374502379,
      "grad_norm": 2.050713860353067,
      "learning_rate": 1.662666026165577e-07,
      "loss": 2.2322,
      "step": 6072
    },
    {
      "epoch": 0.7862251998575913,
      "grad_norm": 2.15933696907368,
      "learning_rate": 1.6607479747887427e-07,
      "loss": 2.4199,
      "step": 6073
    },
    {
      "epoch": 0.7863546622649448,
      "grad_norm": 2.133237622356123,
      "learning_rate": 1.6588308608596587e-07,
      "loss": 2.4496,
      "step": 6074
    },
    {
      "epoch": 0.7864841246722983,
      "grad_norm": 1.3084679715361984,
      "learning_rate": 1.656914684769926e-07,
      "loss": 2.2544,
      "step": 6075
    },
    {
      "epoch": 0.7864841246722983,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5007102489471436,
      "eval_runtime": 14.0418,
      "eval_samples_per_second": 3.133,
      "eval_steps_per_second": 0.214,
      "step": 6075
    },
    {
      "epoch": 0.7866135870796518,
      "grad_norm": 1.4515417691836674,
      "learning_rate": 1.6549994469109667e-07,
      "loss": 2.4021,
      "step": 6076
    },
    {
      "epoch": 0.7867430494870052,
      "grad_norm": 1.3269648480077896,
      "learning_rate": 1.6530851476740092e-07,
      "loss": 2.2122,
      "step": 6077
    },
    {
      "epoch": 0.7868725118943587,
      "grad_norm": 1.2712347927165593,
      "learning_rate": 1.651171787450083e-07,
      "loss": 2.2146,
      "step": 6078
    },
    {
      "epoch": 0.7870019743017121,
      "grad_norm": 2.0675892990491653,
      "learning_rate": 1.6492593666300348e-07,
      "loss": 2.1978,
      "step": 6079
    },
    {
      "epoch": 0.7871314367090656,
      "grad_norm": 2.156715003341984,
      "learning_rate": 1.6473478856045144e-07,
      "loss": 2.2756,
      "step": 6080
    },
    {
      "epoch": 0.7871314367090656,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4994229078292847,
      "eval_runtime": 14.3915,
      "eval_samples_per_second": 3.057,
      "eval_steps_per_second": 0.208,
      "step": 6080
    },
    {
      "epoch": 0.787260899116419,
      "grad_norm": 1.8151402981908307,
      "learning_rate": 1.645437344763978e-07,
      "loss": 2.3232,
      "step": 6081
    },
    {
      "epoch": 0.7873903615237725,
      "grad_norm": 2.3379938434342873,
      "learning_rate": 1.6435277444986956e-07,
      "loss": 2.4546,
      "step": 6082
    },
    {
      "epoch": 0.787519823931126,
      "grad_norm": 2.077111917477973,
      "learning_rate": 1.6416190851987388e-07,
      "loss": 2.3572,
      "step": 6083
    },
    {
      "epoch": 0.7876492863384794,
      "grad_norm": 1.4978347619710501,
      "learning_rate": 1.6397113672539933e-07,
      "loss": 2.4421,
      "step": 6084
    },
    {
      "epoch": 0.787778748745833,
      "grad_norm": 1.508838964561892,
      "learning_rate": 1.6378045910541457e-07,
      "loss": 2.3975,
      "step": 6085
    },
    {
      "epoch": 0.787778748745833,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5037509202957153,
      "eval_runtime": 14.7821,
      "eval_samples_per_second": 2.977,
      "eval_steps_per_second": 0.203,
      "step": 6085
    },
    {
      "epoch": 0.7879082111531864,
      "grad_norm": 2.176982053740583,
      "learning_rate": 1.6358987569886965e-07,
      "loss": 2.6199,
      "step": 6086
    },
    {
      "epoch": 0.7880376735605399,
      "grad_norm": 1.3240284918967633,
      "learning_rate": 1.6339938654469506e-07,
      "loss": 2.2749,
      "step": 6087
    },
    {
      "epoch": 0.7881671359678933,
      "grad_norm": 2.148172982195951,
      "learning_rate": 1.632089916818017e-07,
      "loss": 2.0894,
      "step": 6088
    },
    {
      "epoch": 0.7882965983752468,
      "grad_norm": 1.4095775181701338,
      "learning_rate": 1.6301869114908212e-07,
      "loss": 2.2397,
      "step": 6089
    },
    {
      "epoch": 0.7884260607826002,
      "grad_norm": 1.1630548572233477,
      "learning_rate": 1.6282848498540858e-07,
      "loss": 2.3071,
      "step": 6090
    },
    {
      "epoch": 0.7884260607826002,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5021750926971436,
      "eval_runtime": 14.1054,
      "eval_samples_per_second": 3.119,
      "eval_steps_per_second": 0.213,
      "step": 6090
    },
    {
      "epoch": 0.7885555231899537,
      "grad_norm": 1.4229698546876761,
      "learning_rate": 1.62638373229635e-07,
      "loss": 2.272,
      "step": 6091
    },
    {
      "epoch": 0.7886849855973072,
      "grad_norm": 2.05553390751181,
      "learning_rate": 1.6244835592059534e-07,
      "loss": 2.2585,
      "step": 6092
    },
    {
      "epoch": 0.7888144480046606,
      "grad_norm": 1.7631848787541045,
      "learning_rate": 1.6225843309710429e-07,
      "loss": 2.6106,
      "step": 6093
    },
    {
      "epoch": 0.7889439104120141,
      "grad_norm": 1.3046313223565162,
      "learning_rate": 1.6206860479795757e-07,
      "loss": 2.0957,
      "step": 6094
    },
    {
      "epoch": 0.7890733728193676,
      "grad_norm": 1.649494081694161,
      "learning_rate": 1.618788710619317e-07,
      "loss": 2.4575,
      "step": 6095
    },
    {
      "epoch": 0.7890733728193676,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5007990598678589,
      "eval_runtime": 14.2969,
      "eval_samples_per_second": 3.078,
      "eval_steps_per_second": 0.21,
      "step": 6095
    },
    {
      "epoch": 0.7892028352267211,
      "grad_norm": 2.70446400709977,
      "learning_rate": 1.616892319277833e-07,
      "loss": 2.6511,
      "step": 6096
    },
    {
      "epoch": 0.7893322976340745,
      "grad_norm": 1.9751139149839274,
      "learning_rate": 1.6149968743425054e-07,
      "loss": 2.5085,
      "step": 6097
    },
    {
      "epoch": 0.789461760041428,
      "grad_norm": 2.0019163957443253,
      "learning_rate": 1.61310237620051e-07,
      "loss": 2.4878,
      "step": 6098
    },
    {
      "epoch": 0.7895912224487814,
      "grad_norm": 1.4574557164547899,
      "learning_rate": 1.6112088252388392e-07,
      "loss": 2.2582,
      "step": 6099
    },
    {
      "epoch": 0.7897206848561349,
      "grad_norm": 1.4493033884505937,
      "learning_rate": 1.609316221844292e-07,
      "loss": 2.3171,
      "step": 6100
    },
    {
      "epoch": 0.7897206848561349,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5008434057235718,
      "eval_runtime": 14.1304,
      "eval_samples_per_second": 3.114,
      "eval_steps_per_second": 0.212,
      "step": 6100
    },
    {
      "epoch": 0.7898501472634883,
      "grad_norm": 1.381342483738127,
      "learning_rate": 1.607424566403467e-07,
      "loss": 2.2146,
      "step": 6101
    },
    {
      "epoch": 0.7899796096708418,
      "grad_norm": 1.7650522579979775,
      "learning_rate": 1.605533859302777e-07,
      "loss": 2.377,
      "step": 6102
    },
    {
      "epoch": 0.7901090720781953,
      "grad_norm": 1.141203095347824,
      "learning_rate": 1.6036441009284352e-07,
      "loss": 2.3198,
      "step": 6103
    },
    {
      "epoch": 0.7902385344855487,
      "grad_norm": 3.3857860455406317,
      "learning_rate": 1.601755291666461e-07,
      "loss": 2.3169,
      "step": 6104
    },
    {
      "epoch": 0.7903679968929023,
      "grad_norm": 2.0365630815906415,
      "learning_rate": 1.5998674319026865e-07,
      "loss": 2.2861,
      "step": 6105
    },
    {
      "epoch": 0.7903679968929023,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986904859542847,
      "eval_runtime": 14.4806,
      "eval_samples_per_second": 3.039,
      "eval_steps_per_second": 0.207,
      "step": 6105
    },
    {
      "epoch": 0.7904974593002557,
      "grad_norm": 1.3695382302286623,
      "learning_rate": 1.5979805220227408e-07,
      "loss": 2.0312,
      "step": 6106
    },
    {
      "epoch": 0.7906269217076092,
      "grad_norm": 1.7410987473443396,
      "learning_rate": 1.5960945624120676e-07,
      "loss": 2.4082,
      "step": 6107
    },
    {
      "epoch": 0.7907563841149626,
      "grad_norm": 1.4099151041116424,
      "learning_rate": 1.5942095534559103e-07,
      "loss": 2.3262,
      "step": 6108
    },
    {
      "epoch": 0.7908858465223161,
      "grad_norm": 2.22867655793667,
      "learning_rate": 1.5923254955393187e-07,
      "loss": 2.3796,
      "step": 6109
    },
    {
      "epoch": 0.7910153089296695,
      "grad_norm": 1.9505459585713207,
      "learning_rate": 1.5904423890471525e-07,
      "loss": 2.3708,
      "step": 6110
    },
    {
      "epoch": 0.7910153089296695,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5,
      "eval_runtime": 16.3355,
      "eval_samples_per_second": 2.694,
      "eval_steps_per_second": 0.184,
      "step": 6110
    },
    {
      "epoch": 0.791144771337023,
      "grad_norm": 1.4012679389504141,
      "learning_rate": 1.588560234364072e-07,
      "loss": 2.3118,
      "step": 6111
    },
    {
      "epoch": 0.7912742337443764,
      "grad_norm": 2.5931979312858386,
      "learning_rate": 1.5866790318745487e-07,
      "loss": 2.2876,
      "step": 6112
    },
    {
      "epoch": 0.7914036961517299,
      "grad_norm": 1.6244875495332793,
      "learning_rate": 1.5847987819628541e-07,
      "loss": 2.5779,
      "step": 6113
    },
    {
      "epoch": 0.7915331585590835,
      "grad_norm": 1.2726872583020585,
      "learning_rate": 1.5829194850130652e-07,
      "loss": 2.2617,
      "step": 6114
    },
    {
      "epoch": 0.7916626209664369,
      "grad_norm": 1.5036704262876284,
      "learning_rate": 1.5810411414090717e-07,
      "loss": 2.3987,
      "step": 6115
    },
    {
      "epoch": 0.7916626209664369,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986461400985718,
      "eval_runtime": 14.845,
      "eval_samples_per_second": 2.964,
      "eval_steps_per_second": 0.202,
      "step": 6115
    },
    {
      "epoch": 0.7917920833737904,
      "grad_norm": 1.6929095708158466,
      "learning_rate": 1.5791637515345585e-07,
      "loss": 2.4221,
      "step": 6116
    },
    {
      "epoch": 0.7919215457811438,
      "grad_norm": 1.2109538515779918,
      "learning_rate": 1.5772873157730219e-07,
      "loss": 2.0928,
      "step": 6117
    },
    {
      "epoch": 0.7920510081884973,
      "grad_norm": 1.4096068177328966,
      "learning_rate": 1.5754118345077653e-07,
      "loss": 2.3264,
      "step": 6118
    },
    {
      "epoch": 0.7921804705958507,
      "grad_norm": 1.5250632675031588,
      "learning_rate": 1.5735373081218905e-07,
      "loss": 2.3267,
      "step": 6119
    },
    {
      "epoch": 0.7923099330032042,
      "grad_norm": 2.2773189210692024,
      "learning_rate": 1.5716637369983085e-07,
      "loss": 2.2449,
      "step": 6120
    },
    {
      "epoch": 0.7923099330032042,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986683130264282,
      "eval_runtime": 15.294,
      "eval_samples_per_second": 2.877,
      "eval_steps_per_second": 0.196,
      "step": 6120
    },
    {
      "epoch": 0.7924393954105576,
      "grad_norm": 1.3124534186423893,
      "learning_rate": 1.5697911215197314e-07,
      "loss": 2.4976,
      "step": 6121
    },
    {
      "epoch": 0.7925688578179111,
      "grad_norm": 2.1709334320623666,
      "learning_rate": 1.5679194620686816e-07,
      "loss": 2.4229,
      "step": 6122
    },
    {
      "epoch": 0.7926983202252645,
      "grad_norm": 1.781338773354743,
      "learning_rate": 1.566048759027485e-07,
      "loss": 2.356,
      "step": 6123
    },
    {
      "epoch": 0.7928277826326181,
      "grad_norm": 2.3428167127130335,
      "learning_rate": 1.5641790127782692e-07,
      "loss": 2.4717,
      "step": 6124
    },
    {
      "epoch": 0.7929572450399716,
      "grad_norm": 1.296578328463195,
      "learning_rate": 1.5623102237029646e-07,
      "loss": 2.3322,
      "step": 6125
    },
    {
      "epoch": 0.7929572450399716,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500776767730713,
      "eval_runtime": 14.3348,
      "eval_samples_per_second": 3.069,
      "eval_steps_per_second": 0.209,
      "step": 6125
    },
    {
      "epoch": 0.793086707447325,
      "grad_norm": 1.722932835405144,
      "learning_rate": 1.5604423921833146e-07,
      "loss": 2.5996,
      "step": 6126
    },
    {
      "epoch": 0.7932161698546785,
      "grad_norm": 1.1117891329149647,
      "learning_rate": 1.5585755186008565e-07,
      "loss": 2.2727,
      "step": 6127
    },
    {
      "epoch": 0.7933456322620319,
      "grad_norm": 1.4040584855688887,
      "learning_rate": 1.5567096033369414e-07,
      "loss": 2.3135,
      "step": 6128
    },
    {
      "epoch": 0.7934750946693854,
      "grad_norm": 2.123962151631288,
      "learning_rate": 1.5548446467727172e-07,
      "loss": 2.3286,
      "step": 6129
    },
    {
      "epoch": 0.7936045570767388,
      "grad_norm": 1.214305532013705,
      "learning_rate": 1.5529806492891418e-07,
      "loss": 2.3435,
      "step": 6130
    },
    {
      "epoch": 0.7936045570767388,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5022194385528564,
      "eval_runtime": 13.7346,
      "eval_samples_per_second": 3.204,
      "eval_steps_per_second": 0.218,
      "step": 6130
    },
    {
      "epoch": 0.7937340194840923,
      "grad_norm": 2.3263319730143834,
      "learning_rate": 1.551117611266973e-07,
      "loss": 2.3433,
      "step": 6131
    },
    {
      "epoch": 0.7938634818914457,
      "grad_norm": 2.3992391449933206,
      "learning_rate": 1.5492555330867722e-07,
      "loss": 2.4736,
      "step": 6132
    },
    {
      "epoch": 0.7939929442987992,
      "grad_norm": 1.6800669755646818,
      "learning_rate": 1.5473944151289103e-07,
      "loss": 2.1008,
      "step": 6133
    },
    {
      "epoch": 0.7941224067061528,
      "grad_norm": 1.6445850689757073,
      "learning_rate": 1.5455342577735551e-07,
      "loss": 2.271,
      "step": 6134
    },
    {
      "epoch": 0.7942518691135062,
      "grad_norm": 2.081598306888067,
      "learning_rate": 1.5436750614006847e-07,
      "loss": 2.5806,
      "step": 6135
    },
    {
      "epoch": 0.7942518691135062,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5008434057235718,
      "eval_runtime": 14.2781,
      "eval_samples_per_second": 3.082,
      "eval_steps_per_second": 0.21,
      "step": 6135
    },
    {
      "epoch": 0.7943813315208597,
      "grad_norm": 1.4514280925118617,
      "learning_rate": 1.541816826390076e-07,
      "loss": 2.4148,
      "step": 6136
    },
    {
      "epoch": 0.7945107939282131,
      "grad_norm": 1.3174681619173314,
      "learning_rate": 1.5399595531213092e-07,
      "loss": 2.2729,
      "step": 6137
    },
    {
      "epoch": 0.7946402563355666,
      "grad_norm": 1.6215640398270907,
      "learning_rate": 1.5381032419737727e-07,
      "loss": 2.4543,
      "step": 6138
    },
    {
      "epoch": 0.79476971874292,
      "grad_norm": 1.2365193370324117,
      "learning_rate": 1.5362478933266542e-07,
      "loss": 2.1436,
      "step": 6139
    },
    {
      "epoch": 0.7948991811502735,
      "grad_norm": 2.632926011076279,
      "learning_rate": 1.5343935075589464e-07,
      "loss": 2.4131,
      "step": 6140
    },
    {
      "epoch": 0.7948991811502735,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.499467372894287,
      "eval_runtime": 13.8808,
      "eval_samples_per_second": 3.17,
      "eval_steps_per_second": 0.216,
      "step": 6140
    },
    {
      "epoch": 0.7950286435576269,
      "grad_norm": 1.1284731465856581,
      "learning_rate": 1.532540085049449e-07,
      "loss": 2.311,
      "step": 6141
    },
    {
      "epoch": 0.7951581059649804,
      "grad_norm": 1.2449469392027743,
      "learning_rate": 1.530687626176755e-07,
      "loss": 2.2598,
      "step": 6142
    },
    {
      "epoch": 0.7952875683723338,
      "grad_norm": 1.7788824378976857,
      "learning_rate": 1.5288361313192708e-07,
      "loss": 2.3315,
      "step": 6143
    },
    {
      "epoch": 0.7954170307796874,
      "grad_norm": 1.5167276117801085,
      "learning_rate": 1.5269856008551986e-07,
      "loss": 2.3459,
      "step": 6144
    },
    {
      "epoch": 0.7955464931870408,
      "grad_norm": 1.5919775619850602,
      "learning_rate": 1.5251360351625486e-07,
      "loss": 2.2229,
      "step": 6145
    },
    {
      "epoch": 0.7955464931870408,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964267015457153,
      "eval_runtime": 14.7869,
      "eval_samples_per_second": 2.976,
      "eval_steps_per_second": 0.203,
      "step": 6145
    },
    {
      "epoch": 0.7956759555943943,
      "grad_norm": 1.2648158119597517,
      "learning_rate": 1.5232874346191333e-07,
      "loss": 2.3623,
      "step": 6146
    },
    {
      "epoch": 0.7958054180017478,
      "grad_norm": 1.1042094065120194,
      "learning_rate": 1.5214397996025663e-07,
      "loss": 2.2642,
      "step": 6147
    },
    {
      "epoch": 0.7959348804091012,
      "grad_norm": 2.247255745999499,
      "learning_rate": 1.519593130490262e-07,
      "loss": 2.2715,
      "step": 6148
    },
    {
      "epoch": 0.7960643428164547,
      "grad_norm": 1.8048065815137926,
      "learning_rate": 1.5177474276594432e-07,
      "loss": 2.5674,
      "step": 6149
    },
    {
      "epoch": 0.7961938052238081,
      "grad_norm": 1.3390389414207526,
      "learning_rate": 1.515902691487129e-07,
      "loss": 2.54,
      "step": 6150
    },
    {
      "epoch": 0.7961938052238081,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4993785619735718,
      "eval_runtime": 14.5257,
      "eval_samples_per_second": 3.029,
      "eval_steps_per_second": 0.207,
      "step": 6150
    },
    {
      "epoch": 0.7963232676311616,
      "grad_norm": 1.5342239109903704,
      "learning_rate": 1.5140589223501478e-07,
      "loss": 2.3716,
      "step": 6151
    },
    {
      "epoch": 0.796452730038515,
      "grad_norm": 4.69528803565607,
      "learning_rate": 1.5122161206251249e-07,
      "loss": 2.5566,
      "step": 6152
    },
    {
      "epoch": 0.7965821924458685,
      "grad_norm": 1.8066211949299904,
      "learning_rate": 1.5103742866884874e-07,
      "loss": 2.4558,
      "step": 6153
    },
    {
      "epoch": 0.796711654853222,
      "grad_norm": 1.275134341647327,
      "learning_rate": 1.508533420916472e-07,
      "loss": 2.3179,
      "step": 6154
    },
    {
      "epoch": 0.7968411172605755,
      "grad_norm": 2.072994879982309,
      "learning_rate": 1.506693523685109e-07,
      "loss": 2.176,
      "step": 6155
    },
    {
      "epoch": 0.7968411172605755,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5014426708221436,
      "eval_runtime": 14.5781,
      "eval_samples_per_second": 3.018,
      "eval_steps_per_second": 0.206,
      "step": 6155
    },
    {
      "epoch": 0.796970579667929,
      "grad_norm": 1.5787633460062298,
      "learning_rate": 1.504854595370238e-07,
      "loss": 2.3162,
      "step": 6156
    },
    {
      "epoch": 0.7971000420752824,
      "grad_norm": 1.2148493159576272,
      "learning_rate": 1.503016636347493e-07,
      "loss": 2.3179,
      "step": 6157
    },
    {
      "epoch": 0.7972295044826359,
      "grad_norm": 2.0903271733253805,
      "learning_rate": 1.50117964699232e-07,
      "loss": 2.3411,
      "step": 6158
    },
    {
      "epoch": 0.7973589668899893,
      "grad_norm": 2.3377298729299483,
      "learning_rate": 1.4993436276799568e-07,
      "loss": 2.4409,
      "step": 6159
    },
    {
      "epoch": 0.7974884292973428,
      "grad_norm": 1.2123770940194514,
      "learning_rate": 1.4975085787854484e-07,
      "loss": 2.1208,
      "step": 6160
    },
    {
      "epoch": 0.7974884292973428,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4958717823028564,
      "eval_runtime": 14.0711,
      "eval_samples_per_second": 3.127,
      "eval_steps_per_second": 0.213,
      "step": 6160
    },
    {
      "epoch": 0.7976178917046962,
      "grad_norm": 2.270076638111404,
      "learning_rate": 1.495674500683643e-07,
      "loss": 2.2144,
      "step": 6161
    },
    {
      "epoch": 0.7977473541120497,
      "grad_norm": 4.356558500757882,
      "learning_rate": 1.4938413937491847e-07,
      "loss": 2.5037,
      "step": 6162
    },
    {
      "epoch": 0.7978768165194032,
      "grad_norm": 1.4789728870592507,
      "learning_rate": 1.492009258356527e-07,
      "loss": 2.321,
      "step": 6163
    },
    {
      "epoch": 0.7980062789267567,
      "grad_norm": 3.8056498511021806,
      "learning_rate": 1.4901780948799176e-07,
      "loss": 2.9966,
      "step": 6164
    },
    {
      "epoch": 0.7981357413341101,
      "grad_norm": 1.486241073019933,
      "learning_rate": 1.488347903693408e-07,
      "loss": 2.3691,
      "step": 6165
    },
    {
      "epoch": 0.7981357413341101,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5001331567764282,
      "eval_runtime": 14.1203,
      "eval_samples_per_second": 3.116,
      "eval_steps_per_second": 0.212,
      "step": 6165
    },
    {
      "epoch": 0.7982652037414636,
      "grad_norm": 1.9894754745072447,
      "learning_rate": 1.4865186851708534e-07,
      "loss": 2.4604,
      "step": 6166
    },
    {
      "epoch": 0.798394666148817,
      "grad_norm": 2.433186741899739,
      "learning_rate": 1.4846904396859104e-07,
      "loss": 2.4052,
      "step": 6167
    },
    {
      "epoch": 0.7985241285561705,
      "grad_norm": 1.5172543604318443,
      "learning_rate": 1.4828631676120317e-07,
      "loss": 2.4414,
      "step": 6168
    },
    {
      "epoch": 0.798653590963524,
      "grad_norm": 1.713319160393024,
      "learning_rate": 1.4810368693224802e-07,
      "loss": 2.282,
      "step": 6169
    },
    {
      "epoch": 0.7987830533708774,
      "grad_norm": 1.6772868397887502,
      "learning_rate": 1.4792115451903074e-07,
      "loss": 2.3088,
      "step": 6170
    },
    {
      "epoch": 0.7987830533708774,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986461400985718,
      "eval_runtime": 14.8222,
      "eval_samples_per_second": 2.969,
      "eval_steps_per_second": 0.202,
      "step": 6170
    },
    {
      "epoch": 0.7989125157782309,
      "grad_norm": 2.220546391139406,
      "learning_rate": 1.477387195588375e-07,
      "loss": 2.3459,
      "step": 6171
    },
    {
      "epoch": 0.7990419781855843,
      "grad_norm": 1.8250866516284683,
      "learning_rate": 1.4755638208893468e-07,
      "loss": 2.2881,
      "step": 6172
    },
    {
      "epoch": 0.7991714405929379,
      "grad_norm": 1.1185015553730828,
      "learning_rate": 1.4737414214656798e-07,
      "loss": 2.2446,
      "step": 6173
    },
    {
      "epoch": 0.7993009030002913,
      "grad_norm": 2.6775974898127664,
      "learning_rate": 1.4719199976896394e-07,
      "loss": 2.5151,
      "step": 6174
    },
    {
      "epoch": 0.7994303654076448,
      "grad_norm": 1.5824095031017735,
      "learning_rate": 1.470099549933287e-07,
      "loss": 2.3352,
      "step": 6175
    },
    {
      "epoch": 0.7994303654076448,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986461400985718,
      "eval_runtime": 15.3413,
      "eval_samples_per_second": 2.868,
      "eval_steps_per_second": 0.196,
      "step": 6175
    },
    {
      "epoch": 0.7995598278149982,
      "grad_norm": 2.526927617340138,
      "learning_rate": 1.468280078568484e-07,
      "loss": 2.4373,
      "step": 6176
    },
    {
      "epoch": 0.7996892902223517,
      "grad_norm": 1.4544605430921675,
      "learning_rate": 1.4664615839668974e-07,
      "loss": 2.3767,
      "step": 6177
    },
    {
      "epoch": 0.7998187526297051,
      "grad_norm": 1.7705634737528582,
      "learning_rate": 1.4646440664999887e-07,
      "loss": 2.3916,
      "step": 6178
    },
    {
      "epoch": 0.7999482150370586,
      "grad_norm": 2.18003445685703,
      "learning_rate": 1.4628275265390257e-07,
      "loss": 2.4138,
      "step": 6179
    },
    {
      "epoch": 0.800077677444412,
      "grad_norm": 3.0670119676062964,
      "learning_rate": 1.4610119644550718e-07,
      "loss": 2.4902,
      "step": 6180
    },
    {
      "epoch": 0.800077677444412,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5007990598678589,
      "eval_runtime": 13.7044,
      "eval_samples_per_second": 3.211,
      "eval_steps_per_second": 0.219,
      "step": 6180
    },
    {
      "epoch": 0.8002071398517655,
      "grad_norm": 3.3851435330336477,
      "learning_rate": 1.4591973806189908e-07,
      "loss": 2.8569,
      "step": 6181
    },
    {
      "epoch": 0.800336602259119,
      "grad_norm": 2.098975812003673,
      "learning_rate": 1.4573837754014526e-07,
      "loss": 2.3987,
      "step": 6182
    },
    {
      "epoch": 0.8004660646664725,
      "grad_norm": 1.399760454837553,
      "learning_rate": 1.455571149172917e-07,
      "loss": 2.3491,
      "step": 6183
    },
    {
      "epoch": 0.800595527073826,
      "grad_norm": 1.7284324535569071,
      "learning_rate": 1.4537595023036561e-07,
      "loss": 2.1836,
      "step": 6184
    },
    {
      "epoch": 0.8007249894811794,
      "grad_norm": 1.3966960719229413,
      "learning_rate": 1.4519488351637293e-07,
      "loss": 2.3237,
      "step": 6185
    },
    {
      "epoch": 0.8007249894811794,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4993785619735718,
      "eval_runtime": 14.532,
      "eval_samples_per_second": 3.028,
      "eval_steps_per_second": 0.206,
      "step": 6185
    },
    {
      "epoch": 0.8008544518885329,
      "grad_norm": 2.476369015460651,
      "learning_rate": 1.4501391481230074e-07,
      "loss": 2.3005,
      "step": 6186
    },
    {
      "epoch": 0.8009839142958863,
      "grad_norm": 1.2559412843467503,
      "learning_rate": 1.4483304415511533e-07,
      "loss": 2.2009,
      "step": 6187
    },
    {
      "epoch": 0.8011133767032398,
      "grad_norm": 1.8911858337255594,
      "learning_rate": 1.4465227158176297e-07,
      "loss": 2.3667,
      "step": 6188
    },
    {
      "epoch": 0.8012428391105932,
      "grad_norm": 1.5026642102620302,
      "learning_rate": 1.4447159712917032e-07,
      "loss": 2.2295,
      "step": 6189
    },
    {
      "epoch": 0.8013723015179467,
      "grad_norm": 2.2139346328136305,
      "learning_rate": 1.4429102083424407e-07,
      "loss": 2.1887,
      "step": 6190
    },
    {
      "epoch": 0.8013723015179467,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4958274364471436,
      "eval_runtime": 14.2221,
      "eval_samples_per_second": 3.094,
      "eval_steps_per_second": 0.211,
      "step": 6190
    },
    {
      "epoch": 0.8015017639253001,
      "grad_norm": 1.9892865856060675,
      "learning_rate": 1.441105427338703e-07,
      "loss": 2.3511,
      "step": 6191
    },
    {
      "epoch": 0.8016312263326536,
      "grad_norm": 1.0499738312326028,
      "learning_rate": 1.4393016286491535e-07,
      "loss": 2.238,
      "step": 6192
    },
    {
      "epoch": 0.8017606887400072,
      "grad_norm": 2.1501908537367687,
      "learning_rate": 1.4374988126422533e-07,
      "loss": 2.3462,
      "step": 6193
    },
    {
      "epoch": 0.8018901511473606,
      "grad_norm": 3.735022519512406,
      "learning_rate": 1.4356969796862659e-07,
      "loss": 2.8796,
      "step": 6194
    },
    {
      "epoch": 0.8020196135547141,
      "grad_norm": 1.6034306959702043,
      "learning_rate": 1.4338961301492532e-07,
      "loss": 2.1833,
      "step": 6195
    },
    {
      "epoch": 0.8020196135547141,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500732421875,
      "eval_runtime": 14.5201,
      "eval_samples_per_second": 3.03,
      "eval_steps_per_second": 0.207,
      "step": 6195
    },
    {
      "epoch": 0.8021490759620675,
      "grad_norm": 2.0791454101747857,
      "learning_rate": 1.4320962643990716e-07,
      "loss": 2.6873,
      "step": 6196
    },
    {
      "epoch": 0.802278538369421,
      "grad_norm": 1.8086305881168148,
      "learning_rate": 1.4302973828033848e-07,
      "loss": 2.3369,
      "step": 6197
    },
    {
      "epoch": 0.8024080007767744,
      "grad_norm": 1.4907877786733656,
      "learning_rate": 1.4284994857296488e-07,
      "loss": 2.2207,
      "step": 6198
    },
    {
      "epoch": 0.8025374631841279,
      "grad_norm": 1.4086190468402466,
      "learning_rate": 1.4267025735451182e-07,
      "loss": 2.3857,
      "step": 6199
    },
    {
      "epoch": 0.8026669255914813,
      "grad_norm": 3.1230262128746,
      "learning_rate": 1.424906646616852e-07,
      "loss": 2.4231,
      "step": 6200
    },
    {
      "epoch": 0.8026669255914813,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4999778270721436,
      "eval_runtime": 13.8936,
      "eval_samples_per_second": 3.167,
      "eval_steps_per_second": 0.216,
      "step": 6200
    },
    {
      "epoch": 0.8027963879988348,
      "grad_norm": 1.943849445574589,
      "learning_rate": 1.423111705311702e-07,
      "loss": 2.5109,
      "step": 6201
    },
    {
      "epoch": 0.8029258504061884,
      "grad_norm": 2.1386678798622745,
      "learning_rate": 1.4213177499963246e-07,
      "loss": 2.3232,
      "step": 6202
    },
    {
      "epoch": 0.8030553128135418,
      "grad_norm": 1.4020995198635677,
      "learning_rate": 1.4195247810371703e-07,
      "loss": 2.1711,
      "step": 6203
    },
    {
      "epoch": 0.8031847752208953,
      "grad_norm": 1.486724484433731,
      "learning_rate": 1.4177327988004864e-07,
      "loss": 2.417,
      "step": 6204
    },
    {
      "epoch": 0.8033142376282487,
      "grad_norm": 3.3248342435368192,
      "learning_rate": 1.4159418036523252e-07,
      "loss": 2.5938,
      "step": 6205
    },
    {
      "epoch": 0.8033142376282487,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986017942428589,
      "eval_runtime": 13.7072,
      "eval_samples_per_second": 3.21,
      "eval_steps_per_second": 0.219,
      "step": 6205
    },
    {
      "epoch": 0.8034437000356022,
      "grad_norm": 2.2110859961343845,
      "learning_rate": 1.4141517959585315e-07,
      "loss": 2.5532,
      "step": 6206
    },
    {
      "epoch": 0.8035731624429556,
      "grad_norm": 1.4512526614295482,
      "learning_rate": 1.412362776084753e-07,
      "loss": 2.2158,
      "step": 6207
    },
    {
      "epoch": 0.8037026248503091,
      "grad_norm": 1.9977354722350273,
      "learning_rate": 1.410574744396432e-07,
      "loss": 2.4082,
      "step": 6208
    },
    {
      "epoch": 0.8038320872576625,
      "grad_norm": 2.2175684955777224,
      "learning_rate": 1.4087877012588084e-07,
      "loss": 2.4836,
      "step": 6209
    },
    {
      "epoch": 0.803961549665016,
      "grad_norm": 2.5915435879233573,
      "learning_rate": 1.4070016470369257e-07,
      "loss": 2.415,
      "step": 6210
    },
    {
      "epoch": 0.803961549665016,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972478151321411,
      "eval_runtime": 14.0645,
      "eval_samples_per_second": 3.128,
      "eval_steps_per_second": 0.213,
      "step": 6210
    },
    {
      "epoch": 0.8040910120723694,
      "grad_norm": 2.1549961037032115,
      "learning_rate": 1.4052165820956177e-07,
      "loss": 2.438,
      "step": 6211
    },
    {
      "epoch": 0.804220474479723,
      "grad_norm": 2.5724127751869568,
      "learning_rate": 1.403432506799522e-07,
      "loss": 2.2173,
      "step": 6212
    },
    {
      "epoch": 0.8043499368870765,
      "grad_norm": 2.468218094430767,
      "learning_rate": 1.401649421513074e-07,
      "loss": 2.3452,
      "step": 6213
    },
    {
      "epoch": 0.8044793992944299,
      "grad_norm": 1.2188889868448305,
      "learning_rate": 1.3998673266005035e-07,
      "loss": 2.1982,
      "step": 6214
    },
    {
      "epoch": 0.8046088617017834,
      "grad_norm": 2.0382202235456575,
      "learning_rate": 1.3980862224258395e-07,
      "loss": 2.3364,
      "step": 6215
    },
    {
      "epoch": 0.8046088617017834,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500932216644287,
      "eval_runtime": 15.1166,
      "eval_samples_per_second": 2.911,
      "eval_steps_per_second": 0.198,
      "step": 6215
    },
    {
      "epoch": 0.8047383241091368,
      "grad_norm": 4.0642409601976315,
      "learning_rate": 1.396306109352906e-07,
      "loss": 2.4946,
      "step": 6216
    },
    {
      "epoch": 0.8048677865164903,
      "grad_norm": 2.195532698601248,
      "learning_rate": 1.3945269877453295e-07,
      "loss": 2.4839,
      "step": 6217
    },
    {
      "epoch": 0.8049972489238437,
      "grad_norm": 1.9463459955150941,
      "learning_rate": 1.392748857966534e-07,
      "loss": 2.2126,
      "step": 6218
    },
    {
      "epoch": 0.8051267113311972,
      "grad_norm": 1.7228753645977994,
      "learning_rate": 1.3909717203797358e-07,
      "loss": 2.3926,
      "step": 6219
    },
    {
      "epoch": 0.8052561737385506,
      "grad_norm": 1.435002541132625,
      "learning_rate": 1.3891955753479507e-07,
      "loss": 2.252,
      "step": 6220
    },
    {
      "epoch": 0.8052561737385506,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498002529144287,
      "eval_runtime": 14.0585,
      "eval_samples_per_second": 3.13,
      "eval_steps_per_second": 0.213,
      "step": 6220
    },
    {
      "epoch": 0.8053856361459041,
      "grad_norm": 1.3873675557847818,
      "learning_rate": 1.387420423233994e-07,
      "loss": 2.3364,
      "step": 6221
    },
    {
      "epoch": 0.8055150985532576,
      "grad_norm": 1.5813256350685223,
      "learning_rate": 1.3856462644004753e-07,
      "loss": 2.3149,
      "step": 6222
    },
    {
      "epoch": 0.8056445609606111,
      "grad_norm": 1.4358543420741046,
      "learning_rate": 1.3838730992098037e-07,
      "loss": 2.4104,
      "step": 6223
    },
    {
      "epoch": 0.8057740233679646,
      "grad_norm": 1.6910192863503737,
      "learning_rate": 1.3821009280241827e-07,
      "loss": 2.3755,
      "step": 6224
    },
    {
      "epoch": 0.805903485775318,
      "grad_norm": 1.6057479418021494,
      "learning_rate": 1.380329751205616e-07,
      "loss": 2.196,
      "step": 6225
    },
    {
      "epoch": 0.805903485775318,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972034692764282,
      "eval_runtime": 19.7285,
      "eval_samples_per_second": 2.23,
      "eval_steps_per_second": 0.152,
      "step": 6225
    },
    {
      "epoch": 0.8060329481826715,
      "grad_norm": 1.448748191424554,
      "learning_rate": 1.3785595691159018e-07,
      "loss": 2.6611,
      "step": 6226
    },
    {
      "epoch": 0.8061624105900249,
      "grad_norm": 1.30622228469084,
      "learning_rate": 1.376790382116633e-07,
      "loss": 2.2417,
      "step": 6227
    },
    {
      "epoch": 0.8062918729973784,
      "grad_norm": 1.1270459027688289,
      "learning_rate": 1.3750221905692065e-07,
      "loss": 2.2311,
      "step": 6228
    },
    {
      "epoch": 0.8064213354047318,
      "grad_norm": 1.235283774177082,
      "learning_rate": 1.3732549948348064e-07,
      "loss": 2.2146,
      "step": 6229
    },
    {
      "epoch": 0.8065507978120853,
      "grad_norm": 2.7447854042961697,
      "learning_rate": 1.3714887952744227e-07,
      "loss": 2.3113,
      "step": 6230
    },
    {
      "epoch": 0.8065507978120853,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.6147,
      "eval_samples_per_second": 3.011,
      "eval_steps_per_second": 0.205,
      "step": 6230
    },
    {
      "epoch": 0.8066802602194387,
      "grad_norm": 2.510033060336682,
      "learning_rate": 1.369723592248836e-07,
      "loss": 2.4973,
      "step": 6231
    },
    {
      "epoch": 0.8068097226267923,
      "grad_norm": 1.2585691302725435,
      "learning_rate": 1.3679593861186218e-07,
      "loss": 2.3667,
      "step": 6232
    },
    {
      "epoch": 0.8069391850341457,
      "grad_norm": 1.7096579423023155,
      "learning_rate": 1.3661961772441598e-07,
      "loss": 2.4365,
      "step": 6233
    },
    {
      "epoch": 0.8070686474414992,
      "grad_norm": 1.0830310726128318,
      "learning_rate": 1.364433965985617e-07,
      "loss": 2.2573,
      "step": 6234
    },
    {
      "epoch": 0.8071981098488527,
      "grad_norm": 2.5880546492111005,
      "learning_rate": 1.362672752702962e-07,
      "loss": 2.3967,
      "step": 6235
    },
    {
      "epoch": 0.8071981098488527,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5021750926971436,
      "eval_runtime": 14.4281,
      "eval_samples_per_second": 3.05,
      "eval_steps_per_second": 0.208,
      "step": 6235
    },
    {
      "epoch": 0.8073275722562061,
      "grad_norm": 1.410395699460445,
      "learning_rate": 1.360912537755963e-07,
      "loss": 2.259,
      "step": 6236
    },
    {
      "epoch": 0.8074570346635596,
      "grad_norm": 2.295976613966117,
      "learning_rate": 1.359153321504171e-07,
      "loss": 2.3083,
      "step": 6237
    },
    {
      "epoch": 0.807586497070913,
      "grad_norm": 1.642481760617068,
      "learning_rate": 1.3573951043069489e-07,
      "loss": 2.47,
      "step": 6238
    },
    {
      "epoch": 0.8077159594782665,
      "grad_norm": 1.5561927285141135,
      "learning_rate": 1.3556378865234428e-07,
      "loss": 2.3877,
      "step": 6239
    },
    {
      "epoch": 0.8078454218856199,
      "grad_norm": 1.43831349931753,
      "learning_rate": 1.353881668512603e-07,
      "loss": 2.4724,
      "step": 6240
    },
    {
      "epoch": 0.8078454218856199,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5000888109207153,
      "eval_runtime": 14.436,
      "eval_samples_per_second": 3.048,
      "eval_steps_per_second": 0.208,
      "step": 6240
    },
    {
      "epoch": 0.8079748842929735,
      "grad_norm": 1.57936985747477,
      "learning_rate": 1.3521264506331753e-07,
      "loss": 2.5044,
      "step": 6241
    },
    {
      "epoch": 0.8081043467003269,
      "grad_norm": 1.1381700417949072,
      "learning_rate": 1.350372233243693e-07,
      "loss": 2.2129,
      "step": 6242
    },
    {
      "epoch": 0.8082338091076804,
      "grad_norm": 1.5803923484470352,
      "learning_rate": 1.3486190167024917e-07,
      "loss": 2.4358,
      "step": 6243
    },
    {
      "epoch": 0.8083632715150338,
      "grad_norm": 1.4579328218002998,
      "learning_rate": 1.346866801367705e-07,
      "loss": 2.3018,
      "step": 6244
    },
    {
      "epoch": 0.8084927339223873,
      "grad_norm": 1.2979619486600702,
      "learning_rate": 1.3451155875972544e-07,
      "loss": 2.29,
      "step": 6245
    },
    {
      "epoch": 0.8084927339223873,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4993342161178589,
      "eval_runtime": 16.0826,
      "eval_samples_per_second": 2.736,
      "eval_steps_per_second": 0.187,
      "step": 6245
    },
    {
      "epoch": 0.8086221963297408,
      "grad_norm": 2.046648171025459,
      "learning_rate": 1.3433653757488638e-07,
      "loss": 2.5249,
      "step": 6246
    },
    {
      "epoch": 0.8087516587370942,
      "grad_norm": 1.3396283591167975,
      "learning_rate": 1.3416161661800486e-07,
      "loss": 2.2314,
      "step": 6247
    },
    {
      "epoch": 0.8088811211444477,
      "grad_norm": 1.4208138335817868,
      "learning_rate": 1.3398679592481187e-07,
      "loss": 2.4346,
      "step": 6248
    },
    {
      "epoch": 0.8090105835518011,
      "grad_norm": 1.9371182603836303,
      "learning_rate": 1.3381207553101828e-07,
      "loss": 2.3997,
      "step": 6249
    },
    {
      "epoch": 0.8091400459591546,
      "grad_norm": 1.5543758955550426,
      "learning_rate": 1.3363745547231405e-07,
      "loss": 2.2134,
      "step": 6250
    },
    {
      "epoch": 0.8091400459591546,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979802370071411,
      "eval_runtime": 14.1336,
      "eval_samples_per_second": 3.113,
      "eval_steps_per_second": 0.212,
      "step": 6250
    },
    {
      "epoch": 0.8092695083665081,
      "grad_norm": 2.0387405118195963,
      "learning_rate": 1.3346293578436912e-07,
      "loss": 2.4038,
      "step": 6251
    },
    {
      "epoch": 0.8093989707738616,
      "grad_norm": 2.2434466585684034,
      "learning_rate": 1.3328851650283263e-07,
      "loss": 2.6687,
      "step": 6252
    },
    {
      "epoch": 0.809528433181215,
      "grad_norm": 1.1207425435617016,
      "learning_rate": 1.33114197663333e-07,
      "loss": 2.1802,
      "step": 6253
    },
    {
      "epoch": 0.8096578955885685,
      "grad_norm": 1.3118458079804722,
      "learning_rate": 1.3293997930147872e-07,
      "loss": 2.283,
      "step": 6254
    },
    {
      "epoch": 0.8097873579959219,
      "grad_norm": 2.330512670064463,
      "learning_rate": 1.3276586145285715e-07,
      "loss": 2.479,
      "step": 6255
    },
    {
      "epoch": 0.8097873579959219,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.495805263519287,
      "eval_runtime": 14.6736,
      "eval_samples_per_second": 2.999,
      "eval_steps_per_second": 0.204,
      "step": 6255
    },
    {
      "epoch": 0.8099168204032754,
      "grad_norm": 2.219437394565759,
      "learning_rate": 1.3259184415303563e-07,
      "loss": 2.394,
      "step": 6256
    },
    {
      "epoch": 0.8100462828106288,
      "grad_norm": 2.431686666341144,
      "learning_rate": 1.324179274375604e-07,
      "loss": 2.2617,
      "step": 6257
    },
    {
      "epoch": 0.8101757452179823,
      "grad_norm": 1.5772358118334202,
      "learning_rate": 1.3224411134195785e-07,
      "loss": 2.332,
      "step": 6258
    },
    {
      "epoch": 0.8103052076253358,
      "grad_norm": 2.164202346401528,
      "learning_rate": 1.3207039590173332e-07,
      "loss": 2.3345,
      "step": 6259
    },
    {
      "epoch": 0.8104346700326892,
      "grad_norm": 1.8810949217795152,
      "learning_rate": 1.3189678115237152e-07,
      "loss": 2.5212,
      "step": 6260
    },
    {
      "epoch": 0.8104346700326892,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979137182235718,
      "eval_runtime": 14.1589,
      "eval_samples_per_second": 3.108,
      "eval_steps_per_second": 0.212,
      "step": 6260
    },
    {
      "epoch": 0.8105641324400428,
      "grad_norm": 3.115124798926522,
      "learning_rate": 1.3172326712933683e-07,
      "loss": 2.3335,
      "step": 6261
    },
    {
      "epoch": 0.8106935948473962,
      "grad_norm": 1.7171199322904265,
      "learning_rate": 1.315498538680733e-07,
      "loss": 2.4487,
      "step": 6262
    },
    {
      "epoch": 0.8108230572547497,
      "grad_norm": 2.9788977596153425,
      "learning_rate": 1.3137654140400372e-07,
      "loss": 2.4102,
      "step": 6263
    },
    {
      "epoch": 0.8109525196621031,
      "grad_norm": 1.958154536114569,
      "learning_rate": 1.3120332977253117e-07,
      "loss": 2.5662,
      "step": 6264
    },
    {
      "epoch": 0.8110819820694566,
      "grad_norm": 1.453856583763347,
      "learning_rate": 1.31030219009037e-07,
      "loss": 2.293,
      "step": 6265
    },
    {
      "epoch": 0.8110819820694566,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4965153932571411,
      "eval_runtime": 14.5295,
      "eval_samples_per_second": 3.028,
      "eval_steps_per_second": 0.206,
      "step": 6265
    },
    {
      "epoch": 0.81121144447681,
      "grad_norm": 1.4893094328906442,
      "learning_rate": 1.3085720914888286e-07,
      "loss": 2.5107,
      "step": 6266
    },
    {
      "epoch": 0.8113409068841635,
      "grad_norm": 4.425612894913207,
      "learning_rate": 1.3068430022740983e-07,
      "loss": 2.7544,
      "step": 6267
    },
    {
      "epoch": 0.811470369291517,
      "grad_norm": 1.7791384111662343,
      "learning_rate": 1.3051149227993754e-07,
      "loss": 2.4363,
      "step": 6268
    },
    {
      "epoch": 0.8115998316988704,
      "grad_norm": 1.4961801778060027,
      "learning_rate": 1.30338785341766e-07,
      "loss": 2.4866,
      "step": 6269
    },
    {
      "epoch": 0.8117292941062239,
      "grad_norm": 1.4726579342137567,
      "learning_rate": 1.3016617944817386e-07,
      "loss": 2.3291,
      "step": 6270
    },
    {
      "epoch": 0.8117292941062239,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5000888109207153,
      "eval_runtime": 15.2735,
      "eval_samples_per_second": 2.881,
      "eval_steps_per_second": 0.196,
      "step": 6270
    },
    {
      "epoch": 0.8118587565135774,
      "grad_norm": 1.5362615680151408,
      "learning_rate": 1.299936746344193e-07,
      "loss": 2.3953,
      "step": 6271
    },
    {
      "epoch": 0.8119882189209309,
      "grad_norm": 1.5962930472624859,
      "learning_rate": 1.2982127093574007e-07,
      "loss": 2.5796,
      "step": 6272
    },
    {
      "epoch": 0.8121176813282843,
      "grad_norm": 1.398740965334029,
      "learning_rate": 1.2964896838735295e-07,
      "loss": 2.3535,
      "step": 6273
    },
    {
      "epoch": 0.8122471437356378,
      "grad_norm": 1.7418081881335348,
      "learning_rate": 1.2947676702445455e-07,
      "loss": 2.3938,
      "step": 6274
    },
    {
      "epoch": 0.8123766061429912,
      "grad_norm": 2.046282092118504,
      "learning_rate": 1.2930466688222024e-07,
      "loss": 2.5754,
      "step": 6275
    },
    {
      "epoch": 0.8123766061429912,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957607984542847,
      "eval_runtime": 14.481,
      "eval_samples_per_second": 3.038,
      "eval_steps_per_second": 0.207,
      "step": 6275
    },
    {
      "epoch": 0.8125060685503447,
      "grad_norm": 2.334143625367805,
      "learning_rate": 1.291326679958048e-07,
      "loss": 2.5625,
      "step": 6276
    },
    {
      "epoch": 0.8126355309576981,
      "grad_norm": 2.4934035566855095,
      "learning_rate": 1.2896077040034295e-07,
      "loss": 2.8154,
      "step": 6277
    },
    {
      "epoch": 0.8127649933650516,
      "grad_norm": 1.5819090551958517,
      "learning_rate": 1.2878897413094776e-07,
      "loss": 2.3948,
      "step": 6278
    },
    {
      "epoch": 0.812894455772405,
      "grad_norm": 1.26988767492704,
      "learning_rate": 1.2861727922271257e-07,
      "loss": 2.3611,
      "step": 6279
    },
    {
      "epoch": 0.8130239181797585,
      "grad_norm": 2.0432384546681517,
      "learning_rate": 1.2844568571070927e-07,
      "loss": 2.2361,
      "step": 6280
    },
    {
      "epoch": 0.8130239181797585,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5008434057235718,
      "eval_runtime": 14.0386,
      "eval_samples_per_second": 3.134,
      "eval_steps_per_second": 0.214,
      "step": 6280
    },
    {
      "epoch": 0.8131533805871121,
      "grad_norm": 3.42417390692355,
      "learning_rate": 1.2827419362998924e-07,
      "loss": 2.8511,
      "step": 6281
    },
    {
      "epoch": 0.8132828429944655,
      "grad_norm": 3.3870531176022407,
      "learning_rate": 1.2810280301558341e-07,
      "loss": 2.2778,
      "step": 6282
    },
    {
      "epoch": 0.813412305401819,
      "grad_norm": 1.7018346516498715,
      "learning_rate": 1.2793151390250163e-07,
      "loss": 2.478,
      "step": 6283
    },
    {
      "epoch": 0.8135417678091724,
      "grad_norm": 2.3822412806207915,
      "learning_rate": 1.277603263257333e-07,
      "loss": 2.2595,
      "step": 6284
    },
    {
      "epoch": 0.8136712302165259,
      "grad_norm": 1.6448255526328308,
      "learning_rate": 1.2758924032024695e-07,
      "loss": 2.417,
      "step": 6285
    },
    {
      "epoch": 0.8136712302165259,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.2849,
      "eval_samples_per_second": 3.08,
      "eval_steps_per_second": 0.21,
      "step": 6285
    },
    {
      "epoch": 0.8138006926238793,
      "grad_norm": 3.1391670066861432,
      "learning_rate": 1.2741825592099037e-07,
      "loss": 2.4766,
      "step": 6286
    },
    {
      "epoch": 0.8139301550312328,
      "grad_norm": 1.9230502837718988,
      "learning_rate": 1.2724737316289058e-07,
      "loss": 2.3713,
      "step": 6287
    },
    {
      "epoch": 0.8140596174385862,
      "grad_norm": 1.838057915567865,
      "learning_rate": 1.2707659208085358e-07,
      "loss": 2.3872,
      "step": 6288
    },
    {
      "epoch": 0.8141890798459397,
      "grad_norm": 1.3936634498286318,
      "learning_rate": 1.2690591270976515e-07,
      "loss": 2.4773,
      "step": 6289
    },
    {
      "epoch": 0.8143185422532933,
      "grad_norm": 1.375837943148604,
      "learning_rate": 1.2673533508449008e-07,
      "loss": 2.3242,
      "step": 6290
    },
    {
      "epoch": 0.8143185422532933,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972256422042847,
      "eval_runtime": 14.1293,
      "eval_samples_per_second": 3.114,
      "eval_steps_per_second": 0.212,
      "step": 6290
    },
    {
      "epoch": 0.8144480046606467,
      "grad_norm": 1.5427035381192669,
      "learning_rate": 1.2656485923987223e-07,
      "loss": 2.3506,
      "step": 6291
    },
    {
      "epoch": 0.8145774670680002,
      "grad_norm": 1.1785225735342413,
      "learning_rate": 1.2639448521073446e-07,
      "loss": 2.521,
      "step": 6292
    },
    {
      "epoch": 0.8147069294753536,
      "grad_norm": 1.3263459914557014,
      "learning_rate": 1.2622421303187963e-07,
      "loss": 2.4124,
      "step": 6293
    },
    {
      "epoch": 0.8148363918827071,
      "grad_norm": 2.1394412925821373,
      "learning_rate": 1.2605404273808873e-07,
      "loss": 2.4636,
      "step": 6294
    },
    {
      "epoch": 0.8149658542900605,
      "grad_norm": 2.028831192488588,
      "learning_rate": 1.2588397436412306e-07,
      "loss": 2.4758,
      "step": 6295
    },
    {
      "epoch": 0.8149658542900605,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4994007349014282,
      "eval_runtime": 14.9777,
      "eval_samples_per_second": 2.938,
      "eval_steps_per_second": 0.2,
      "step": 6295
    },
    {
      "epoch": 0.815095316697414,
      "grad_norm": 1.9907033736587316,
      "learning_rate": 1.25714007944722e-07,
      "loss": 2.2,
      "step": 6296
    },
    {
      "epoch": 0.8152247791047674,
      "grad_norm": 1.0472666627340403,
      "learning_rate": 1.2554414351460507e-07,
      "loss": 2.3074,
      "step": 6297
    },
    {
      "epoch": 0.8153542415121209,
      "grad_norm": 2.9257359065708703,
      "learning_rate": 1.2537438110847033e-07,
      "loss": 2.5596,
      "step": 6298
    },
    {
      "epoch": 0.8154837039194743,
      "grad_norm": 1.2165676070447224,
      "learning_rate": 1.25204720760995e-07,
      "loss": 2.2349,
      "step": 6299
    },
    {
      "epoch": 0.8156131663268279,
      "grad_norm": 1.4149763568977256,
      "learning_rate": 1.2503516250683607e-07,
      "loss": 2.3835,
      "step": 6300
    },
    {
      "epoch": 0.8156131663268279,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957164525985718,
      "eval_runtime": 14.0742,
      "eval_samples_per_second": 3.126,
      "eval_steps_per_second": 0.213,
      "step": 6300
    },
    {
      "epoch": 0.8157426287341814,
      "grad_norm": 3.1148233771089364,
      "learning_rate": 1.2486570638062883e-07,
      "loss": 2.3169,
      "step": 6301
    },
    {
      "epoch": 0.8158720911415348,
      "grad_norm": 1.8478945745385917,
      "learning_rate": 1.246963524169885e-07,
      "loss": 2.3478,
      "step": 6302
    },
    {
      "epoch": 0.8160015535488883,
      "grad_norm": 3.208744695371423,
      "learning_rate": 1.2452710065050887e-07,
      "loss": 2.4253,
      "step": 6303
    },
    {
      "epoch": 0.8161310159562417,
      "grad_norm": 2.084172645021899,
      "learning_rate": 1.243579511157629e-07,
      "loss": 2.3889,
      "step": 6304
    },
    {
      "epoch": 0.8162604783635952,
      "grad_norm": 2.062815310308983,
      "learning_rate": 1.2418890384730312e-07,
      "loss": 2.4033,
      "step": 6305
    },
    {
      "epoch": 0.8162604783635952,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4958274364471436,
      "eval_runtime": 13.7397,
      "eval_samples_per_second": 3.202,
      "eval_steps_per_second": 0.218,
      "step": 6305
    },
    {
      "epoch": 0.8163899407709486,
      "grad_norm": 1.319709058617649,
      "learning_rate": 1.2401995887966055e-07,
      "loss": 2.4094,
      "step": 6306
    },
    {
      "epoch": 0.8165194031783021,
      "grad_norm": 1.4474955083374486,
      "learning_rate": 1.238511162473458e-07,
      "loss": 2.3557,
      "step": 6307
    },
    {
      "epoch": 0.8166488655856555,
      "grad_norm": 2.026216934287568,
      "learning_rate": 1.2368237598484872e-07,
      "loss": 2.3748,
      "step": 6308
    },
    {
      "epoch": 0.816778327993009,
      "grad_norm": 1.6297930342281466,
      "learning_rate": 1.235137381266372e-07,
      "loss": 2.1533,
      "step": 6309
    },
    {
      "epoch": 0.8169077904003625,
      "grad_norm": 2.292826571351977,
      "learning_rate": 1.2334520270715953e-07,
      "loss": 2.4448,
      "step": 6310
    },
    {
      "epoch": 0.8169077904003625,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986461400985718,
      "eval_runtime": 14.1434,
      "eval_samples_per_second": 3.111,
      "eval_steps_per_second": 0.212,
      "step": 6310
    },
    {
      "epoch": 0.817037252807716,
      "grad_norm": 1.474053267399202,
      "learning_rate": 1.2317676976084213e-07,
      "loss": 2.1426,
      "step": 6311
    },
    {
      "epoch": 0.8171667152150695,
      "grad_norm": 3.198823635565862,
      "learning_rate": 1.2300843932209102e-07,
      "loss": 2.4983,
      "step": 6312
    },
    {
      "epoch": 0.8172961776224229,
      "grad_norm": 1.6076498391802823,
      "learning_rate": 1.228402114252912e-07,
      "loss": 2.1465,
      "step": 6313
    },
    {
      "epoch": 0.8174256400297764,
      "grad_norm": 1.918009796029203,
      "learning_rate": 1.2267208610480656e-07,
      "loss": 2.3655,
      "step": 6314
    },
    {
      "epoch": 0.8175551024371298,
      "grad_norm": 1.2798263326878114,
      "learning_rate": 1.2250406339497984e-07,
      "loss": 2.2139,
      "step": 6315
    },
    {
      "epoch": 0.8175551024371298,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5001553297042847,
      "eval_runtime": 14.0525,
      "eval_samples_per_second": 3.131,
      "eval_steps_per_second": 0.213,
      "step": 6315
    },
    {
      "epoch": 0.8176845648444833,
      "grad_norm": 3.2846196386402675,
      "learning_rate": 1.2233614333013343e-07,
      "loss": 2.2935,
      "step": 6316
    },
    {
      "epoch": 0.8178140272518367,
      "grad_norm": 2.643592477792786,
      "learning_rate": 1.221683259445682e-07,
      "loss": 2.375,
      "step": 6317
    },
    {
      "epoch": 0.8179434896591902,
      "grad_norm": 2.0023751912121788,
      "learning_rate": 1.2200061127256448e-07,
      "loss": 2.1931,
      "step": 6318
    },
    {
      "epoch": 0.8180729520665436,
      "grad_norm": 1.7552013474575578,
      "learning_rate": 1.2183299934838118e-07,
      "loss": 2.3896,
      "step": 6319
    },
    {
      "epoch": 0.8182024144738972,
      "grad_norm": 2.3558165674252574,
      "learning_rate": 1.216654902062564e-07,
      "loss": 2.6265,
      "step": 6320
    },
    {
      "epoch": 0.8182024144738972,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4980690479278564,
      "eval_runtime": 14.8566,
      "eval_samples_per_second": 2.962,
      "eval_steps_per_second": 0.202,
      "step": 6320
    },
    {
      "epoch": 0.8183318768812506,
      "grad_norm": 1.4932324440789273,
      "learning_rate": 1.214980838804075e-07,
      "loss": 2.3643,
      "step": 6321
    },
    {
      "epoch": 0.8184613392886041,
      "grad_norm": 2.5255086955296844,
      "learning_rate": 1.213307804050304e-07,
      "loss": 2.3972,
      "step": 6322
    },
    {
      "epoch": 0.8185908016959575,
      "grad_norm": 4.728126543569728,
      "learning_rate": 1.2116357981430036e-07,
      "loss": 2.9607,
      "step": 6323
    },
    {
      "epoch": 0.818720264103311,
      "grad_norm": 1.7559297891187966,
      "learning_rate": 1.2099648214237134e-07,
      "loss": 2.3367,
      "step": 6324
    },
    {
      "epoch": 0.8188497265106645,
      "grad_norm": 1.5159677540468217,
      "learning_rate": 1.2082948742337668e-07,
      "loss": 2.2708,
      "step": 6325
    },
    {
      "epoch": 0.8188497265106645,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5014426708221436,
      "eval_runtime": 14.8541,
      "eval_samples_per_second": 2.962,
      "eval_steps_per_second": 0.202,
      "step": 6325
    },
    {
      "epoch": 0.8189791889180179,
      "grad_norm": 2.1311950130000246,
      "learning_rate": 1.2066259569142827e-07,
      "loss": 2.4731,
      "step": 6326
    },
    {
      "epoch": 0.8191086513253714,
      "grad_norm": 1.4084495877590246,
      "learning_rate": 1.2049580698061696e-07,
      "loss": 2.4939,
      "step": 6327
    },
    {
      "epoch": 0.8192381137327248,
      "grad_norm": 3.383438293522866,
      "learning_rate": 1.203291213250131e-07,
      "loss": 2.4492,
      "step": 6328
    },
    {
      "epoch": 0.8193675761400784,
      "grad_norm": 2.6568288185072153,
      "learning_rate": 1.2016253875866532e-07,
      "loss": 2.5332,
      "step": 6329
    },
    {
      "epoch": 0.8194970385474318,
      "grad_norm": 1.5536524314424307,
      "learning_rate": 1.1999605931560166e-07,
      "loss": 2.2972,
      "step": 6330
    },
    {
      "epoch": 0.8194970385474318,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4992009401321411,
      "eval_runtime": 13.7537,
      "eval_samples_per_second": 3.199,
      "eval_steps_per_second": 0.218,
      "step": 6330
    },
    {
      "epoch": 0.8196265009547853,
      "grad_norm": 1.2488211101956501,
      "learning_rate": 1.1982968302982887e-07,
      "loss": 2.4124,
      "step": 6331
    },
    {
      "epoch": 0.8197559633621387,
      "grad_norm": 1.203962667877106,
      "learning_rate": 1.1966340993533256e-07,
      "loss": 2.3562,
      "step": 6332
    },
    {
      "epoch": 0.8198854257694922,
      "grad_norm": 2.14800605479246,
      "learning_rate": 1.1949724006607752e-07,
      "loss": 2.3933,
      "step": 6333
    },
    {
      "epoch": 0.8200148881768456,
      "grad_norm": 1.4157104140600718,
      "learning_rate": 1.1933117345600722e-07,
      "loss": 2.2075,
      "step": 6334
    },
    {
      "epoch": 0.8201443505841991,
      "grad_norm": 1.5516360090138508,
      "learning_rate": 1.1916521013904413e-07,
      "loss": 2.2917,
      "step": 6335
    },
    {
      "epoch": 0.8201443505841991,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500776767730713,
      "eval_runtime": 13.8856,
      "eval_samples_per_second": 3.169,
      "eval_steps_per_second": 0.216,
      "step": 6335
    },
    {
      "epoch": 0.8202738129915526,
      "grad_norm": 1.6796408387198547,
      "learning_rate": 1.1899935014908999e-07,
      "loss": 2.3203,
      "step": 6336
    },
    {
      "epoch": 0.820403275398906,
      "grad_norm": 1.8900740461739876,
      "learning_rate": 1.1883359352002445e-07,
      "loss": 2.384,
      "step": 6337
    },
    {
      "epoch": 0.8205327378062595,
      "grad_norm": 1.667028364745508,
      "learning_rate": 1.1866794028570695e-07,
      "loss": 2.2772,
      "step": 6338
    },
    {
      "epoch": 0.820662200213613,
      "grad_norm": 1.9198259530486241,
      "learning_rate": 1.185023904799757e-07,
      "loss": 2.3115,
      "step": 6339
    },
    {
      "epoch": 0.8207916626209665,
      "grad_norm": 1.7635794553885937,
      "learning_rate": 1.183369441366472e-07,
      "loss": 2.3157,
      "step": 6340
    },
    {
      "epoch": 0.8207916626209665,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5022639036178589,
      "eval_runtime": 14.2424,
      "eval_samples_per_second": 3.089,
      "eval_steps_per_second": 0.211,
      "step": 6340
    },
    {
      "epoch": 0.8209211250283199,
      "grad_norm": 2.521564124682436,
      "learning_rate": 1.1817160128951774e-07,
      "loss": 2.4697,
      "step": 6341
    },
    {
      "epoch": 0.8210505874356734,
      "grad_norm": 1.992318842451766,
      "learning_rate": 1.1800636197236155e-07,
      "loss": 2.2249,
      "step": 6342
    },
    {
      "epoch": 0.8211800498430268,
      "grad_norm": 5.126400349138914,
      "learning_rate": 1.1784122621893213e-07,
      "loss": 2.626,
      "step": 6343
    },
    {
      "epoch": 0.8213095122503803,
      "grad_norm": 2.7019548229363672,
      "learning_rate": 1.1767619406296199e-07,
      "loss": 2.5398,
      "step": 6344
    },
    {
      "epoch": 0.8214389746577337,
      "grad_norm": 3.0664055281732048,
      "learning_rate": 1.175112655381621e-07,
      "loss": 2.2334,
      "step": 6345
    },
    {
      "epoch": 0.8214389746577337,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972478151321411,
      "eval_runtime": 14.8137,
      "eval_samples_per_second": 2.97,
      "eval_steps_per_second": 0.203,
      "step": 6345
    },
    {
      "epoch": 0.8215684370650872,
      "grad_norm": 1.5686917620892498,
      "learning_rate": 1.1734644067822271e-07,
      "loss": 2.2732,
      "step": 6346
    },
    {
      "epoch": 0.8216978994724407,
      "grad_norm": 3.5494625171075658,
      "learning_rate": 1.1718171951681241e-07,
      "loss": 2.5874,
      "step": 6347
    },
    {
      "epoch": 0.8218273618797941,
      "grad_norm": 1.7442651376418326,
      "learning_rate": 1.1701710208757885e-07,
      "loss": 2.1843,
      "step": 6348
    },
    {
      "epoch": 0.8219568242871477,
      "grad_norm": 5.105265885110529,
      "learning_rate": 1.1685258842414858e-07,
      "loss": 2.3115,
      "step": 6349
    },
    {
      "epoch": 0.8220862866945011,
      "grad_norm": 2.269954966843938,
      "learning_rate": 1.1668817856012669e-07,
      "loss": 2.4819,
      "step": 6350
    },
    {
      "epoch": 0.8220862866945011,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986461400985718,
      "eval_runtime": 14.3959,
      "eval_samples_per_second": 3.056,
      "eval_steps_per_second": 0.208,
      "step": 6350
    },
    {
      "epoch": 0.8222157491018546,
      "grad_norm": 1.7022721319988725,
      "learning_rate": 1.1652387252909757e-07,
      "loss": 2.3052,
      "step": 6351
    },
    {
      "epoch": 0.822345211509208,
      "grad_norm": 1.8262557412178944,
      "learning_rate": 1.163596703646236e-07,
      "loss": 2.3816,
      "step": 6352
    },
    {
      "epoch": 0.8224746739165615,
      "grad_norm": 2.43558720565118,
      "learning_rate": 1.1619557210024681e-07,
      "loss": 2.6401,
      "step": 6353
    },
    {
      "epoch": 0.8226041363239149,
      "grad_norm": 2.24869568377646,
      "learning_rate": 1.1603157776948739e-07,
      "loss": 2.8328,
      "step": 6354
    },
    {
      "epoch": 0.8227335987312684,
      "grad_norm": 1.8826967942782387,
      "learning_rate": 1.158676874058444e-07,
      "loss": 2.3718,
      "step": 6355
    },
    {
      "epoch": 0.8227335987312684,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4994229078292847,
      "eval_runtime": 14.6804,
      "eval_samples_per_second": 2.997,
      "eval_steps_per_second": 0.204,
      "step": 6355
    },
    {
      "epoch": 0.8228630611386218,
      "grad_norm": 1.4439797336157976,
      "learning_rate": 1.1570390104279586e-07,
      "loss": 2.4099,
      "step": 6356
    },
    {
      "epoch": 0.8229925235459753,
      "grad_norm": 2.110129876706644,
      "learning_rate": 1.1554021871379863e-07,
      "loss": 2.3296,
      "step": 6357
    },
    {
      "epoch": 0.8231219859533287,
      "grad_norm": 2.8215050398732227,
      "learning_rate": 1.1537664045228811e-07,
      "loss": 2.5105,
      "step": 6358
    },
    {
      "epoch": 0.8232514483606823,
      "grad_norm": 2.0882115659103486,
      "learning_rate": 1.1521316629167826e-07,
      "loss": 2.4209,
      "step": 6359
    },
    {
      "epoch": 0.8233809107680358,
      "grad_norm": 1.9000026590855885,
      "learning_rate": 1.1504979626536204e-07,
      "loss": 2.3137,
      "step": 6360
    },
    {
      "epoch": 0.8233809107680358,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4965598583221436,
      "eval_runtime": 14.0624,
      "eval_samples_per_second": 3.129,
      "eval_steps_per_second": 0.213,
      "step": 6360
    },
    {
      "epoch": 0.8235103731753892,
      "grad_norm": 3.5913305057742786,
      "learning_rate": 1.1488653040671108e-07,
      "loss": 2.6326,
      "step": 6361
    },
    {
      "epoch": 0.8236398355827427,
      "grad_norm": 1.3011725229583542,
      "learning_rate": 1.1472336874907598e-07,
      "loss": 2.3154,
      "step": 6362
    },
    {
      "epoch": 0.8237692979900961,
      "grad_norm": 1.534195125871146,
      "learning_rate": 1.1456031132578551e-07,
      "loss": 2.4561,
      "step": 6363
    },
    {
      "epoch": 0.8238987603974496,
      "grad_norm": 1.5063020634263042,
      "learning_rate": 1.1439735817014783e-07,
      "loss": 2.6113,
      "step": 6364
    },
    {
      "epoch": 0.824028222804803,
      "grad_norm": 2.61783421848311,
      "learning_rate": 1.1423450931544907e-07,
      "loss": 2.4829,
      "step": 6365
    },
    {
      "epoch": 0.824028222804803,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972478151321411,
      "eval_runtime": 14.2782,
      "eval_samples_per_second": 3.082,
      "eval_steps_per_second": 0.21,
      "step": 6365
    },
    {
      "epoch": 0.8241576852121565,
      "grad_norm": 1.3929034740784199,
      "learning_rate": 1.1407176479495442e-07,
      "loss": 2.1797,
      "step": 6366
    },
    {
      "epoch": 0.8242871476195099,
      "grad_norm": 1.4291142234319245,
      "learning_rate": 1.1390912464190801e-07,
      "loss": 2.5068,
      "step": 6367
    },
    {
      "epoch": 0.8244166100268634,
      "grad_norm": 2.1614160281194974,
      "learning_rate": 1.1374658888953206e-07,
      "loss": 2.3751,
      "step": 6368
    },
    {
      "epoch": 0.824546072434217,
      "grad_norm": 1.1936741709563778,
      "learning_rate": 1.1358415757102816e-07,
      "loss": 2.3489,
      "step": 6369
    },
    {
      "epoch": 0.8246755348415704,
      "grad_norm": 2.5059238635481105,
      "learning_rate": 1.1342183071957593e-07,
      "loss": 2.3833,
      "step": 6370
    },
    {
      "epoch": 0.8246755348415704,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4951393604278564,
      "eval_runtime": 14.8639,
      "eval_samples_per_second": 2.96,
      "eval_steps_per_second": 0.202,
      "step": 6370
    },
    {
      "epoch": 0.8248049972489239,
      "grad_norm": 2.0832520132129795,
      "learning_rate": 1.1325960836833391e-07,
      "loss": 2.0908,
      "step": 6371
    },
    {
      "epoch": 0.8249344596562773,
      "grad_norm": 1.3783147514475196,
      "learning_rate": 1.130974905504395e-07,
      "loss": 2.4692,
      "step": 6372
    },
    {
      "epoch": 0.8250639220636308,
      "grad_norm": 1.8844940682759106,
      "learning_rate": 1.1293547729900829e-07,
      "loss": 2.1514,
      "step": 6373
    },
    {
      "epoch": 0.8251933844709842,
      "grad_norm": 1.2574216983158233,
      "learning_rate": 1.1277356864713506e-07,
      "loss": 2.2532,
      "step": 6374
    },
    {
      "epoch": 0.8253228468783377,
      "grad_norm": 1.2285115747458222,
      "learning_rate": 1.1261176462789277e-07,
      "loss": 2.3296,
      "step": 6375
    },
    {
      "epoch": 0.8253228468783377,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5007990598678589,
      "eval_runtime": 14.5752,
      "eval_samples_per_second": 3.019,
      "eval_steps_per_second": 0.206,
      "step": 6375
    },
    {
      "epoch": 0.8254523092856911,
      "grad_norm": 1.6530484342683356,
      "learning_rate": 1.1245006527433308e-07,
      "loss": 2.3611,
      "step": 6376
    },
    {
      "epoch": 0.8255817716930446,
      "grad_norm": 2.0151341225551045,
      "learning_rate": 1.1228847061948663e-07,
      "loss": 2.2935,
      "step": 6377
    },
    {
      "epoch": 0.8257112341003982,
      "grad_norm": 2.4675896195890084,
      "learning_rate": 1.1212698069636208e-07,
      "loss": 2.366,
      "step": 6378
    },
    {
      "epoch": 0.8258406965077516,
      "grad_norm": 2.221841954849348,
      "learning_rate": 1.1196559553794721e-07,
      "loss": 2.45,
      "step": 6379
    },
    {
      "epoch": 0.8259701589151051,
      "grad_norm": 1.4876343686037203,
      "learning_rate": 1.1180431517720844e-07,
      "loss": 2.4241,
      "step": 6380
    },
    {
      "epoch": 0.8259701589151051,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.495849609375,
      "eval_runtime": 13.978,
      "eval_samples_per_second": 3.148,
      "eval_steps_per_second": 0.215,
      "step": 6380
    },
    {
      "epoch": 0.8260996213224585,
      "grad_norm": 1.9454769615917034,
      "learning_rate": 1.1164313964709003e-07,
      "loss": 2.4382,
      "step": 6381
    },
    {
      "epoch": 0.826229083729812,
      "grad_norm": 1.4187495933312828,
      "learning_rate": 1.1148206898051572e-07,
      "loss": 2.4478,
      "step": 6382
    },
    {
      "epoch": 0.8263585461371654,
      "grad_norm": 1.7853762717898694,
      "learning_rate": 1.1132110321038725e-07,
      "loss": 2.291,
      "step": 6383
    },
    {
      "epoch": 0.8264880085445189,
      "grad_norm": 2.457562277746307,
      "learning_rate": 1.1116024236958525e-07,
      "loss": 2.4897,
      "step": 6384
    },
    {
      "epoch": 0.8266174709518723,
      "grad_norm": 1.5155581734655637,
      "learning_rate": 1.1099948649096894e-07,
      "loss": 2.4282,
      "step": 6385
    },
    {
      "epoch": 0.8266174709518723,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979137182235718,
      "eval_runtime": 14.3687,
      "eval_samples_per_second": 3.062,
      "eval_steps_per_second": 0.209,
      "step": 6385
    },
    {
      "epoch": 0.8267469333592258,
      "grad_norm": 2.6334341768112672,
      "learning_rate": 1.1083883560737585e-07,
      "loss": 2.4365,
      "step": 6386
    },
    {
      "epoch": 0.8268763957665792,
      "grad_norm": 2.5672186012764358,
      "learning_rate": 1.1067828975162196e-07,
      "loss": 2.7646,
      "step": 6387
    },
    {
      "epoch": 0.8270058581739328,
      "grad_norm": 1.6284770364380279,
      "learning_rate": 1.1051784895650238e-07,
      "loss": 2.3625,
      "step": 6388
    },
    {
      "epoch": 0.8271353205812862,
      "grad_norm": 1.4061592856712424,
      "learning_rate": 1.1035751325479009e-07,
      "loss": 2.2429,
      "step": 6389
    },
    {
      "epoch": 0.8272647829886397,
      "grad_norm": 2.4616930379219726,
      "learning_rate": 1.1019728267923718e-07,
      "loss": 2.3237,
      "step": 6390
    },
    {
      "epoch": 0.8272647829886397,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4956942796707153,
      "eval_runtime": 13.445,
      "eval_samples_per_second": 3.273,
      "eval_steps_per_second": 0.223,
      "step": 6390
    },
    {
      "epoch": 0.8273942453959932,
      "grad_norm": 1.5232611395883022,
      "learning_rate": 1.1003715726257372e-07,
      "loss": 2.4048,
      "step": 6391
    },
    {
      "epoch": 0.8275237078033466,
      "grad_norm": 1.8488047209916623,
      "learning_rate": 1.0987713703750887e-07,
      "loss": 2.3242,
      "step": 6392
    },
    {
      "epoch": 0.8276531702107001,
      "grad_norm": 2.23346436212358,
      "learning_rate": 1.0971722203672985e-07,
      "loss": 2.2783,
      "step": 6393
    },
    {
      "epoch": 0.8277826326180535,
      "grad_norm": 1.5317558912886233,
      "learning_rate": 1.0955741229290238e-07,
      "loss": 2.6125,
      "step": 6394
    },
    {
      "epoch": 0.827912095025407,
      "grad_norm": 1.4226477397582635,
      "learning_rate": 1.0939770783867112e-07,
      "loss": 2.4043,
      "step": 6395
    },
    {
      "epoch": 0.827912095025407,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972256422042847,
      "eval_runtime": 13.0766,
      "eval_samples_per_second": 3.365,
      "eval_steps_per_second": 0.229,
      "step": 6395
    },
    {
      "epoch": 0.8280415574327604,
      "grad_norm": 2.6032292403679276,
      "learning_rate": 1.0923810870665873e-07,
      "loss": 2.522,
      "step": 6396
    },
    {
      "epoch": 0.8281710198401139,
      "grad_norm": 1.5169909012805793,
      "learning_rate": 1.0907861492946676e-07,
      "loss": 2.3389,
      "step": 6397
    },
    {
      "epoch": 0.8283004822474674,
      "grad_norm": 1.1286836291653919,
      "learning_rate": 1.089192265396749e-07,
      "loss": 2.2214,
      "step": 6398
    },
    {
      "epoch": 0.8284299446548209,
      "grad_norm": 1.5813920773612495,
      "learning_rate": 1.0875994356984145e-07,
      "loss": 2.4224,
      "step": 6399
    },
    {
      "epoch": 0.8285594070621743,
      "grad_norm": 1.6031188879209641,
      "learning_rate": 1.0860076605250333e-07,
      "loss": 2.3916,
      "step": 6400
    },
    {
      "epoch": 0.8285594070621743,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986904859542847,
      "eval_runtime": 13.0961,
      "eval_samples_per_second": 3.36,
      "eval_steps_per_second": 0.229,
      "step": 6400
    },
    {
      "epoch": 0.8286888694695278,
      "grad_norm": 4.133883665012408,
      "learning_rate": 1.0844169402017545e-07,
      "loss": 2.8918,
      "step": 6401
    },
    {
      "epoch": 0.8288183318768813,
      "grad_norm": 3.50089292383131,
      "learning_rate": 1.0828272750535184e-07,
      "loss": 2.1743,
      "step": 6402
    },
    {
      "epoch": 0.8289477942842347,
      "grad_norm": 1.792265394416634,
      "learning_rate": 1.0812386654050481e-07,
      "loss": 2.3909,
      "step": 6403
    },
    {
      "epoch": 0.8290772566915882,
      "grad_norm": 1.1207067169576135,
      "learning_rate": 1.0796511115808428e-07,
      "loss": 2.2256,
      "step": 6404
    },
    {
      "epoch": 0.8292067190989416,
      "grad_norm": 1.7885467551223386,
      "learning_rate": 1.0780646139051984e-07,
      "loss": 2.6963,
      "step": 6405
    },
    {
      "epoch": 0.8292067190989416,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972034692764282,
      "eval_runtime": 15.8546,
      "eval_samples_per_second": 2.775,
      "eval_steps_per_second": 0.189,
      "step": 6405
    },
    {
      "epoch": 0.8293361815062951,
      "grad_norm": 2.245053247878789,
      "learning_rate": 1.076479172702185e-07,
      "loss": 2.5806,
      "step": 6406
    },
    {
      "epoch": 0.8294656439136485,
      "grad_norm": 1.252553684160652,
      "learning_rate": 1.0748947882956627e-07,
      "loss": 2.4224,
      "step": 6407
    },
    {
      "epoch": 0.8295951063210021,
      "grad_norm": 1.4356691870665468,
      "learning_rate": 1.0733114610092777e-07,
      "loss": 2.3816,
      "step": 6408
    },
    {
      "epoch": 0.8297245687283555,
      "grad_norm": 1.6939817722741666,
      "learning_rate": 1.0717291911664512e-07,
      "loss": 2.3059,
      "step": 6409
    },
    {
      "epoch": 0.829854031135709,
      "grad_norm": 1.8007081509821632,
      "learning_rate": 1.0701479790903954e-07,
      "loss": 2.4708,
      "step": 6410
    },
    {
      "epoch": 0.829854031135709,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4993785619735718,
      "eval_runtime": 14.2826,
      "eval_samples_per_second": 3.081,
      "eval_steps_per_second": 0.21,
      "step": 6410
    },
    {
      "epoch": 0.8299834935430624,
      "grad_norm": 1.5954342764989775,
      "learning_rate": 1.0685678251041074e-07,
      "loss": 2.196,
      "step": 6411
    },
    {
      "epoch": 0.8301129559504159,
      "grad_norm": 1.509963918739043,
      "learning_rate": 1.0669887295303626e-07,
      "loss": 2.5137,
      "step": 6412
    },
    {
      "epoch": 0.8302424183577694,
      "grad_norm": 2.453019985642702,
      "learning_rate": 1.0654106926917263e-07,
      "loss": 2.3425,
      "step": 6413
    },
    {
      "epoch": 0.8303718807651228,
      "grad_norm": 1.5370109058625236,
      "learning_rate": 1.0638337149105421e-07,
      "loss": 2.5103,
      "step": 6414
    },
    {
      "epoch": 0.8305013431724763,
      "grad_norm": 1.9255064265127184,
      "learning_rate": 1.0622577965089399e-07,
      "loss": 2.3008,
      "step": 6415
    },
    {
      "epoch": 0.8305013431724763,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5000888109207153,
      "eval_runtime": 13.7882,
      "eval_samples_per_second": 3.191,
      "eval_steps_per_second": 0.218,
      "step": 6415
    },
    {
      "epoch": 0.8306308055798297,
      "grad_norm": 2.1989925651908173,
      "learning_rate": 1.0606829378088344e-07,
      "loss": 2.104,
      "step": 6416
    },
    {
      "epoch": 0.8307602679871833,
      "grad_norm": 1.6521423458566424,
      "learning_rate": 1.0591091391319203e-07,
      "loss": 2.3623,
      "step": 6417
    },
    {
      "epoch": 0.8308897303945367,
      "grad_norm": 1.7258684115741765,
      "learning_rate": 1.0575364007996802e-07,
      "loss": 2.3477,
      "step": 6418
    },
    {
      "epoch": 0.8310191928018902,
      "grad_norm": 1.6109080646447245,
      "learning_rate": 1.055964723133377e-07,
      "loss": 2.4294,
      "step": 6419
    },
    {
      "epoch": 0.8311486552092436,
      "grad_norm": 1.5139435020263514,
      "learning_rate": 1.0543941064540562e-07,
      "loss": 2.3738,
      "step": 6420
    },
    {
      "epoch": 0.8311486552092436,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972034692764282,
      "eval_runtime": 14.2108,
      "eval_samples_per_second": 3.096,
      "eval_steps_per_second": 0.211,
      "step": 6420
    },
    {
      "epoch": 0.8312781176165971,
      "grad_norm": 1.6740019597672209,
      "learning_rate": 1.0528245510825499e-07,
      "loss": 2.3313,
      "step": 6421
    },
    {
      "epoch": 0.8314075800239505,
      "grad_norm": 3.340182350999748,
      "learning_rate": 1.0512560573394704e-07,
      "loss": 2.5273,
      "step": 6422
    },
    {
      "epoch": 0.831537042431304,
      "grad_norm": 2.323489780503788,
      "learning_rate": 1.0496886255452157e-07,
      "loss": 2.5625,
      "step": 6423
    },
    {
      "epoch": 0.8316665048386574,
      "grad_norm": 1.112693611364247,
      "learning_rate": 1.0481222560199636e-07,
      "loss": 2.178,
      "step": 6424
    },
    {
      "epoch": 0.8317959672460109,
      "grad_norm": 2.0998704426055372,
      "learning_rate": 1.0465569490836796e-07,
      "loss": 2.3674,
      "step": 6425
    },
    {
      "epoch": 0.8317959672460109,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950505495071411,
      "eval_runtime": 14.302,
      "eval_samples_per_second": 3.077,
      "eval_steps_per_second": 0.21,
      "step": 6425
    },
    {
      "epoch": 0.8319254296533644,
      "grad_norm": 1.40469361930793,
      "learning_rate": 1.0449927050561075e-07,
      "loss": 2.5125,
      "step": 6426
    },
    {
      "epoch": 0.8320548920607179,
      "grad_norm": 3.6172029908607652,
      "learning_rate": 1.0434295242567747e-07,
      "loss": 2.3716,
      "step": 6427
    },
    {
      "epoch": 0.8321843544680714,
      "grad_norm": 2.4029236695488843,
      "learning_rate": 1.0418674070049941e-07,
      "loss": 2.1892,
      "step": 6428
    },
    {
      "epoch": 0.8323138168754248,
      "grad_norm": 2.711887532567513,
      "learning_rate": 1.0403063536198611e-07,
      "loss": 2.9238,
      "step": 6429
    },
    {
      "epoch": 0.8324432792827783,
      "grad_norm": 2.186038441931193,
      "learning_rate": 1.0387463644202491e-07,
      "loss": 2.1677,
      "step": 6430
    },
    {
      "epoch": 0.8324432792827783,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.497114658355713,
      "eval_runtime": 13.6863,
      "eval_samples_per_second": 3.215,
      "eval_steps_per_second": 0.219,
      "step": 6430
    },
    {
      "epoch": 0.8325727416901317,
      "grad_norm": 1.3687318854130055,
      "learning_rate": 1.0371874397248231e-07,
      "loss": 2.45,
      "step": 6431
    },
    {
      "epoch": 0.8327022040974852,
      "grad_norm": 1.312477260591721,
      "learning_rate": 1.0356295798520182e-07,
      "loss": 2.4282,
      "step": 6432
    },
    {
      "epoch": 0.8328316665048386,
      "grad_norm": 1.025468656356217,
      "learning_rate": 1.0340727851200619e-07,
      "loss": 2.238,
      "step": 6433
    },
    {
      "epoch": 0.8329611289121921,
      "grad_norm": 1.4178525785579699,
      "learning_rate": 1.0325170558469627e-07,
      "loss": 2.2163,
      "step": 6434
    },
    {
      "epoch": 0.8330905913195455,
      "grad_norm": 1.4385261104497653,
      "learning_rate": 1.0309623923505076e-07,
      "loss": 2.4458,
      "step": 6435
    },
    {
      "epoch": 0.8330905913195455,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978693723678589,
      "eval_runtime": 14.7108,
      "eval_samples_per_second": 2.991,
      "eval_steps_per_second": 0.204,
      "step": 6435
    },
    {
      "epoch": 0.833220053726899,
      "grad_norm": 2.2515834221024322,
      "learning_rate": 1.0294087949482695e-07,
      "loss": 2.2119,
      "step": 6436
    },
    {
      "epoch": 0.8333495161342526,
      "grad_norm": 5.180428223063512,
      "learning_rate": 1.0278562639576023e-07,
      "loss": 2.3623,
      "step": 6437
    },
    {
      "epoch": 0.833478978541606,
      "grad_norm": 2.2399196869550106,
      "learning_rate": 1.026304799695639e-07,
      "loss": 2.3987,
      "step": 6438
    },
    {
      "epoch": 0.8336084409489595,
      "grad_norm": 3.975352821810764,
      "learning_rate": 1.024754402479302e-07,
      "loss": 2.229,
      "step": 6439
    },
    {
      "epoch": 0.8337379033563129,
      "grad_norm": 1.5059969237912278,
      "learning_rate": 1.0232050726252877e-07,
      "loss": 2.3799,
      "step": 6440
    },
    {
      "epoch": 0.8337379033563129,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4965153932571411,
      "eval_runtime": 14.3793,
      "eval_samples_per_second": 3.06,
      "eval_steps_per_second": 0.209,
      "step": 6440
    },
    {
      "epoch": 0.8338673657636664,
      "grad_norm": 1.5173152415254207,
      "learning_rate": 1.0216568104500807e-07,
      "loss": 2.3354,
      "step": 6441
    },
    {
      "epoch": 0.8339968281710198,
      "grad_norm": 2.6131880246119636,
      "learning_rate": 1.0201096162699445e-07,
      "loss": 2.4905,
      "step": 6442
    },
    {
      "epoch": 0.8341262905783733,
      "grad_norm": 3.2120368669527424,
      "learning_rate": 1.0185634904009222e-07,
      "loss": 2.5732,
      "step": 6443
    },
    {
      "epoch": 0.8342557529857267,
      "grad_norm": 1.3179830029889634,
      "learning_rate": 1.0170184331588444e-07,
      "loss": 2.3423,
      "step": 6444
    },
    {
      "epoch": 0.8343852153930802,
      "grad_norm": 2.19358825725728,
      "learning_rate": 1.0154744448593184e-07,
      "loss": 2.4502,
      "step": 6445
    },
    {
      "epoch": 0.8343852153930802,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4992454051971436,
      "eval_runtime": 13.9088,
      "eval_samples_per_second": 3.163,
      "eval_steps_per_second": 0.216,
      "step": 6445
    },
    {
      "epoch": 0.8345146778004336,
      "grad_norm": 1.6287275827762187,
      "learning_rate": 1.0139315258177379e-07,
      "loss": 2.3049,
      "step": 6446
    },
    {
      "epoch": 0.8346441402077872,
      "grad_norm": 1.429665237215633,
      "learning_rate": 1.0123896763492728e-07,
      "loss": 2.323,
      "step": 6447
    },
    {
      "epoch": 0.8347736026151407,
      "grad_norm": 1.4350009006769175,
      "learning_rate": 1.010848896768877e-07,
      "loss": 2.3994,
      "step": 6448
    },
    {
      "epoch": 0.8349030650224941,
      "grad_norm": 1.7655805310349582,
      "learning_rate": 1.0093091873912878e-07,
      "loss": 2.5006,
      "step": 6449
    },
    {
      "epoch": 0.8350325274298476,
      "grad_norm": 2.255536594214907,
      "learning_rate": 1.0077705485310202e-07,
      "loss": 2.3765,
      "step": 6450
    },
    {
      "epoch": 0.8350325274298476,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498579502105713,
      "eval_runtime": 13.663,
      "eval_samples_per_second": 3.22,
      "eval_steps_per_second": 0.22,
      "step": 6450
    },
    {
      "epoch": 0.835161989837201,
      "grad_norm": 1.8054078504276907,
      "learning_rate": 1.0062329805023724e-07,
      "loss": 2.4846,
      "step": 6451
    },
    {
      "epoch": 0.8352914522445545,
      "grad_norm": 3.037454279473438,
      "learning_rate": 1.0046964836194266e-07,
      "loss": 2.5586,
      "step": 6452
    },
    {
      "epoch": 0.8354209146519079,
      "grad_norm": 1.2787395669862127,
      "learning_rate": 1.0031610581960415e-07,
      "loss": 2.2812,
      "step": 6453
    },
    {
      "epoch": 0.8355503770592614,
      "grad_norm": 3.6193811337637163,
      "learning_rate": 1.0016267045458584e-07,
      "loss": 2.3008,
      "step": 6454
    },
    {
      "epoch": 0.8356798394666148,
      "grad_norm": 1.4207908201958475,
      "learning_rate": 1.0000934229822991e-07,
      "loss": 2.4453,
      "step": 6455
    },
    {
      "epoch": 0.8356798394666148,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986017942428589,
      "eval_runtime": 14.7564,
      "eval_samples_per_second": 2.982,
      "eval_steps_per_second": 0.203,
      "step": 6455
    },
    {
      "epoch": 0.8358093018739683,
      "grad_norm": 1.4162014983415658,
      "learning_rate": 9.985612138185698e-08,
      "loss": 2.4373,
      "step": 6456
    },
    {
      "epoch": 0.8359387642813219,
      "grad_norm": 1.3531289973247176,
      "learning_rate": 9.970300773676549e-08,
      "loss": 2.3115,
      "step": 6457
    },
    {
      "epoch": 0.8360682266886753,
      "grad_norm": 1.6191023171546748,
      "learning_rate": 9.955000139423202e-08,
      "loss": 2.1587,
      "step": 6458
    },
    {
      "epoch": 0.8361976890960288,
      "grad_norm": 2.336679131520185,
      "learning_rate": 9.939710238551094e-08,
      "loss": 2.4019,
      "step": 6459
    },
    {
      "epoch": 0.8363271515033822,
      "grad_norm": 1.345522844325268,
      "learning_rate": 9.92443107418354e-08,
      "loss": 2.4321,
      "step": 6460
    },
    {
      "epoch": 0.8363271515033822,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978693723678589,
      "eval_runtime": 15.2064,
      "eval_samples_per_second": 2.894,
      "eval_steps_per_second": 0.197,
      "step": 6460
    },
    {
      "epoch": 0.8364566139107357,
      "grad_norm": 2.266673357236779,
      "learning_rate": 9.909162649441577e-08,
      "loss": 2.3085,
      "step": 6461
    },
    {
      "epoch": 0.8365860763180891,
      "grad_norm": 1.32516756877342,
      "learning_rate": 9.893904967444128e-08,
      "loss": 2.197,
      "step": 6462
    },
    {
      "epoch": 0.8367155387254426,
      "grad_norm": 2.525947931911717,
      "learning_rate": 9.878658031307851e-08,
      "loss": 2.4266,
      "step": 6463
    },
    {
      "epoch": 0.836845001132796,
      "grad_norm": 3.5691086435279207,
      "learning_rate": 9.86342184414727e-08,
      "loss": 2.6842,
      "step": 6464
    },
    {
      "epoch": 0.8369744635401495,
      "grad_norm": 2.135162784958368,
      "learning_rate": 9.848196409074674e-08,
      "loss": 2.4607,
      "step": 6465
    },
    {
      "epoch": 0.8369744635401495,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978249073028564,
      "eval_runtime": 13.5468,
      "eval_samples_per_second": 3.248,
      "eval_steps_per_second": 0.221,
      "step": 6465
    },
    {
      "epoch": 0.837103925947503,
      "grad_norm": 1.8846414014849577,
      "learning_rate": 9.83298172920015e-08,
      "loss": 2.582,
      "step": 6466
    },
    {
      "epoch": 0.8372333883548565,
      "grad_norm": 1.638006688951118,
      "learning_rate": 9.817777807631635e-08,
      "loss": 2.5054,
      "step": 6467
    },
    {
      "epoch": 0.83736285076221,
      "grad_norm": 1.4933553032556124,
      "learning_rate": 9.802584647474814e-08,
      "loss": 2.4329,
      "step": 6468
    },
    {
      "epoch": 0.8374923131695634,
      "grad_norm": 1.5737481583229285,
      "learning_rate": 9.787402251833208e-08,
      "loss": 2.167,
      "step": 6469
    },
    {
      "epoch": 0.8376217755769169,
      "grad_norm": 2.137994404294818,
      "learning_rate": 9.772230623808138e-08,
      "loss": 2.3762,
      "step": 6470
    },
    {
      "epoch": 0.8376217755769169,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4958274364471436,
      "eval_runtime": 13.6285,
      "eval_samples_per_second": 3.229,
      "eval_steps_per_second": 0.22,
      "step": 6470
    },
    {
      "epoch": 0.8377512379842703,
      "grad_norm": 2.2310719545142677,
      "learning_rate": 9.757069766498691e-08,
      "loss": 2.6445,
      "step": 6471
    },
    {
      "epoch": 0.8378807003916238,
      "grad_norm": 2.05354638777211,
      "learning_rate": 9.741919683001803e-08,
      "loss": 2.2329,
      "step": 6472
    },
    {
      "epoch": 0.8380101627989772,
      "grad_norm": 3.5270950270715233,
      "learning_rate": 9.72678037641217e-08,
      "loss": 2.6211,
      "step": 6473
    },
    {
      "epoch": 0.8381396252063307,
      "grad_norm": 1.9027877599330625,
      "learning_rate": 9.711651849822305e-08,
      "loss": 2.3047,
      "step": 6474
    },
    {
      "epoch": 0.8382690876136841,
      "grad_norm": 2.6428275990740673,
      "learning_rate": 9.696534106322548e-08,
      "loss": 2.4868,
      "step": 6475
    },
    {
      "epoch": 0.8382690876136841,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5,
      "eval_runtime": 14.4964,
      "eval_samples_per_second": 3.035,
      "eval_steps_per_second": 0.207,
      "step": 6475
    },
    {
      "epoch": 0.8383985500210377,
      "grad_norm": 2.3408093495437226,
      "learning_rate": 9.681427149000958e-08,
      "loss": 2.4219,
      "step": 6476
    },
    {
      "epoch": 0.8385280124283911,
      "grad_norm": 2.088435308795428,
      "learning_rate": 9.66633098094346e-08,
      "loss": 2.4595,
      "step": 6477
    },
    {
      "epoch": 0.8386574748357446,
      "grad_norm": 1.1944995132995673,
      "learning_rate": 9.651245605233742e-08,
      "loss": 2.103,
      "step": 6478
    },
    {
      "epoch": 0.838786937243098,
      "grad_norm": 1.5030855341472356,
      "learning_rate": 9.636171024953296e-08,
      "loss": 2.1825,
      "step": 6479
    },
    {
      "epoch": 0.8389163996504515,
      "grad_norm": 1.9242026297233268,
      "learning_rate": 9.621107243181438e-08,
      "loss": 2.3271,
      "step": 6480
    },
    {
      "epoch": 0.8389163996504515,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.499267578125,
      "eval_runtime": 13.8401,
      "eval_samples_per_second": 3.179,
      "eval_steps_per_second": 0.217,
      "step": 6480
    },
    {
      "epoch": 0.839045862057805,
      "grad_norm": 2.123819128620499,
      "learning_rate": 9.606054262995233e-08,
      "loss": 2.3103,
      "step": 6481
    },
    {
      "epoch": 0.8391753244651584,
      "grad_norm": 1.8110541653102248,
      "learning_rate": 9.591012087469542e-08,
      "loss": 2.6023,
      "step": 6482
    },
    {
      "epoch": 0.8393047868725119,
      "grad_norm": 1.8685853805122463,
      "learning_rate": 9.575980719677067e-08,
      "loss": 2.3406,
      "step": 6483
    },
    {
      "epoch": 0.8394342492798653,
      "grad_norm": 2.070165888529964,
      "learning_rate": 9.560960162688232e-08,
      "loss": 2.5867,
      "step": 6484
    },
    {
      "epoch": 0.8395637116872188,
      "grad_norm": 1.1295464189418838,
      "learning_rate": 9.54595041957132e-08,
      "loss": 2.3257,
      "step": 6485
    },
    {
      "epoch": 0.8395637116872188,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4963600635528564,
      "eval_runtime": 14.3865,
      "eval_samples_per_second": 3.058,
      "eval_steps_per_second": 0.209,
      "step": 6485
    },
    {
      "epoch": 0.8396931740945723,
      "grad_norm": 2.7487103684316105,
      "learning_rate": 9.530951493392373e-08,
      "loss": 2.4204,
      "step": 6486
    },
    {
      "epoch": 0.8398226365019258,
      "grad_norm": 2.9153054178015214,
      "learning_rate": 9.515963387215194e-08,
      "loss": 2.7673,
      "step": 6487
    },
    {
      "epoch": 0.8399520989092792,
      "grad_norm": 3.653676950453402,
      "learning_rate": 9.500986104101446e-08,
      "loss": 2.4565,
      "step": 6488
    },
    {
      "epoch": 0.8400815613166327,
      "grad_norm": 1.639118775242146,
      "learning_rate": 9.486019647110517e-08,
      "loss": 2.4131,
      "step": 6489
    },
    {
      "epoch": 0.8402110237239862,
      "grad_norm": 1.1148424040470422,
      "learning_rate": 9.471064019299634e-08,
      "loss": 2.3469,
      "step": 6490
    },
    {
      "epoch": 0.8402110237239862,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978693723678589,
      "eval_runtime": 14.1919,
      "eval_samples_per_second": 3.1,
      "eval_steps_per_second": 0.211,
      "step": 6490
    },
    {
      "epoch": 0.8403404861313396,
      "grad_norm": 1.603860514646031,
      "learning_rate": 9.456119223723755e-08,
      "loss": 2.4788,
      "step": 6491
    },
    {
      "epoch": 0.8404699485386931,
      "grad_norm": 1.6274864027338574,
      "learning_rate": 9.441185263435695e-08,
      "loss": 2.1736,
      "step": 6492
    },
    {
      "epoch": 0.8405994109460465,
      "grad_norm": 1.3533335848384915,
      "learning_rate": 9.426262141485995e-08,
      "loss": 2.2622,
      "step": 6493
    },
    {
      "epoch": 0.8407288733534,
      "grad_norm": 3.1789856121972573,
      "learning_rate": 9.411349860922993e-08,
      "loss": 2.5908,
      "step": 6494
    },
    {
      "epoch": 0.8408583357607534,
      "grad_norm": 2.0659538759969194,
      "learning_rate": 9.396448424792856e-08,
      "loss": 2.6143,
      "step": 6495
    },
    {
      "epoch": 0.8408583357607534,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4994007349014282,
      "eval_runtime": 13.5354,
      "eval_samples_per_second": 3.251,
      "eval_steps_per_second": 0.222,
      "step": 6495
    },
    {
      "epoch": 0.840987798168107,
      "grad_norm": 1.6950779115093808,
      "learning_rate": 9.38155783613947e-08,
      "loss": 2.3954,
      "step": 6496
    },
    {
      "epoch": 0.8411172605754604,
      "grad_norm": 3.0319684867384806,
      "learning_rate": 9.366678098004574e-08,
      "loss": 2.4626,
      "step": 6497
    },
    {
      "epoch": 0.8412467229828139,
      "grad_norm": 1.477483873017304,
      "learning_rate": 9.35180921342764e-08,
      "loss": 2.4211,
      "step": 6498
    },
    {
      "epoch": 0.8413761853901673,
      "grad_norm": 2.228960723318929,
      "learning_rate": 9.336951185445926e-08,
      "loss": 2.28,
      "step": 6499
    },
    {
      "epoch": 0.8415056477975208,
      "grad_norm": 1.9539828618597388,
      "learning_rate": 9.322104017094512e-08,
      "loss": 2.3633,
      "step": 6500
    },
    {
      "epoch": 0.8415056477975208,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4991344213485718,
      "eval_runtime": 13.5498,
      "eval_samples_per_second": 3.247,
      "eval_steps_per_second": 0.221,
      "step": 6500
    },
    {
      "epoch": 0.8416351102048742,
      "grad_norm": 1.3259543552425912,
      "learning_rate": 9.307267711406206e-08,
      "loss": 2.343,
      "step": 6501
    },
    {
      "epoch": 0.8417645726122277,
      "grad_norm": 1.4418420274137762,
      "learning_rate": 9.292442271411635e-08,
      "loss": 2.4102,
      "step": 6502
    },
    {
      "epoch": 0.8418940350195812,
      "grad_norm": 2.7925919099498744,
      "learning_rate": 9.277627700139229e-08,
      "loss": 2.4409,
      "step": 6503
    },
    {
      "epoch": 0.8420234974269346,
      "grad_norm": 3.2387106365662044,
      "learning_rate": 9.262824000615102e-08,
      "loss": 2.314,
      "step": 6504
    },
    {
      "epoch": 0.8421529598342882,
      "grad_norm": 1.8635814648159303,
      "learning_rate": 9.248031175863245e-08,
      "loss": 2.3154,
      "step": 6505
    },
    {
      "epoch": 0.8421529598342882,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5000221729278564,
      "eval_runtime": 13.6189,
      "eval_samples_per_second": 3.231,
      "eval_steps_per_second": 0.22,
      "step": 6505
    },
    {
      "epoch": 0.8422824222416416,
      "grad_norm": 2.4410969275031555,
      "learning_rate": 9.233249228905401e-08,
      "loss": 2.3711,
      "step": 6506
    },
    {
      "epoch": 0.8424118846489951,
      "grad_norm": 1.4817949391045466,
      "learning_rate": 9.21847816276105e-08,
      "loss": 2.3257,
      "step": 6507
    },
    {
      "epoch": 0.8425413470563485,
      "grad_norm": 2.0258687338134194,
      "learning_rate": 9.203717980447537e-08,
      "loss": 2.1184,
      "step": 6508
    },
    {
      "epoch": 0.842670809463702,
      "grad_norm": 1.7264698812002046,
      "learning_rate": 9.188968684979864e-08,
      "loss": 2.2495,
      "step": 6509
    },
    {
      "epoch": 0.8428002718710554,
      "grad_norm": 1.4558195482246246,
      "learning_rate": 9.1742302793709e-08,
      "loss": 2.2571,
      "step": 6510
    },
    {
      "epoch": 0.8428002718710554,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972034692764282,
      "eval_runtime": 13.8019,
      "eval_samples_per_second": 3.188,
      "eval_steps_per_second": 0.217,
      "step": 6510
    },
    {
      "epoch": 0.8429297342784089,
      "grad_norm": 1.566367849584917,
      "learning_rate": 9.159502766631287e-08,
      "loss": 2.2446,
      "step": 6511
    },
    {
      "epoch": 0.8430591966857623,
      "grad_norm": 2.5926903019890455,
      "learning_rate": 9.144786149769373e-08,
      "loss": 2.333,
      "step": 6512
    },
    {
      "epoch": 0.8431886590931158,
      "grad_norm": 2.114919991502529,
      "learning_rate": 9.130080431791366e-08,
      "loss": 2.5859,
      "step": 6513
    },
    {
      "epoch": 0.8433181215004693,
      "grad_norm": 2.503769759007235,
      "learning_rate": 9.11538561570119e-08,
      "loss": 2.1353,
      "step": 6514
    },
    {
      "epoch": 0.8434475839078228,
      "grad_norm": 2.021167943015097,
      "learning_rate": 9.100701704500545e-08,
      "loss": 2.3098,
      "step": 6515
    },
    {
      "epoch": 0.8434475839078228,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.49853515625,
      "eval_runtime": 14.0454,
      "eval_samples_per_second": 3.133,
      "eval_steps_per_second": 0.214,
      "step": 6515
    },
    {
      "epoch": 0.8435770463151763,
      "grad_norm": 2.16966049767526,
      "learning_rate": 9.086028701188943e-08,
      "loss": 2.3318,
      "step": 6516
    },
    {
      "epoch": 0.8437065087225297,
      "grad_norm": 2.8273270822500804,
      "learning_rate": 9.07136660876362e-08,
      "loss": 2.6746,
      "step": 6517
    },
    {
      "epoch": 0.8438359711298832,
      "grad_norm": 1.3446635160909532,
      "learning_rate": 9.056715430219631e-08,
      "loss": 2.2957,
      "step": 6518
    },
    {
      "epoch": 0.8439654335372366,
      "grad_norm": 2.0053251293641376,
      "learning_rate": 9.042075168549747e-08,
      "loss": 2.4177,
      "step": 6519
    },
    {
      "epoch": 0.8440948959445901,
      "grad_norm": 1.8918982052194862,
      "learning_rate": 9.027445826744565e-08,
      "loss": 2.3977,
      "step": 6520
    },
    {
      "epoch": 0.8440948959445901,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4992897510528564,
      "eval_runtime": 14.0505,
      "eval_samples_per_second": 3.132,
      "eval_steps_per_second": 0.214,
      "step": 6520
    },
    {
      "epoch": 0.8442243583519435,
      "grad_norm": 3.2638395475813797,
      "learning_rate": 9.012827407792413e-08,
      "loss": 2.5859,
      "step": 6521
    },
    {
      "epoch": 0.844353820759297,
      "grad_norm": 1.1584025231346888,
      "learning_rate": 8.998219914679384e-08,
      "loss": 2.2974,
      "step": 6522
    },
    {
      "epoch": 0.8444832831666504,
      "grad_norm": 1.6002007773791291,
      "learning_rate": 8.983623350389368e-08,
      "loss": 2.2749,
      "step": 6523
    },
    {
      "epoch": 0.8446127455740039,
      "grad_norm": 1.5895397898338828,
      "learning_rate": 8.969037717904027e-08,
      "loss": 2.3691,
      "step": 6524
    },
    {
      "epoch": 0.8447422079813575,
      "grad_norm": 2.0815608784226822,
      "learning_rate": 8.954463020202753e-08,
      "loss": 2.5276,
      "step": 6525
    },
    {
      "epoch": 0.8447422079813575,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4984685182571411,
      "eval_runtime": 14.5154,
      "eval_samples_per_second": 3.031,
      "eval_steps_per_second": 0.207,
      "step": 6525
    },
    {
      "epoch": 0.8448716703887109,
      "grad_norm": 2.3871542043188496,
      "learning_rate": 8.93989926026273e-08,
      "loss": 2.4087,
      "step": 6526
    },
    {
      "epoch": 0.8450011327960644,
      "grad_norm": 2.0524388281270056,
      "learning_rate": 8.925346441058878e-08,
      "loss": 2.396,
      "step": 6527
    },
    {
      "epoch": 0.8451305952034178,
      "grad_norm": 2.5308945057287793,
      "learning_rate": 8.910804565563925e-08,
      "loss": 2.2397,
      "step": 6528
    },
    {
      "epoch": 0.8452600576107713,
      "grad_norm": 1.2536553032983182,
      "learning_rate": 8.89627363674836e-08,
      "loss": 2.1331,
      "step": 6529
    },
    {
      "epoch": 0.8453895200181247,
      "grad_norm": 1.7762497606313208,
      "learning_rate": 8.881753657580386e-08,
      "loss": 2.2786,
      "step": 6530
    },
    {
      "epoch": 0.8453895200181247,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.49853515625,
      "eval_runtime": 14.0646,
      "eval_samples_per_second": 3.128,
      "eval_steps_per_second": 0.213,
      "step": 6530
    },
    {
      "epoch": 0.8455189824254782,
      "grad_norm": 2.964287127326882,
      "learning_rate": 8.867244631026034e-08,
      "loss": 2.3127,
      "step": 6531
    },
    {
      "epoch": 0.8456484448328316,
      "grad_norm": 3.319152351668377,
      "learning_rate": 8.852746560049051e-08,
      "loss": 2.4539,
      "step": 6532
    },
    {
      "epoch": 0.8457779072401851,
      "grad_norm": 1.6225686188504977,
      "learning_rate": 8.83825944761095e-08,
      "loss": 2.3101,
      "step": 6533
    },
    {
      "epoch": 0.8459073696475385,
      "grad_norm": 1.7803476185344367,
      "learning_rate": 8.823783296671044e-08,
      "loss": 2.3115,
      "step": 6534
    },
    {
      "epoch": 0.8460368320548921,
      "grad_norm": 2.045672519888185,
      "learning_rate": 8.809318110186357e-08,
      "loss": 2.2305,
      "step": 6535
    },
    {
      "epoch": 0.8460368320548921,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.501265048980713,
      "eval_runtime": 13.5854,
      "eval_samples_per_second": 3.239,
      "eval_steps_per_second": 0.221,
      "step": 6535
    },
    {
      "epoch": 0.8461662944622456,
      "grad_norm": 1.975831045634424,
      "learning_rate": 8.79486389111172e-08,
      "loss": 2.4536,
      "step": 6536
    },
    {
      "epoch": 0.846295756869599,
      "grad_norm": 1.1154771045361036,
      "learning_rate": 8.78042064239968e-08,
      "loss": 2.28,
      "step": 6537
    },
    {
      "epoch": 0.8464252192769525,
      "grad_norm": 2.992747146713637,
      "learning_rate": 8.765988367000565e-08,
      "loss": 2.3772,
      "step": 6538
    },
    {
      "epoch": 0.8465546816843059,
      "grad_norm": 1.3438032785430103,
      "learning_rate": 8.751567067862475e-08,
      "loss": 2.3877,
      "step": 6539
    },
    {
      "epoch": 0.8466841440916594,
      "grad_norm": 1.292350606048276,
      "learning_rate": 8.737156747931227e-08,
      "loss": 2.3428,
      "step": 6540
    },
    {
      "epoch": 0.8466841440916594,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5000221729278564,
      "eval_runtime": 13.6721,
      "eval_samples_per_second": 3.218,
      "eval_steps_per_second": 0.219,
      "step": 6540
    },
    {
      "epoch": 0.8468136064990128,
      "grad_norm": 1.3521108405047466,
      "learning_rate": 8.722757410150453e-08,
      "loss": 2.2734,
      "step": 6541
    },
    {
      "epoch": 0.8469430689063663,
      "grad_norm": 2.4058565450452765,
      "learning_rate": 8.708369057461488e-08,
      "loss": 2.4351,
      "step": 6542
    },
    {
      "epoch": 0.8470725313137197,
      "grad_norm": 1.7071696905699472,
      "learning_rate": 8.693991692803439e-08,
      "loss": 2.4612,
      "step": 6543
    },
    {
      "epoch": 0.8472019937210733,
      "grad_norm": 1.8158029047366344,
      "learning_rate": 8.679625319113205e-08,
      "loss": 2.4351,
      "step": 6544
    },
    {
      "epoch": 0.8473314561284268,
      "grad_norm": 2.308366482130721,
      "learning_rate": 8.665269939325365e-08,
      "loss": 2.2588,
      "step": 6545
    },
    {
      "epoch": 0.8473314561284268,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.497025966644287,
      "eval_runtime": 13.4986,
      "eval_samples_per_second": 3.26,
      "eval_steps_per_second": 0.222,
      "step": 6545
    },
    {
      "epoch": 0.8474609185357802,
      "grad_norm": 4.451751948532977,
      "learning_rate": 8.650925556372328e-08,
      "loss": 2.3818,
      "step": 6546
    },
    {
      "epoch": 0.8475903809431337,
      "grad_norm": 2.2877676550449353,
      "learning_rate": 8.636592173184237e-08,
      "loss": 2.4968,
      "step": 6547
    },
    {
      "epoch": 0.8477198433504871,
      "grad_norm": 1.5277204605513772,
      "learning_rate": 8.622269792688938e-08,
      "loss": 2.4919,
      "step": 6548
    },
    {
      "epoch": 0.8478493057578406,
      "grad_norm": 1.5636627127805616,
      "learning_rate": 8.60795841781209e-08,
      "loss": 2.2617,
      "step": 6549
    },
    {
      "epoch": 0.847978768165194,
      "grad_norm": 4.2296920407784375,
      "learning_rate": 8.59365805147707e-08,
      "loss": 2.3198,
      "step": 6550
    },
    {
      "epoch": 0.847978768165194,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978915452957153,
      "eval_runtime": 14.1425,
      "eval_samples_per_second": 3.111,
      "eval_steps_per_second": 0.212,
      "step": 6550
    },
    {
      "epoch": 0.8481082305725475,
      "grad_norm": 1.332680801780174,
      "learning_rate": 8.579368696605021e-08,
      "loss": 2.3081,
      "step": 6551
    },
    {
      "epoch": 0.8482376929799009,
      "grad_norm": 2.779204567943503,
      "learning_rate": 8.565090356114851e-08,
      "loss": 2.2917,
      "step": 6552
    },
    {
      "epoch": 0.8483671553872544,
      "grad_norm": 4.235294498534441,
      "learning_rate": 8.550823032923191e-08,
      "loss": 2.3481,
      "step": 6553
    },
    {
      "epoch": 0.8484966177946079,
      "grad_norm": 2.5930879672563005,
      "learning_rate": 8.536566729944406e-08,
      "loss": 2.3984,
      "step": 6554
    },
    {
      "epoch": 0.8486260802019614,
      "grad_norm": 1.6926972085102159,
      "learning_rate": 8.522321450090673e-08,
      "loss": 2.417,
      "step": 6555
    },
    {
      "epoch": 0.8486260802019614,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498579502105713,
      "eval_runtime": 15.1432,
      "eval_samples_per_second": 2.906,
      "eval_steps_per_second": 0.198,
      "step": 6555
    },
    {
      "epoch": 0.8487555426093149,
      "grad_norm": 1.4755699823751198,
      "learning_rate": 8.508087196271851e-08,
      "loss": 2.4583,
      "step": 6556
    },
    {
      "epoch": 0.8488850050166683,
      "grad_norm": 1.6616283466143034,
      "learning_rate": 8.493863971395594e-08,
      "loss": 2.3877,
      "step": 6557
    },
    {
      "epoch": 0.8490144674240218,
      "grad_norm": 1.9845445460265752,
      "learning_rate": 8.479651778367266e-08,
      "loss": 2.5459,
      "step": 6558
    },
    {
      "epoch": 0.8491439298313752,
      "grad_norm": 1.4785273182922332,
      "learning_rate": 8.465450620090014e-08,
      "loss": 2.2258,
      "step": 6559
    },
    {
      "epoch": 0.8492733922387287,
      "grad_norm": 2.0323793655426,
      "learning_rate": 8.451260499464701e-08,
      "loss": 2.4294,
      "step": 6560
    },
    {
      "epoch": 0.8492733922387287,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5007545948028564,
      "eval_runtime": 13.9743,
      "eval_samples_per_second": 3.149,
      "eval_steps_per_second": 0.215,
      "step": 6560
    },
    {
      "epoch": 0.8494028546460821,
      "grad_norm": 1.684013440587913,
      "learning_rate": 8.437081419389939e-08,
      "loss": 2.3379,
      "step": 6561
    },
    {
      "epoch": 0.8495323170534356,
      "grad_norm": 1.6807115087927769,
      "learning_rate": 8.422913382762119e-08,
      "loss": 2.3931,
      "step": 6562
    },
    {
      "epoch": 0.849661779460789,
      "grad_norm": 1.4753970675663166,
      "learning_rate": 8.40875639247532e-08,
      "loss": 2.1882,
      "step": 6563
    },
    {
      "epoch": 0.8497912418681426,
      "grad_norm": 1.8688579009748316,
      "learning_rate": 8.394610451421419e-08,
      "loss": 2.436,
      "step": 6564
    },
    {
      "epoch": 0.849920704275496,
      "grad_norm": 2.2596132532911466,
      "learning_rate": 8.380475562489999e-08,
      "loss": 2.2864,
      "step": 6565
    },
    {
      "epoch": 0.849920704275496,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.496382236480713,
      "eval_runtime": 15.4174,
      "eval_samples_per_second": 2.854,
      "eval_steps_per_second": 0.195,
      "step": 6565
    },
    {
      "epoch": 0.8500501666828495,
      "grad_norm": 2.237402806840521,
      "learning_rate": 8.366351728568381e-08,
      "loss": 2.79,
      "step": 6566
    },
    {
      "epoch": 0.850179629090203,
      "grad_norm": 1.8105416369856966,
      "learning_rate": 8.352238952541677e-08,
      "loss": 2.344,
      "step": 6567
    },
    {
      "epoch": 0.8503090914975564,
      "grad_norm": 1.9674130505675813,
      "learning_rate": 8.338137237292682e-08,
      "loss": 2.3169,
      "step": 6568
    },
    {
      "epoch": 0.8504385539049099,
      "grad_norm": 1.4080536621367579,
      "learning_rate": 8.324046585701961e-08,
      "loss": 2.4047,
      "step": 6569
    },
    {
      "epoch": 0.8505680163122633,
      "grad_norm": 1.2238358240747322,
      "learning_rate": 8.309967000647859e-08,
      "loss": 2.4175,
      "step": 6570
    },
    {
      "epoch": 0.8505680163122633,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.497758388519287,
      "eval_runtime": 13.8179,
      "eval_samples_per_second": 3.184,
      "eval_steps_per_second": 0.217,
      "step": 6570
    },
    {
      "epoch": 0.8506974787196168,
      "grad_norm": 1.4394485178164549,
      "learning_rate": 8.295898485006344e-08,
      "loss": 2.2981,
      "step": 6571
    },
    {
      "epoch": 0.8508269411269702,
      "grad_norm": 2.8709935283050654,
      "learning_rate": 8.281841041651244e-08,
      "loss": 2.6245,
      "step": 6572
    },
    {
      "epoch": 0.8509564035343237,
      "grad_norm": 2.0917170872260242,
      "learning_rate": 8.267794673454052e-08,
      "loss": 2.4221,
      "step": 6573
    },
    {
      "epoch": 0.8510858659416772,
      "grad_norm": 1.6158921248350986,
      "learning_rate": 8.253759383284037e-08,
      "loss": 2.1921,
      "step": 6574
    },
    {
      "epoch": 0.8512153283490307,
      "grad_norm": 1.9901127000146681,
      "learning_rate": 8.239735174008209e-08,
      "loss": 2.2681,
      "step": 6575
    },
    {
      "epoch": 0.8512153283490307,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957829713821411,
      "eval_runtime": 14.1554,
      "eval_samples_per_second": 3.108,
      "eval_steps_per_second": 0.212,
      "step": 6575
    },
    {
      "epoch": 0.8513447907563841,
      "grad_norm": 2.18364016590462,
      "learning_rate": 8.225722048491258e-08,
      "loss": 2.3252,
      "step": 6576
    },
    {
      "epoch": 0.8514742531637376,
      "grad_norm": 1.4391037089945609,
      "learning_rate": 8.211720009595663e-08,
      "loss": 2.3042,
      "step": 6577
    },
    {
      "epoch": 0.851603715571091,
      "grad_norm": 1.9606021614539075,
      "learning_rate": 8.197729060181646e-08,
      "loss": 2.4482,
      "step": 6578
    },
    {
      "epoch": 0.8517331779784445,
      "grad_norm": 1.885092716148146,
      "learning_rate": 8.183749203107105e-08,
      "loss": 2.2485,
      "step": 6579
    },
    {
      "epoch": 0.851862640385798,
      "grad_norm": 1.1966103608416532,
      "learning_rate": 8.169780441227751e-08,
      "loss": 2.428,
      "step": 6580
    },
    {
      "epoch": 0.851862640385798,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964710474014282,
      "eval_runtime": 14.3668,
      "eval_samples_per_second": 3.063,
      "eval_steps_per_second": 0.209,
      "step": 6580
    },
    {
      "epoch": 0.8519921027931514,
      "grad_norm": 1.2957432268188154,
      "learning_rate": 8.15582277739696e-08,
      "loss": 2.1587,
      "step": 6581
    },
    {
      "epoch": 0.8521215652005049,
      "grad_norm": 2.176315218841233,
      "learning_rate": 8.141876214465862e-08,
      "loss": 2.4551,
      "step": 6582
    },
    {
      "epoch": 0.8522510276078583,
      "grad_norm": 2.019386925982672,
      "learning_rate": 8.12794075528335e-08,
      "loss": 2.4011,
      "step": 6583
    },
    {
      "epoch": 0.8523804900152119,
      "grad_norm": 3.1691583603241917,
      "learning_rate": 8.114016402695997e-08,
      "loss": 2.6421,
      "step": 6584
    },
    {
      "epoch": 0.8525099524225653,
      "grad_norm": 2.134935088096232,
      "learning_rate": 8.100103159548165e-08,
      "loss": 2.3665,
      "step": 6585
    },
    {
      "epoch": 0.8525099524225653,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957829713821411,
      "eval_runtime": 20.7121,
      "eval_samples_per_second": 2.124,
      "eval_steps_per_second": 0.145,
      "step": 6585
    },
    {
      "epoch": 0.8526394148299188,
      "grad_norm": 1.4413197206552746,
      "learning_rate": 8.086201028681892e-08,
      "loss": 2.4827,
      "step": 6586
    },
    {
      "epoch": 0.8527688772372722,
      "grad_norm": 1.19904193420047,
      "learning_rate": 8.072310012936964e-08,
      "loss": 2.291,
      "step": 6587
    },
    {
      "epoch": 0.8528983396446257,
      "grad_norm": 1.8884830551462888,
      "learning_rate": 8.05843011515093e-08,
      "loss": 2.396,
      "step": 6588
    },
    {
      "epoch": 0.8530278020519791,
      "grad_norm": 2.1765376138593675,
      "learning_rate": 8.044561338159007e-08,
      "loss": 2.2468,
      "step": 6589
    },
    {
      "epoch": 0.8531572644593326,
      "grad_norm": 2.645903634177425,
      "learning_rate": 8.030703684794208e-08,
      "loss": 2.3191,
      "step": 6590
    },
    {
      "epoch": 0.8531572644593326,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.497270107269287,
      "eval_runtime": 20.6437,
      "eval_samples_per_second": 2.131,
      "eval_steps_per_second": 0.145,
      "step": 6590
    },
    {
      "epoch": 0.853286726866686,
      "grad_norm": 1.7761040813914157,
      "learning_rate": 8.016857157887214e-08,
      "loss": 2.3232,
      "step": 6591
    },
    {
      "epoch": 0.8534161892740395,
      "grad_norm": 2.907370819672574,
      "learning_rate": 8.003021760266482e-08,
      "loss": 2.4331,
      "step": 6592
    },
    {
      "epoch": 0.8535456516813931,
      "grad_norm": 1.9621918112269952,
      "learning_rate": 7.989197494758155e-08,
      "loss": 2.4325,
      "step": 6593
    },
    {
      "epoch": 0.8536751140887465,
      "grad_norm": 1.1107233976492266,
      "learning_rate": 7.975384364186108e-08,
      "loss": 2.2961,
      "step": 6594
    },
    {
      "epoch": 0.8538045764961,
      "grad_norm": 1.8042133984389783,
      "learning_rate": 7.961582371371984e-08,
      "loss": 2.2388,
      "step": 6595
    },
    {
      "epoch": 0.8538045764961,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979137182235718,
      "eval_runtime": 14.1021,
      "eval_samples_per_second": 3.12,
      "eval_steps_per_second": 0.213,
      "step": 6595
    },
    {
      "epoch": 0.8539340389034534,
      "grad_norm": 1.4439991292614032,
      "learning_rate": 7.947791519135092e-08,
      "loss": 2.3792,
      "step": 6596
    },
    {
      "epoch": 0.8540635013108069,
      "grad_norm": 1.2676295708067427,
      "learning_rate": 7.934011810292505e-08,
      "loss": 2.3293,
      "step": 6597
    },
    {
      "epoch": 0.8541929637181603,
      "grad_norm": 1.4310416848210628,
      "learning_rate": 7.92024324765904e-08,
      "loss": 2.4097,
      "step": 6598
    },
    {
      "epoch": 0.8543224261255138,
      "grad_norm": 1.4984133539034448,
      "learning_rate": 7.906485834047135e-08,
      "loss": 2.1934,
      "step": 6599
    },
    {
      "epoch": 0.8544518885328672,
      "grad_norm": 2.6466724781204167,
      "learning_rate": 7.892739572267064e-08,
      "loss": 2.5369,
      "step": 6600
    },
    {
      "epoch": 0.8544518885328672,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.49853515625,
      "eval_runtime": 13.8267,
      "eval_samples_per_second": 3.182,
      "eval_steps_per_second": 0.217,
      "step": 6600
    },
    {
      "epoch": 0.8545813509402207,
      "grad_norm": 1.91746015441227,
      "learning_rate": 7.879004465126786e-08,
      "loss": 2.3949,
      "step": 6601
    },
    {
      "epoch": 0.8547108133475741,
      "grad_norm": 2.3182520639332855,
      "learning_rate": 7.865280515431957e-08,
      "loss": 2.27,
      "step": 6602
    },
    {
      "epoch": 0.8548402757549277,
      "grad_norm": 1.5024747837125307,
      "learning_rate": 7.851567725985997e-08,
      "loss": 2.4561,
      "step": 6603
    },
    {
      "epoch": 0.8549697381622812,
      "grad_norm": 1.4492388772804083,
      "learning_rate": 7.837866099589974e-08,
      "loss": 2.4192,
      "step": 6604
    },
    {
      "epoch": 0.8550992005696346,
      "grad_norm": 3.33924252541541,
      "learning_rate": 7.824175639042759e-08,
      "loss": 2.3782,
      "step": 6605
    },
    {
      "epoch": 0.8550992005696346,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.499467372894287,
      "eval_runtime": 13.8932,
      "eval_samples_per_second": 3.167,
      "eval_steps_per_second": 0.216,
      "step": 6605
    },
    {
      "epoch": 0.8552286629769881,
      "grad_norm": 3.3559383969194463,
      "learning_rate": 7.810496347140907e-08,
      "loss": 2.5371,
      "step": 6606
    },
    {
      "epoch": 0.8553581253843415,
      "grad_norm": 1.3101877312417733,
      "learning_rate": 7.796828226678668e-08,
      "loss": 2.2766,
      "step": 6607
    },
    {
      "epoch": 0.855487587791695,
      "grad_norm": 1.3990719601509942,
      "learning_rate": 7.78317128044806e-08,
      "loss": 2.4614,
      "step": 6608
    },
    {
      "epoch": 0.8556170501990484,
      "grad_norm": 1.6327505814759302,
      "learning_rate": 7.76952551123878e-08,
      "loss": 2.3713,
      "step": 6609
    },
    {
      "epoch": 0.8557465126064019,
      "grad_norm": 1.8005477359662325,
      "learning_rate": 7.755890921838233e-08,
      "loss": 2.479,
      "step": 6610
    },
    {
      "epoch": 0.8557465126064019,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964045286178589,
      "eval_runtime": 15.0772,
      "eval_samples_per_second": 2.918,
      "eval_steps_per_second": 0.199,
      "step": 6610
    },
    {
      "epoch": 0.8558759750137553,
      "grad_norm": 1.8673360100529317,
      "learning_rate": 7.742267515031598e-08,
      "loss": 2.2236,
      "step": 6611
    },
    {
      "epoch": 0.8560054374211088,
      "grad_norm": 1.488865352619389,
      "learning_rate": 7.728655293601693e-08,
      "loss": 2.323,
      "step": 6612
    },
    {
      "epoch": 0.8561348998284624,
      "grad_norm": 2.4792871443144997,
      "learning_rate": 7.715054260329119e-08,
      "loss": 2.4717,
      "step": 6613
    },
    {
      "epoch": 0.8562643622358158,
      "grad_norm": 2.7874324571707567,
      "learning_rate": 7.70146441799216e-08,
      "loss": 2.3867,
      "step": 6614
    },
    {
      "epoch": 0.8563938246431693,
      "grad_norm": 2.0496795992850987,
      "learning_rate": 7.687885769366791e-08,
      "loss": 2.4673,
      "step": 6615
    },
    {
      "epoch": 0.8563938246431693,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986239671707153,
      "eval_runtime": 13.8288,
      "eval_samples_per_second": 3.182,
      "eval_steps_per_second": 0.217,
      "step": 6615
    },
    {
      "epoch": 0.8565232870505227,
      "grad_norm": 2.2504258093393696,
      "learning_rate": 7.674318317226759e-08,
      "loss": 2.2751,
      "step": 6616
    },
    {
      "epoch": 0.8566527494578762,
      "grad_norm": 1.783309652459419,
      "learning_rate": 7.660762064343463e-08,
      "loss": 2.2981,
      "step": 6617
    },
    {
      "epoch": 0.8567822118652296,
      "grad_norm": 1.2477501546588274,
      "learning_rate": 7.647217013486058e-08,
      "loss": 2.2598,
      "step": 6618
    },
    {
      "epoch": 0.8569116742725831,
      "grad_norm": 1.599085079126057,
      "learning_rate": 7.633683167421406e-08,
      "loss": 2.2395,
      "step": 6619
    },
    {
      "epoch": 0.8570411366799365,
      "grad_norm": 1.5007528903720948,
      "learning_rate": 7.620160528914053e-08,
      "loss": 2.4846,
      "step": 6620
    },
    {
      "epoch": 0.8570411366799365,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957386255264282,
      "eval_runtime": 14.4218,
      "eval_samples_per_second": 3.051,
      "eval_steps_per_second": 0.208,
      "step": 6620
    },
    {
      "epoch": 0.85717059908729,
      "grad_norm": 2.247987149619524,
      "learning_rate": 7.606649100726283e-08,
      "loss": 2.4062,
      "step": 6621
    },
    {
      "epoch": 0.8573000614946434,
      "grad_norm": 2.011566519304433,
      "learning_rate": 7.593148885618052e-08,
      "loss": 2.6294,
      "step": 6622
    },
    {
      "epoch": 0.857429523901997,
      "grad_norm": 2.3806236406277876,
      "learning_rate": 7.579659886347076e-08,
      "loss": 2.4233,
      "step": 6623
    },
    {
      "epoch": 0.8575589863093505,
      "grad_norm": 1.0506828312725818,
      "learning_rate": 7.566182105668763e-08,
      "loss": 2.071,
      "step": 6624
    },
    {
      "epoch": 0.8576884487167039,
      "grad_norm": 1.8997382080487009,
      "learning_rate": 7.552715546336214e-08,
      "loss": 2.5034,
      "step": 6625
    },
    {
      "epoch": 0.8576884487167039,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979802370071411,
      "eval_runtime": 14.5319,
      "eval_samples_per_second": 3.028,
      "eval_steps_per_second": 0.206,
      "step": 6625
    },
    {
      "epoch": 0.8578179111240574,
      "grad_norm": 1.1481669993995705,
      "learning_rate": 7.539260211100234e-08,
      "loss": 2.1968,
      "step": 6626
    },
    {
      "epoch": 0.8579473735314108,
      "grad_norm": 1.490292122150186,
      "learning_rate": 7.525816102709367e-08,
      "loss": 2.3159,
      "step": 6627
    },
    {
      "epoch": 0.8580768359387643,
      "grad_norm": 1.4973001998170343,
      "learning_rate": 7.512383223909833e-08,
      "loss": 2.4431,
      "step": 6628
    },
    {
      "epoch": 0.8582062983461177,
      "grad_norm": 1.6000629422857742,
      "learning_rate": 7.498961577445577e-08,
      "loss": 2.2461,
      "step": 6629
    },
    {
      "epoch": 0.8583357607534712,
      "grad_norm": 2.0629435424889464,
      "learning_rate": 7.48555116605823e-08,
      "loss": 2.7925,
      "step": 6630
    },
    {
      "epoch": 0.8583357607534712,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4951171875,
      "eval_runtime": 14.859,
      "eval_samples_per_second": 2.961,
      "eval_steps_per_second": 0.202,
      "step": 6630
    },
    {
      "epoch": 0.8584652231608246,
      "grad_norm": 1.8964111486105253,
      "learning_rate": 7.472151992487166e-08,
      "loss": 2.448,
      "step": 6631
    },
    {
      "epoch": 0.8585946855681782,
      "grad_norm": 2.2718942043757604,
      "learning_rate": 7.458764059469415e-08,
      "loss": 2.2598,
      "step": 6632
    },
    {
      "epoch": 0.8587241479755316,
      "grad_norm": 1.9176820788679079,
      "learning_rate": 7.44538736973973e-08,
      "loss": 2.2473,
      "step": 6633
    },
    {
      "epoch": 0.8588536103828851,
      "grad_norm": 1.5155302775959094,
      "learning_rate": 7.432021926030598e-08,
      "loss": 2.3923,
      "step": 6634
    },
    {
      "epoch": 0.8589830727902386,
      "grad_norm": 4.963229203085601,
      "learning_rate": 7.418667731072144e-08,
      "loss": 2.6514,
      "step": 6635
    },
    {
      "epoch": 0.8589830727902386,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4951171875,
      "eval_runtime": 13.5948,
      "eval_samples_per_second": 3.237,
      "eval_steps_per_second": 0.221,
      "step": 6635
    },
    {
      "epoch": 0.859112535197592,
      "grad_norm": 1.919372632336787,
      "learning_rate": 7.405324787592274e-08,
      "loss": 2.3503,
      "step": 6636
    },
    {
      "epoch": 0.8592419976049455,
      "grad_norm": 1.3871793180860774,
      "learning_rate": 7.391993098316529e-08,
      "loss": 2.1494,
      "step": 6637
    },
    {
      "epoch": 0.8593714600122989,
      "grad_norm": 1.6644484785152112,
      "learning_rate": 7.378672665968172e-08,
      "loss": 2.2834,
      "step": 6638
    },
    {
      "epoch": 0.8595009224196524,
      "grad_norm": 1.8542750543387743,
      "learning_rate": 7.36536349326819e-08,
      "loss": 2.4065,
      "step": 6639
    },
    {
      "epoch": 0.8596303848270058,
      "grad_norm": 1.4092349047145702,
      "learning_rate": 7.352065582935233e-08,
      "loss": 2.3184,
      "step": 6640
    },
    {
      "epoch": 0.8596303848270058,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972034692764282,
      "eval_runtime": 13.8724,
      "eval_samples_per_second": 3.172,
      "eval_steps_per_second": 0.216,
      "step": 6640
    },
    {
      "epoch": 0.8597598472343593,
      "grad_norm": 1.5609488746223876,
      "learning_rate": 7.338778937685677e-08,
      "loss": 2.4194,
      "step": 6641
    },
    {
      "epoch": 0.8598893096417128,
      "grad_norm": 1.409179470791384,
      "learning_rate": 7.325503560233622e-08,
      "loss": 2.1055,
      "step": 6642
    },
    {
      "epoch": 0.8600187720490663,
      "grad_norm": 1.2215535484344948,
      "learning_rate": 7.312239453290772e-08,
      "loss": 2.2432,
      "step": 6643
    },
    {
      "epoch": 0.8601482344564197,
      "grad_norm": 1.4873162733115775,
      "learning_rate": 7.298986619566628e-08,
      "loss": 2.3604,
      "step": 6644
    },
    {
      "epoch": 0.8602776968637732,
      "grad_norm": 2.122053465202789,
      "learning_rate": 7.28574506176833e-08,
      "loss": 2.3068,
      "step": 6645
    },
    {
      "epoch": 0.8602776968637732,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979137182235718,
      "eval_runtime": 14.5146,
      "eval_samples_per_second": 3.031,
      "eval_steps_per_second": 0.207,
      "step": 6645
    },
    {
      "epoch": 0.8604071592711267,
      "grad_norm": 3.1898086590477694,
      "learning_rate": 7.272514782600742e-08,
      "loss": 2.4614,
      "step": 6646
    },
    {
      "epoch": 0.8605366216784801,
      "grad_norm": 1.3091026421283558,
      "learning_rate": 7.259295784766431e-08,
      "loss": 2.3464,
      "step": 6647
    },
    {
      "epoch": 0.8606660840858336,
      "grad_norm": 1.3026709897836444,
      "learning_rate": 7.246088070965641e-08,
      "loss": 2.3071,
      "step": 6648
    },
    {
      "epoch": 0.860795546493187,
      "grad_norm": 1.72972720740264,
      "learning_rate": 7.232891643896296e-08,
      "loss": 2.3364,
      "step": 6649
    },
    {
      "epoch": 0.8609250089005405,
      "grad_norm": 3.2077939199248324,
      "learning_rate": 7.219706506254052e-08,
      "loss": 2.2295,
      "step": 6650
    },
    {
      "epoch": 0.8609250089005405,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4962712526321411,
      "eval_runtime": 13.3678,
      "eval_samples_per_second": 3.291,
      "eval_steps_per_second": 0.224,
      "step": 6650
    },
    {
      "epoch": 0.8610544713078939,
      "grad_norm": 1.3060665403929674,
      "learning_rate": 7.206532660732227e-08,
      "loss": 2.3586,
      "step": 6651
    },
    {
      "epoch": 0.8611839337152475,
      "grad_norm": 1.7815142420066077,
      "learning_rate": 7.193370110021865e-08,
      "loss": 2.2949,
      "step": 6652
    },
    {
      "epoch": 0.8613133961226009,
      "grad_norm": 1.5681106973761263,
      "learning_rate": 7.180218856811673e-08,
      "loss": 2.2166,
      "step": 6653
    },
    {
      "epoch": 0.8614428585299544,
      "grad_norm": 1.3938777034923548,
      "learning_rate": 7.167078903788045e-08,
      "loss": 2.3921,
      "step": 6654
    },
    {
      "epoch": 0.8615723209373078,
      "grad_norm": 4.808381169883879,
      "learning_rate": 7.153950253635113e-08,
      "loss": 2.6716,
      "step": 6655
    },
    {
      "epoch": 0.8615723209373078,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498579502105713,
      "eval_runtime": 15.08,
      "eval_samples_per_second": 2.918,
      "eval_steps_per_second": 0.199,
      "step": 6655
    },
    {
      "epoch": 0.8617017833446613,
      "grad_norm": 1.5097288496512302,
      "learning_rate": 7.140832909034639e-08,
      "loss": 2.3577,
      "step": 6656
    },
    {
      "epoch": 0.8618312457520148,
      "grad_norm": 2.8709418748852906,
      "learning_rate": 7.127726872666135e-08,
      "loss": 2.3718,
      "step": 6657
    },
    {
      "epoch": 0.8619607081593682,
      "grad_norm": 1.9944739230712052,
      "learning_rate": 7.114632147206758e-08,
      "loss": 2.4968,
      "step": 6658
    },
    {
      "epoch": 0.8620901705667217,
      "grad_norm": 2.219423702053198,
      "learning_rate": 7.101548735331378e-08,
      "loss": 2.2529,
      "step": 6659
    },
    {
      "epoch": 0.8622196329740751,
      "grad_norm": 1.7419824075160113,
      "learning_rate": 7.088476639712551e-08,
      "loss": 2.4932,
      "step": 6660
    },
    {
      "epoch": 0.8622196329740751,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.495649814605713,
      "eval_runtime": 13.8326,
      "eval_samples_per_second": 3.181,
      "eval_steps_per_second": 0.217,
      "step": 6660
    },
    {
      "epoch": 0.8623490953814286,
      "grad_norm": 3.0449880880269626,
      "learning_rate": 7.075415863020497e-08,
      "loss": 2.375,
      "step": 6661
    },
    {
      "epoch": 0.8624785577887821,
      "grad_norm": 1.7712876580479961,
      "learning_rate": 7.062366407923168e-08,
      "loss": 2.5535,
      "step": 6662
    },
    {
      "epoch": 0.8626080201961356,
      "grad_norm": 1.4868857233321155,
      "learning_rate": 7.049328277086154e-08,
      "loss": 2.4163,
      "step": 6663
    },
    {
      "epoch": 0.862737482603489,
      "grad_norm": 2.828166011141021,
      "learning_rate": 7.036301473172794e-08,
      "loss": 2.4304,
      "step": 6664
    },
    {
      "epoch": 0.8628669450108425,
      "grad_norm": 1.9820238185600705,
      "learning_rate": 7.023285998844054e-08,
      "loss": 2.4917,
      "step": 6665
    },
    {
      "epoch": 0.8628669450108425,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.499267578125,
      "eval_runtime": 13.9697,
      "eval_samples_per_second": 3.15,
      "eval_steps_per_second": 0.215,
      "step": 6665
    },
    {
      "epoch": 0.8629964074181959,
      "grad_norm": 1.643282848319218,
      "learning_rate": 7.010281856758596e-08,
      "loss": 2.2725,
      "step": 6666
    },
    {
      "epoch": 0.8631258698255494,
      "grad_norm": 3.094323303257011,
      "learning_rate": 6.997289049572802e-08,
      "loss": 2.2539,
      "step": 6667
    },
    {
      "epoch": 0.8632553322329029,
      "grad_norm": 1.5211016580596106,
      "learning_rate": 6.984307579940702e-08,
      "loss": 2.3838,
      "step": 6668
    },
    {
      "epoch": 0.8633847946402563,
      "grad_norm": 1.6884303073669678,
      "learning_rate": 6.97133745051403e-08,
      "loss": 2.2773,
      "step": 6669
    },
    {
      "epoch": 0.8635142570476098,
      "grad_norm": 3.0494919365178546,
      "learning_rate": 6.958378663942222e-08,
      "loss": 2.2134,
      "step": 6670
    },
    {
      "epoch": 0.8635142570476098,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4965153932571411,
      "eval_runtime": 14.3478,
      "eval_samples_per_second": 3.067,
      "eval_steps_per_second": 0.209,
      "step": 6670
    },
    {
      "epoch": 0.8636437194549632,
      "grad_norm": 3.2495713404244317,
      "learning_rate": 6.945431222872328e-08,
      "loss": 2.406,
      "step": 6671
    },
    {
      "epoch": 0.8637731818623168,
      "grad_norm": 1.3206205069732337,
      "learning_rate": 6.93249512994914e-08,
      "loss": 2.4033,
      "step": 6672
    },
    {
      "epoch": 0.8639026442696702,
      "grad_norm": 1.6309333651349487,
      "learning_rate": 6.919570387815142e-08,
      "loss": 2.4502,
      "step": 6673
    },
    {
      "epoch": 0.8640321066770237,
      "grad_norm": 1.6215884944324817,
      "learning_rate": 6.906656999110445e-08,
      "loss": 2.3093,
      "step": 6674
    },
    {
      "epoch": 0.8641615690843771,
      "grad_norm": 2.7176432304682074,
      "learning_rate": 6.893754966472897e-08,
      "loss": 2.2837,
      "step": 6675
    },
    {
      "epoch": 0.8641615690843771,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498490810394287,
      "eval_runtime": 15.4896,
      "eval_samples_per_second": 2.841,
      "eval_steps_per_second": 0.194,
      "step": 6675
    },
    {
      "epoch": 0.8642910314917306,
      "grad_norm": 3.6147734963821443,
      "learning_rate": 6.880864292537967e-08,
      "loss": 2.3394,
      "step": 6676
    },
    {
      "epoch": 0.864420493899084,
      "grad_norm": 1.9261856659341565,
      "learning_rate": 6.867984979938853e-08,
      "loss": 2.4463,
      "step": 6677
    },
    {
      "epoch": 0.8645499563064375,
      "grad_norm": 1.2587535992701346,
      "learning_rate": 6.855117031306431e-08,
      "loss": 2.4731,
      "step": 6678
    },
    {
      "epoch": 0.864679418713791,
      "grad_norm": 1.49590193668734,
      "learning_rate": 6.842260449269209e-08,
      "loss": 2.4614,
      "step": 6679
    },
    {
      "epoch": 0.8648088811211444,
      "grad_norm": 2.341721144541663,
      "learning_rate": 6.829415236453439e-08,
      "loss": 2.6802,
      "step": 6680
    },
    {
      "epoch": 0.8648088811211444,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964267015457153,
      "eval_runtime": 14.1092,
      "eval_samples_per_second": 3.119,
      "eval_steps_per_second": 0.213,
      "step": 6680
    },
    {
      "epoch": 0.864938343528498,
      "grad_norm": 2.418843730853219,
      "learning_rate": 6.81658139548299e-08,
      "loss": 2.4199,
      "step": 6681
    },
    {
      "epoch": 0.8650678059358514,
      "grad_norm": 6.258683298257318,
      "learning_rate": 6.803758928979436e-08,
      "loss": 3.4131,
      "step": 6682
    },
    {
      "epoch": 0.8651972683432049,
      "grad_norm": 1.8592163292031774,
      "learning_rate": 6.790947839562044e-08,
      "loss": 2.353,
      "step": 6683
    },
    {
      "epoch": 0.8653267307505583,
      "grad_norm": 3.149712906113063,
      "learning_rate": 6.778148129847714e-08,
      "loss": 2.4346,
      "step": 6684
    },
    {
      "epoch": 0.8654561931579118,
      "grad_norm": 1.2635828765716453,
      "learning_rate": 6.76535980245107e-08,
      "loss": 2.2256,
      "step": 6685
    },
    {
      "epoch": 0.8654561931579118,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.0168,
      "eval_samples_per_second": 3.139,
      "eval_steps_per_second": 0.214,
      "step": 6685
    },
    {
      "epoch": 0.8655856555652652,
      "grad_norm": 2.123001861538419,
      "learning_rate": 6.752582859984362e-08,
      "loss": 2.4702,
      "step": 6686
    },
    {
      "epoch": 0.8657151179726187,
      "grad_norm": 1.4411008922578532,
      "learning_rate": 6.739817305057573e-08,
      "loss": 2.3286,
      "step": 6687
    },
    {
      "epoch": 0.8658445803799721,
      "grad_norm": 2.024492899125188,
      "learning_rate": 6.727063140278302e-08,
      "loss": 2.5676,
      "step": 6688
    },
    {
      "epoch": 0.8659740427873256,
      "grad_norm": 1.2306431037272239,
      "learning_rate": 6.714320368251835e-08,
      "loss": 2.2649,
      "step": 6689
    },
    {
      "epoch": 0.866103505194679,
      "grad_norm": 1.2588539105575596,
      "learning_rate": 6.701588991581174e-08,
      "loss": 2.394,
      "step": 6690
    },
    {
      "epoch": 0.866103505194679,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986017942428589,
      "eval_runtime": 15.1974,
      "eval_samples_per_second": 2.895,
      "eval_steps_per_second": 0.197,
      "step": 6690
    },
    {
      "epoch": 0.8662329676020326,
      "grad_norm": 1.426028810653981,
      "learning_rate": 6.68886901286693e-08,
      "loss": 2.4453,
      "step": 6691
    },
    {
      "epoch": 0.8663624300093861,
      "grad_norm": 2.3146376481760207,
      "learning_rate": 6.676160434707444e-08,
      "loss": 2.4929,
      "step": 6692
    },
    {
      "epoch": 0.8664918924167395,
      "grad_norm": 2.169533302902575,
      "learning_rate": 6.663463259698688e-08,
      "loss": 2.3918,
      "step": 6693
    },
    {
      "epoch": 0.866621354824093,
      "grad_norm": 3.3913541751770624,
      "learning_rate": 6.650777490434301e-08,
      "loss": 2.6057,
      "step": 6694
    },
    {
      "epoch": 0.8667508172314464,
      "grad_norm": 1.6259640231818908,
      "learning_rate": 6.638103129505621e-08,
      "loss": 2.21,
      "step": 6695
    },
    {
      "epoch": 0.8667508172314464,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4958717823028564,
      "eval_runtime": 14.3498,
      "eval_samples_per_second": 3.066,
      "eval_steps_per_second": 0.209,
      "step": 6695
    },
    {
      "epoch": 0.8668802796387999,
      "grad_norm": 1.8977601147542396,
      "learning_rate": 6.62544017950166e-08,
      "loss": 2.2977,
      "step": 6696
    },
    {
      "epoch": 0.8670097420461533,
      "grad_norm": 1.4919689917223402,
      "learning_rate": 6.612788643009058e-08,
      "loss": 2.541,
      "step": 6697
    },
    {
      "epoch": 0.8671392044535068,
      "grad_norm": 1.918917568538702,
      "learning_rate": 6.600148522612183e-08,
      "loss": 2.395,
      "step": 6698
    },
    {
      "epoch": 0.8672686668608602,
      "grad_norm": 2.445324326197817,
      "learning_rate": 6.587519820892986e-08,
      "loss": 2.5247,
      "step": 6699
    },
    {
      "epoch": 0.8673981292682137,
      "grad_norm": 1.8177422252010311,
      "learning_rate": 6.574902540431159e-08,
      "loss": 2.2876,
      "step": 6700
    },
    {
      "epoch": 0.8673981292682137,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971369504928589,
      "eval_runtime": 15.1819,
      "eval_samples_per_second": 2.898,
      "eval_steps_per_second": 0.198,
      "step": 6700
    },
    {
      "epoch": 0.8675275916755673,
      "grad_norm": 1.4113452390545302,
      "learning_rate": 6.56229668380405e-08,
      "loss": 2.3835,
      "step": 6701
    },
    {
      "epoch": 0.8676570540829207,
      "grad_norm": 2.9581728796058404,
      "learning_rate": 6.549702253586636e-08,
      "loss": 2.3403,
      "step": 6702
    },
    {
      "epoch": 0.8677865164902742,
      "grad_norm": 1.6860442297937641,
      "learning_rate": 6.537119252351613e-08,
      "loss": 2.2998,
      "step": 6703
    },
    {
      "epoch": 0.8679159788976276,
      "grad_norm": 3.3056083276683887,
      "learning_rate": 6.524547682669295e-08,
      "loss": 2.6072,
      "step": 6704
    },
    {
      "epoch": 0.8680454413049811,
      "grad_norm": 1.3738842325250389,
      "learning_rate": 6.511987547107669e-08,
      "loss": 2.2642,
      "step": 6705
    },
    {
      "epoch": 0.8680454413049811,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4956942796707153,
      "eval_runtime": 13.7513,
      "eval_samples_per_second": 3.2,
      "eval_steps_per_second": 0.218,
      "step": 6705
    },
    {
      "epoch": 0.8681749037123345,
      "grad_norm": 1.8744006896497827,
      "learning_rate": 6.499438848232433e-08,
      "loss": 2.4712,
      "step": 6706
    },
    {
      "epoch": 0.868304366119688,
      "grad_norm": 1.7345172057711467,
      "learning_rate": 6.486901588606875e-08,
      "loss": 2.5117,
      "step": 6707
    },
    {
      "epoch": 0.8684338285270414,
      "grad_norm": 4.480451394271401,
      "learning_rate": 6.474375770792011e-08,
      "loss": 2.3818,
      "step": 6708
    },
    {
      "epoch": 0.8685632909343949,
      "grad_norm": 3.3076567599340265,
      "learning_rate": 6.461861397346482e-08,
      "loss": 2.334,
      "step": 6709
    },
    {
      "epoch": 0.8686927533417483,
      "grad_norm": 2.23589586447218,
      "learning_rate": 6.449358470826596e-08,
      "loss": 2.2439,
      "step": 6710
    },
    {
      "epoch": 0.8686927533417483,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4965598583221436,
      "eval_runtime": 13.6012,
      "eval_samples_per_second": 3.235,
      "eval_steps_per_second": 0.221,
      "step": 6710
    },
    {
      "epoch": 0.8688222157491019,
      "grad_norm": 1.7633218572490406,
      "learning_rate": 6.436866993786352e-08,
      "loss": 2.521,
      "step": 6711
    },
    {
      "epoch": 0.8689516781564554,
      "grad_norm": 1.6903969599263293,
      "learning_rate": 6.424386968777355e-08,
      "loss": 2.365,
      "step": 6712
    },
    {
      "epoch": 0.8690811405638088,
      "grad_norm": 2.494849740128448,
      "learning_rate": 6.41191839834892e-08,
      "loss": 2.261,
      "step": 6713
    },
    {
      "epoch": 0.8692106029711623,
      "grad_norm": 1.7985873314679863,
      "learning_rate": 6.399461285048037e-08,
      "loss": 2.2898,
      "step": 6714
    },
    {
      "epoch": 0.8693400653785157,
      "grad_norm": 1.1845342386497755,
      "learning_rate": 6.38701563141926e-08,
      "loss": 2.2742,
      "step": 6715
    },
    {
      "epoch": 0.8693400653785157,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.493696689605713,
      "eval_runtime": 14.6525,
      "eval_samples_per_second": 3.003,
      "eval_steps_per_second": 0.205,
      "step": 6715
    },
    {
      "epoch": 0.8694695277858692,
      "grad_norm": 2.8367812605520326,
      "learning_rate": 6.374581440004912e-08,
      "loss": 2.4934,
      "step": 6716
    },
    {
      "epoch": 0.8695989901932226,
      "grad_norm": 1.297741587629704,
      "learning_rate": 6.362158713344911e-08,
      "loss": 2.3435,
      "step": 6717
    },
    {
      "epoch": 0.8697284526005761,
      "grad_norm": 1.0787507772483806,
      "learning_rate": 6.349747453976845e-08,
      "loss": 2.2175,
      "step": 6718
    },
    {
      "epoch": 0.8698579150079295,
      "grad_norm": 2.8228350545199863,
      "learning_rate": 6.337347664435993e-08,
      "loss": 2.395,
      "step": 6719
    },
    {
      "epoch": 0.8699873774152831,
      "grad_norm": 2.1531608728093468,
      "learning_rate": 6.324959347255244e-08,
      "loss": 2.4441,
      "step": 6720
    },
    {
      "epoch": 0.8699873774152831,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4956721067428589,
      "eval_runtime": 13.705,
      "eval_samples_per_second": 3.21,
      "eval_steps_per_second": 0.219,
      "step": 6720
    },
    {
      "epoch": 0.8701168398226365,
      "grad_norm": 1.5408867219279547,
      "learning_rate": 6.312582504965158e-08,
      "loss": 2.3599,
      "step": 6721
    },
    {
      "epoch": 0.87024630222999,
      "grad_norm": 2.5062948685091047,
      "learning_rate": 6.300217140093968e-08,
      "loss": 2.4946,
      "step": 6722
    },
    {
      "epoch": 0.8703757646373435,
      "grad_norm": 3.739691038319502,
      "learning_rate": 6.287863255167532e-08,
      "loss": 2.5645,
      "step": 6723
    },
    {
      "epoch": 0.8705052270446969,
      "grad_norm": 1.9680246115938134,
      "learning_rate": 6.27552085270941e-08,
      "loss": 2.517,
      "step": 6724
    },
    {
      "epoch": 0.8706346894520504,
      "grad_norm": 2.1199852540363575,
      "learning_rate": 6.263189935240756e-08,
      "loss": 2.2751,
      "step": 6725
    },
    {
      "epoch": 0.8706346894520504,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971591234207153,
      "eval_runtime": 14.1899,
      "eval_samples_per_second": 3.101,
      "eval_steps_per_second": 0.211,
      "step": 6725
    },
    {
      "epoch": 0.8707641518594038,
      "grad_norm": 1.9392810375556566,
      "learning_rate": 6.250870505280441e-08,
      "loss": 2.4363,
      "step": 6726
    },
    {
      "epoch": 0.8708936142667573,
      "grad_norm": 1.470228373888008,
      "learning_rate": 6.238562565344949e-08,
      "loss": 2.2487,
      "step": 6727
    },
    {
      "epoch": 0.8710230766741107,
      "grad_norm": 1.8541662871864155,
      "learning_rate": 6.226266117948396e-08,
      "loss": 2.5835,
      "step": 6728
    },
    {
      "epoch": 0.8711525390814642,
      "grad_norm": 1.3166860730606136,
      "learning_rate": 6.213981165602623e-08,
      "loss": 2.2578,
      "step": 6729
    },
    {
      "epoch": 0.8712820014888177,
      "grad_norm": 2.2233620269026555,
      "learning_rate": 6.201707710817052e-08,
      "loss": 2.2661,
      "step": 6730
    },
    {
      "epoch": 0.8712820014888177,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500688076019287,
      "eval_runtime": 14.6611,
      "eval_samples_per_second": 3.001,
      "eval_steps_per_second": 0.205,
      "step": 6730
    },
    {
      "epoch": 0.8714114638961712,
      "grad_norm": 2.01201731499581,
      "learning_rate": 6.189445756098811e-08,
      "loss": 2.4124,
      "step": 6731
    },
    {
      "epoch": 0.8715409263035246,
      "grad_norm": 1.4653296658699253,
      "learning_rate": 6.177195303952631e-08,
      "loss": 2.4182,
      "step": 6732
    },
    {
      "epoch": 0.8716703887108781,
      "grad_norm": 1.295969178583342,
      "learning_rate": 6.164956356880907e-08,
      "loss": 2.1621,
      "step": 6733
    },
    {
      "epoch": 0.8717998511182316,
      "grad_norm": 1.1827208234732027,
      "learning_rate": 6.152728917383728e-08,
      "loss": 2.2689,
      "step": 6734
    },
    {
      "epoch": 0.871929313525585,
      "grad_norm": 1.9456854403210608,
      "learning_rate": 6.140512987958751e-08,
      "loss": 2.3398,
      "step": 6735
    },
    {
      "epoch": 0.871929313525585,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4970703125,
      "eval_runtime": 14.3118,
      "eval_samples_per_second": 3.074,
      "eval_steps_per_second": 0.21,
      "step": 6735
    },
    {
      "epoch": 0.8720587759329385,
      "grad_norm": 1.4939730722222275,
      "learning_rate": 6.12830857110136e-08,
      "loss": 2.2427,
      "step": 6736
    },
    {
      "epoch": 0.8721882383402919,
      "grad_norm": 1.5103520766035068,
      "learning_rate": 6.116115669304569e-08,
      "loss": 2.4314,
      "step": 6737
    },
    {
      "epoch": 0.8723177007476454,
      "grad_norm": 1.8580683651981682,
      "learning_rate": 6.103934285058975e-08,
      "loss": 2.5381,
      "step": 6738
    },
    {
      "epoch": 0.8724471631549988,
      "grad_norm": 2.1825185467165467,
      "learning_rate": 6.091764420852908e-08,
      "loss": 2.3928,
      "step": 6739
    },
    {
      "epoch": 0.8725766255623524,
      "grad_norm": 1.994433186729432,
      "learning_rate": 6.079606079172296e-08,
      "loss": 2.3441,
      "step": 6740
    },
    {
      "epoch": 0.8725766255623524,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.5621,
      "eval_samples_per_second": 3.022,
      "eval_steps_per_second": 0.206,
      "step": 6740
    },
    {
      "epoch": 0.8727060879697058,
      "grad_norm": 1.641103674320105,
      "learning_rate": 6.067459262500728e-08,
      "loss": 2.3308,
      "step": 6741
    },
    {
      "epoch": 0.8728355503770593,
      "grad_norm": 1.631849082225562,
      "learning_rate": 6.055323973319469e-08,
      "loss": 2.1851,
      "step": 6742
    },
    {
      "epoch": 0.8729650127844127,
      "grad_norm": 1.616586959545823,
      "learning_rate": 6.043200214107342e-08,
      "loss": 2.2988,
      "step": 6743
    },
    {
      "epoch": 0.8730944751917662,
      "grad_norm": 1.8153477763680517,
      "learning_rate": 6.031087987340895e-08,
      "loss": 2.3359,
      "step": 6744
    },
    {
      "epoch": 0.8732239375991196,
      "grad_norm": 1.2726504952075033,
      "learning_rate": 6.018987295494313e-08,
      "loss": 2.3091,
      "step": 6745
    },
    {
      "epoch": 0.8732239375991196,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971591234207153,
      "eval_runtime": 13.9031,
      "eval_samples_per_second": 3.165,
      "eval_steps_per_second": 0.216,
      "step": 6745
    },
    {
      "epoch": 0.8733534000064731,
      "grad_norm": 4.465389909271955,
      "learning_rate": 6.006898141039386e-08,
      "loss": 2.4573,
      "step": 6746
    },
    {
      "epoch": 0.8734828624138266,
      "grad_norm": 2.4611450773474592,
      "learning_rate": 5.994820526445587e-08,
      "loss": 2.6506,
      "step": 6747
    },
    {
      "epoch": 0.87361232482118,
      "grad_norm": 1.7613136560798546,
      "learning_rate": 5.982754454180007e-08,
      "loss": 2.356,
      "step": 6748
    },
    {
      "epoch": 0.8737417872285335,
      "grad_norm": 2.404495628018289,
      "learning_rate": 5.970699926707361e-08,
      "loss": 2.4065,
      "step": 6749
    },
    {
      "epoch": 0.873871249635887,
      "grad_norm": 1.3126209330673706,
      "learning_rate": 5.958656946490074e-08,
      "loss": 2.3899,
      "step": 6750
    },
    {
      "epoch": 0.873871249635887,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.495161533355713,
      "eval_runtime": 15.6843,
      "eval_samples_per_second": 2.805,
      "eval_steps_per_second": 0.191,
      "step": 6750
    },
    {
      "epoch": 0.8740007120432405,
      "grad_norm": 2.2086063703509726,
      "learning_rate": 5.946625515988125e-08,
      "loss": 2.4753,
      "step": 6751
    },
    {
      "epoch": 0.8741301744505939,
      "grad_norm": 1.3743774949038223,
      "learning_rate": 5.9346056376592204e-08,
      "loss": 2.3318,
      "step": 6752
    },
    {
      "epoch": 0.8742596368579474,
      "grad_norm": 2.2715550593388385,
      "learning_rate": 5.922597313958632e-08,
      "loss": 2.533,
      "step": 6753
    },
    {
      "epoch": 0.8743890992653008,
      "grad_norm": 1.7538062983161529,
      "learning_rate": 5.910600547339308e-08,
      "loss": 2.156,
      "step": 6754
    },
    {
      "epoch": 0.8745185616726543,
      "grad_norm": 2.311886330068569,
      "learning_rate": 5.898615340251851e-08,
      "loss": 2.9456,
      "step": 6755
    },
    {
      "epoch": 0.8745185616726543,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.499311923980713,
      "eval_runtime": 14.5473,
      "eval_samples_per_second": 3.025,
      "eval_steps_per_second": 0.206,
      "step": 6755
    },
    {
      "epoch": 0.8746480240800077,
      "grad_norm": 1.321627514671924,
      "learning_rate": 5.886641695144454e-08,
      "loss": 2.3308,
      "step": 6756
    },
    {
      "epoch": 0.8747774864873612,
      "grad_norm": 1.6868890964968488,
      "learning_rate": 5.8746796144630026e-08,
      "loss": 2.4585,
      "step": 6757
    },
    {
      "epoch": 0.8749069488947147,
      "grad_norm": 1.869483961231482,
      "learning_rate": 5.862729100650971e-08,
      "loss": 2.387,
      "step": 6758
    },
    {
      "epoch": 0.8750364113020681,
      "grad_norm": 1.5965096175788616,
      "learning_rate": 5.850790156149519e-08,
      "loss": 2.4253,
      "step": 6759
    },
    {
      "epoch": 0.8751658737094217,
      "grad_norm": 2.01826973845739,
      "learning_rate": 5.838862783397403e-08,
      "loss": 2.3745,
      "step": 6760
    },
    {
      "epoch": 0.8751658737094217,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950062036514282,
      "eval_runtime": 14.3968,
      "eval_samples_per_second": 3.056,
      "eval_steps_per_second": 0.208,
      "step": 6760
    },
    {
      "epoch": 0.8752953361167751,
      "grad_norm": 1.2557740455679882,
      "learning_rate": 5.8269469848310224e-08,
      "loss": 2.2773,
      "step": 6761
    },
    {
      "epoch": 0.8754247985241286,
      "grad_norm": 1.2105687212851428,
      "learning_rate": 5.815042762884443e-08,
      "loss": 2.1643,
      "step": 6762
    },
    {
      "epoch": 0.875554260931482,
      "grad_norm": 2.3780251169723514,
      "learning_rate": 5.8031501199893174e-08,
      "loss": 2.6351,
      "step": 6763
    },
    {
      "epoch": 0.8756837233388355,
      "grad_norm": 1.3686426958774904,
      "learning_rate": 5.791269058574987e-08,
      "loss": 2.3872,
      "step": 6764
    },
    {
      "epoch": 0.8758131857461889,
      "grad_norm": 2.020189550657089,
      "learning_rate": 5.779399581068379e-08,
      "loss": 2.3906,
      "step": 6765
    },
    {
      "epoch": 0.8758131857461889,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.497114658355713,
      "eval_runtime": 15.1632,
      "eval_samples_per_second": 2.902,
      "eval_steps_per_second": 0.198,
      "step": 6765
    },
    {
      "epoch": 0.8759426481535424,
      "grad_norm": 1.972570207724478,
      "learning_rate": 5.767541689894074e-08,
      "loss": 2.3276,
      "step": 6766
    },
    {
      "epoch": 0.8760721105608958,
      "grad_norm": 3.690751774115214,
      "learning_rate": 5.755695387474294e-08,
      "loss": 2.2214,
      "step": 6767
    },
    {
      "epoch": 0.8762015729682493,
      "grad_norm": 2.1374501330175417,
      "learning_rate": 5.7438606762288964e-08,
      "loss": 2.425,
      "step": 6768
    },
    {
      "epoch": 0.8763310353756029,
      "grad_norm": 2.070583221051381,
      "learning_rate": 5.7320375585753416e-08,
      "loss": 2.407,
      "step": 6769
    },
    {
      "epoch": 0.8764604977829563,
      "grad_norm": 2.1761599122714403,
      "learning_rate": 5.720226036928768e-08,
      "loss": 2.5735,
      "step": 6770
    },
    {
      "epoch": 0.8764604977829563,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498579502105713,
      "eval_runtime": 14.1958,
      "eval_samples_per_second": 3.1,
      "eval_steps_per_second": 0.211,
      "step": 6770
    },
    {
      "epoch": 0.8765899601903098,
      "grad_norm": 1.7398946188110953,
      "learning_rate": 5.708426113701877e-08,
      "loss": 2.405,
      "step": 6771
    },
    {
      "epoch": 0.8767194225976632,
      "grad_norm": 1.3869566375129447,
      "learning_rate": 5.696637791305068e-08,
      "loss": 2.3115,
      "step": 6772
    },
    {
      "epoch": 0.8768488850050167,
      "grad_norm": 1.8084579157259326,
      "learning_rate": 5.684861072146351e-08,
      "loss": 2.7219,
      "step": 6773
    },
    {
      "epoch": 0.8769783474123701,
      "grad_norm": 6.700630691213659,
      "learning_rate": 5.673095958631338e-08,
      "loss": 3.1399,
      "step": 6774
    },
    {
      "epoch": 0.8771078098197236,
      "grad_norm": 1.9381461751639362,
      "learning_rate": 5.6613424531633197e-08,
      "loss": 2.6365,
      "step": 6775
    },
    {
      "epoch": 0.8771078098197236,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.49853515625,
      "eval_runtime": 13.7755,
      "eval_samples_per_second": 3.194,
      "eval_steps_per_second": 0.218,
      "step": 6775
    },
    {
      "epoch": 0.877237272227077,
      "grad_norm": 1.7307214565159195,
      "learning_rate": 5.649600558143169e-08,
      "loss": 2.6199,
      "step": 6776
    },
    {
      "epoch": 0.8773667346344305,
      "grad_norm": 2.3790148434202543,
      "learning_rate": 5.6378702759693966e-08,
      "loss": 2.4351,
      "step": 6777
    },
    {
      "epoch": 0.8774961970417839,
      "grad_norm": 3.106152250041979,
      "learning_rate": 5.626151609038177e-08,
      "loss": 2.418,
      "step": 6778
    },
    {
      "epoch": 0.8776256594491375,
      "grad_norm": 1.4183043220992424,
      "learning_rate": 5.614444559743262e-08,
      "loss": 2.2131,
      "step": 6779
    },
    {
      "epoch": 0.877755121856491,
      "grad_norm": 1.2146629735113779,
      "learning_rate": 5.602749130476072e-08,
      "loss": 2.2922,
      "step": 6780
    },
    {
      "epoch": 0.877755121856491,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957386255264282,
      "eval_runtime": 13.7841,
      "eval_samples_per_second": 3.192,
      "eval_steps_per_second": 0.218,
      "step": 6780
    },
    {
      "epoch": 0.8778845842638444,
      "grad_norm": 1.4819708460149876,
      "learning_rate": 5.591065323625626e-08,
      "loss": 2.4458,
      "step": 6781
    },
    {
      "epoch": 0.8780140466711979,
      "grad_norm": 3.34377513442525,
      "learning_rate": 5.579393141578571e-08,
      "loss": 2.4927,
      "step": 6782
    },
    {
      "epoch": 0.8781435090785513,
      "grad_norm": 3.0271503684934227,
      "learning_rate": 5.567732586719207e-08,
      "loss": 2.2695,
      "step": 6783
    },
    {
      "epoch": 0.8782729714859048,
      "grad_norm": 1.5971368293002015,
      "learning_rate": 5.556083661429415e-08,
      "loss": 2.394,
      "step": 6784
    },
    {
      "epoch": 0.8784024338932582,
      "grad_norm": 2.195654431220393,
      "learning_rate": 5.544446368088742e-08,
      "loss": 2.2937,
      "step": 6785
    },
    {
      "epoch": 0.8784024338932582,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978915452957153,
      "eval_runtime": 14.5783,
      "eval_samples_per_second": 3.018,
      "eval_steps_per_second": 0.206,
      "step": 6785
    },
    {
      "epoch": 0.8785318963006117,
      "grad_norm": 1.4724710394139269,
      "learning_rate": 5.532820709074328e-08,
      "loss": 2.2842,
      "step": 6786
    },
    {
      "epoch": 0.8786613587079651,
      "grad_norm": 1.2931475058509387,
      "learning_rate": 5.5212066867609687e-08,
      "loss": 2.2429,
      "step": 6787
    },
    {
      "epoch": 0.8787908211153186,
      "grad_norm": 1.5520712508654095,
      "learning_rate": 5.5096043035210435e-08,
      "loss": 2.1504,
      "step": 6788
    },
    {
      "epoch": 0.8789202835226722,
      "grad_norm": 1.6392619538079143,
      "learning_rate": 5.49801356172458e-08,
      "loss": 2.3875,
      "step": 6789
    },
    {
      "epoch": 0.8790497459300256,
      "grad_norm": 1.202033276643043,
      "learning_rate": 5.486434463739219e-08,
      "loss": 2.342,
      "step": 6790
    },
    {
      "epoch": 0.8790497459300256,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986017942428589,
      "eval_runtime": 14.3802,
      "eval_samples_per_second": 3.06,
      "eval_steps_per_second": 0.209,
      "step": 6790
    },
    {
      "epoch": 0.8791792083373791,
      "grad_norm": 1.7293585132868499,
      "learning_rate": 5.474867011930243e-08,
      "loss": 2.1604,
      "step": 6791
    },
    {
      "epoch": 0.8793086707447325,
      "grad_norm": 1.4487881241094938,
      "learning_rate": 5.4633112086605254e-08,
      "loss": 2.2876,
      "step": 6792
    },
    {
      "epoch": 0.879438133152086,
      "grad_norm": 2.048277615854377,
      "learning_rate": 5.45176705629058e-08,
      "loss": 2.478,
      "step": 6793
    },
    {
      "epoch": 0.8795675955594394,
      "grad_norm": 2.380244877642461,
      "learning_rate": 5.4402345571785145e-08,
      "loss": 2.2944,
      "step": 6794
    },
    {
      "epoch": 0.8796970579667929,
      "grad_norm": 1.489663261776139,
      "learning_rate": 5.4287137136800826e-08,
      "loss": 2.2495,
      "step": 6795
    },
    {
      "epoch": 0.8796970579667929,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.493607997894287,
      "eval_runtime": 14.571,
      "eval_samples_per_second": 3.02,
      "eval_steps_per_second": 0.206,
      "step": 6795
    },
    {
      "epoch": 0.8798265203741463,
      "grad_norm": 2.9769008748587784,
      "learning_rate": 5.41720452814868e-08,
      "loss": 2.3701,
      "step": 6796
    },
    {
      "epoch": 0.8799559827814998,
      "grad_norm": 2.3774783432529976,
      "learning_rate": 5.405707002935245e-08,
      "loss": 2.6128,
      "step": 6797
    },
    {
      "epoch": 0.8800854451888532,
      "grad_norm": 2.796212901722403,
      "learning_rate": 5.3942211403884214e-08,
      "loss": 2.4675,
      "step": 6798
    },
    {
      "epoch": 0.8802149075962068,
      "grad_norm": 1.3879262284780065,
      "learning_rate": 5.382746942854414e-08,
      "loss": 2.2344,
      "step": 6799
    },
    {
      "epoch": 0.8803443700035603,
      "grad_norm": 1.2444795640800748,
      "learning_rate": 5.371284412677043e-08,
      "loss": 2.4526,
      "step": 6800
    },
    {
      "epoch": 0.8803443700035603,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4963600635528564,
      "eval_runtime": 14.5985,
      "eval_samples_per_second": 3.014,
      "eval_steps_per_second": 0.206,
      "step": 6800
    },
    {
      "epoch": 0.8804738324109137,
      "grad_norm": 1.272375416587221,
      "learning_rate": 5.3598335521977826e-08,
      "loss": 2.25,
      "step": 6801
    },
    {
      "epoch": 0.8806032948182672,
      "grad_norm": 1.8228716382863823,
      "learning_rate": 5.348394363755692e-08,
      "loss": 2.4907,
      "step": 6802
    },
    {
      "epoch": 0.8807327572256206,
      "grad_norm": 1.4814843194662073,
      "learning_rate": 5.336966849687472e-08,
      "loss": 2.2861,
      "step": 6803
    },
    {
      "epoch": 0.8808622196329741,
      "grad_norm": 2.2646095197360254,
      "learning_rate": 5.3255510123274224e-08,
      "loss": 2.2917,
      "step": 6804
    },
    {
      "epoch": 0.8809916820403275,
      "grad_norm": 1.812269115029419,
      "learning_rate": 5.314146854007433e-08,
      "loss": 2.2893,
      "step": 6805
    },
    {
      "epoch": 0.8809916820403275,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4956942796707153,
      "eval_runtime": 14.3365,
      "eval_samples_per_second": 3.069,
      "eval_steps_per_second": 0.209,
      "step": 6805
    },
    {
      "epoch": 0.881121144447681,
      "grad_norm": 1.4031238110066964,
      "learning_rate": 5.302754377057066e-08,
      "loss": 2.3474,
      "step": 6806
    },
    {
      "epoch": 0.8812506068550344,
      "grad_norm": 2.0274924805058334,
      "learning_rate": 5.2913735838034385e-08,
      "loss": 2.4805,
      "step": 6807
    },
    {
      "epoch": 0.881380069262388,
      "grad_norm": 1.7515371439213783,
      "learning_rate": 5.2800044765713295e-08,
      "loss": 2.1887,
      "step": 6808
    },
    {
      "epoch": 0.8815095316697414,
      "grad_norm": 2.430880768328812,
      "learning_rate": 5.268647057683117e-08,
      "loss": 2.6685,
      "step": 6809
    },
    {
      "epoch": 0.8816389940770949,
      "grad_norm": 1.3275899288501505,
      "learning_rate": 5.2573013294587515e-08,
      "loss": 2.4717,
      "step": 6810
    },
    {
      "epoch": 0.8816389940770949,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972256422042847,
      "eval_runtime": 13.8922,
      "eval_samples_per_second": 3.167,
      "eval_steps_per_second": 0.216,
      "step": 6810
    },
    {
      "epoch": 0.8817684564844483,
      "grad_norm": 1.852054907065753,
      "learning_rate": 5.2459672942158575e-08,
      "loss": 2.3359,
      "step": 6811
    },
    {
      "epoch": 0.8818979188918018,
      "grad_norm": 1.697480310156964,
      "learning_rate": 5.2346449542696185e-08,
      "loss": 2.2881,
      "step": 6812
    },
    {
      "epoch": 0.8820273812991553,
      "grad_norm": 2.202499180085729,
      "learning_rate": 5.223334311932865e-08,
      "loss": 2.4719,
      "step": 6813
    },
    {
      "epoch": 0.8821568437065087,
      "grad_norm": 1.5761265031072722,
      "learning_rate": 5.212035369516033e-08,
      "loss": 2.4033,
      "step": 6814
    },
    {
      "epoch": 0.8822863061138622,
      "grad_norm": 2.0088869742015913,
      "learning_rate": 5.2007481293271514e-08,
      "loss": 2.4524,
      "step": 6815
    },
    {
      "epoch": 0.8822863061138622,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972922801971436,
      "eval_runtime": 14.346,
      "eval_samples_per_second": 3.067,
      "eval_steps_per_second": 0.209,
      "step": 6815
    },
    {
      "epoch": 0.8824157685212156,
      "grad_norm": 2.2596026701442646,
      "learning_rate": 5.189472593671856e-08,
      "loss": 2.6572,
      "step": 6816
    },
    {
      "epoch": 0.8825452309285691,
      "grad_norm": 1.9062682312606882,
      "learning_rate": 5.1782087648534275e-08,
      "loss": 2.3262,
      "step": 6817
    },
    {
      "epoch": 0.8826746933359226,
      "grad_norm": 1.3008849475205415,
      "learning_rate": 5.1669566451727005e-08,
      "loss": 2.1953,
      "step": 6818
    },
    {
      "epoch": 0.8828041557432761,
      "grad_norm": 1.5609962547581708,
      "learning_rate": 5.1557162369281836e-08,
      "loss": 2.2661,
      "step": 6819
    },
    {
      "epoch": 0.8829336181506295,
      "grad_norm": 2.4739657629797867,
      "learning_rate": 5.144487542415942e-08,
      "loss": 2.4871,
      "step": 6820
    },
    {
      "epoch": 0.8829336181506295,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4977805614471436,
      "eval_runtime": 13.637,
      "eval_samples_per_second": 3.227,
      "eval_steps_per_second": 0.22,
      "step": 6820
    },
    {
      "epoch": 0.883063080557983,
      "grad_norm": 1.6040042906766272,
      "learning_rate": 5.1332705639296485e-08,
      "loss": 2.4102,
      "step": 6821
    },
    {
      "epoch": 0.8831925429653364,
      "grad_norm": 1.6887012402620727,
      "learning_rate": 5.12206530376063e-08,
      "loss": 2.3594,
      "step": 6822
    },
    {
      "epoch": 0.8833220053726899,
      "grad_norm": 1.3405311912088704,
      "learning_rate": 5.1108717641977504e-08,
      "loss": 2.3254,
      "step": 6823
    },
    {
      "epoch": 0.8834514677800434,
      "grad_norm": 2.5960548001675563,
      "learning_rate": 5.099689947527554e-08,
      "loss": 2.3899,
      "step": 6824
    },
    {
      "epoch": 0.8835809301873968,
      "grad_norm": 1.2979289484318808,
      "learning_rate": 5.088519856034126e-08,
      "loss": 2.3182,
      "step": 6825
    },
    {
      "epoch": 0.8835809301873968,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964710474014282,
      "eval_runtime": 14.2333,
      "eval_samples_per_second": 3.091,
      "eval_steps_per_second": 0.211,
      "step": 6825
    },
    {
      "epoch": 0.8837103925947503,
      "grad_norm": 1.8178412323124833,
      "learning_rate": 5.077361491999208e-08,
      "loss": 2.1327,
      "step": 6826
    },
    {
      "epoch": 0.8838398550021037,
      "grad_norm": 1.5244985575952903,
      "learning_rate": 5.0662148577021106e-08,
      "loss": 2.3545,
      "step": 6827
    },
    {
      "epoch": 0.8839693174094573,
      "grad_norm": 2.362733321399371,
      "learning_rate": 5.055079955419741e-08,
      "loss": 2.3171,
      "step": 6828
    },
    {
      "epoch": 0.8840987798168107,
      "grad_norm": 1.206699504106358,
      "learning_rate": 5.043956787426661e-08,
      "loss": 2.4153,
      "step": 6829
    },
    {
      "epoch": 0.8842282422241642,
      "grad_norm": 1.5704394450802655,
      "learning_rate": 5.032845355994978e-08,
      "loss": 2.3965,
      "step": 6830
    },
    {
      "epoch": 0.8842282422241642,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4956942796707153,
      "eval_runtime": 13.5494,
      "eval_samples_per_second": 3.247,
      "eval_steps_per_second": 0.221,
      "step": 6830
    },
    {
      "epoch": 0.8843577046315176,
      "grad_norm": 1.5112196749551587,
      "learning_rate": 5.021745663394445e-08,
      "loss": 2.3868,
      "step": 6831
    },
    {
      "epoch": 0.8844871670388711,
      "grad_norm": 1.3908879176411566,
      "learning_rate": 5.010657711892394e-08,
      "loss": 2.3821,
      "step": 6832
    },
    {
      "epoch": 0.8846166294462245,
      "grad_norm": 1.3499601569912798,
      "learning_rate": 4.9995815037537433e-08,
      "loss": 2.4922,
      "step": 6833
    },
    {
      "epoch": 0.884746091853578,
      "grad_norm": 1.6594522414697797,
      "learning_rate": 4.9885170412410715e-08,
      "loss": 2.4341,
      "step": 6834
    },
    {
      "epoch": 0.8848755542609315,
      "grad_norm": 1.2013526688641463,
      "learning_rate": 4.9774643266144796e-08,
      "loss": 2.3118,
      "step": 6835
    },
    {
      "epoch": 0.8848755542609315,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964488744735718,
      "eval_runtime": 14.9143,
      "eval_samples_per_second": 2.95,
      "eval_steps_per_second": 0.201,
      "step": 6835
    },
    {
      "epoch": 0.8850050166682849,
      "grad_norm": 1.6973190897588857,
      "learning_rate": 4.966423362131732e-08,
      "loss": 2.262,
      "step": 6836
    },
    {
      "epoch": 0.8851344790756384,
      "grad_norm": 2.164161001163814,
      "learning_rate": 4.9553941500481917e-08,
      "loss": 2.4458,
      "step": 6837
    },
    {
      "epoch": 0.8852639414829919,
      "grad_norm": 1.3206649933939947,
      "learning_rate": 4.9443766926167425e-08,
      "loss": 2.5044,
      "step": 6838
    },
    {
      "epoch": 0.8853934038903454,
      "grad_norm": 1.5635126332503226,
      "learning_rate": 4.933370992087953e-08,
      "loss": 2.3807,
      "step": 6839
    },
    {
      "epoch": 0.8855228662976988,
      "grad_norm": 2.674186711804766,
      "learning_rate": 4.922377050709977e-08,
      "loss": 2.4448,
      "step": 6840
    },
    {
      "epoch": 0.8855228662976988,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978693723678589,
      "eval_runtime": 13.5552,
      "eval_samples_per_second": 3.246,
      "eval_steps_per_second": 0.221,
      "step": 6840
    },
    {
      "epoch": 0.8856523287050523,
      "grad_norm": 1.1852061910867218,
      "learning_rate": 4.911394870728524e-08,
      "loss": 2.3462,
      "step": 6841
    },
    {
      "epoch": 0.8857817911124057,
      "grad_norm": 1.6833025405061035,
      "learning_rate": 4.9004244543869524e-08,
      "loss": 2.561,
      "step": 6842
    },
    {
      "epoch": 0.8859112535197592,
      "grad_norm": 1.6137981062566404,
      "learning_rate": 4.889465803926158e-08,
      "loss": 2.3879,
      "step": 6843
    },
    {
      "epoch": 0.8860407159271126,
      "grad_norm": 1.4002543212038663,
      "learning_rate": 4.8785189215846826e-08,
      "loss": 2.4956,
      "step": 6844
    },
    {
      "epoch": 0.8861701783344661,
      "grad_norm": 2.259311525048003,
      "learning_rate": 4.867583809598668e-08,
      "loss": 2.415,
      "step": 6845
    },
    {
      "epoch": 0.8861701783344661,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4921208620071411,
      "eval_runtime": 13.4811,
      "eval_samples_per_second": 3.264,
      "eval_steps_per_second": 0.223,
      "step": 6845
    },
    {
      "epoch": 0.8862996407418195,
      "grad_norm": 1.1739139988353178,
      "learning_rate": 4.8566604702018e-08,
      "loss": 2.0773,
      "step": 6846
    },
    {
      "epoch": 0.8864291031491731,
      "grad_norm": 1.6920353589218537,
      "learning_rate": 4.845748905625419e-08,
      "loss": 2.2161,
      "step": 6847
    },
    {
      "epoch": 0.8865585655565266,
      "grad_norm": 2.1925823450327813,
      "learning_rate": 4.8348491180984224e-08,
      "loss": 2.3091,
      "step": 6848
    },
    {
      "epoch": 0.88668802796388,
      "grad_norm": 1.7912374780015778,
      "learning_rate": 4.823961109847307e-08,
      "loss": 2.3779,
      "step": 6849
    },
    {
      "epoch": 0.8868174903712335,
      "grad_norm": 1.5919658250717248,
      "learning_rate": 4.813084883096182e-08,
      "loss": 2.3093,
      "step": 6850
    },
    {
      "epoch": 0.8868174903712335,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986683130264282,
      "eval_runtime": 14.755,
      "eval_samples_per_second": 2.982,
      "eval_steps_per_second": 0.203,
      "step": 6850
    },
    {
      "epoch": 0.8869469527785869,
      "grad_norm": 2.4835319244556575,
      "learning_rate": 4.802220440066723e-08,
      "loss": 2.4106,
      "step": 6851
    },
    {
      "epoch": 0.8870764151859404,
      "grad_norm": 1.9092603332812002,
      "learning_rate": 4.791367782978222e-08,
      "loss": 2.6543,
      "step": 6852
    },
    {
      "epoch": 0.8872058775932938,
      "grad_norm": 1.438123227684614,
      "learning_rate": 4.7805269140475464e-08,
      "loss": 2.4385,
      "step": 6853
    },
    {
      "epoch": 0.8873353400006473,
      "grad_norm": 1.409676616809633,
      "learning_rate": 4.7696978354891816e-08,
      "loss": 2.395,
      "step": 6854
    },
    {
      "epoch": 0.8874648024080007,
      "grad_norm": 1.9197598074861877,
      "learning_rate": 4.75888054951517e-08,
      "loss": 2.2854,
      "step": 6855
    },
    {
      "epoch": 0.8874648024080007,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950283765792847,
      "eval_runtime": 13.7904,
      "eval_samples_per_second": 3.191,
      "eval_steps_per_second": 0.218,
      "step": 6855
    },
    {
      "epoch": 0.8875942648153542,
      "grad_norm": 1.5964177352374567,
      "learning_rate": 4.748075058335161e-08,
      "loss": 2.4468,
      "step": 6856
    },
    {
      "epoch": 0.8877237272227078,
      "grad_norm": 2.18707695800178,
      "learning_rate": 4.737281364156403e-08,
      "loss": 2.4023,
      "step": 6857
    },
    {
      "epoch": 0.8878531896300612,
      "grad_norm": 1.5138236784357733,
      "learning_rate": 4.7264994691837153e-08,
      "loss": 2.2424,
      "step": 6858
    },
    {
      "epoch": 0.8879826520374147,
      "grad_norm": 1.9180195132690427,
      "learning_rate": 4.7157293756195375e-08,
      "loss": 2.314,
      "step": 6859
    },
    {
      "epoch": 0.8881121144447681,
      "grad_norm": 1.74251428073698,
      "learning_rate": 4.704971085663867e-08,
      "loss": 2.4175,
      "step": 6860
    },
    {
      "epoch": 0.8881121144447681,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.496382236480713,
      "eval_runtime": 14.1415,
      "eval_samples_per_second": 3.111,
      "eval_steps_per_second": 0.212,
      "step": 6860
    },
    {
      "epoch": 0.8882415768521216,
      "grad_norm": 2.451772939019097,
      "learning_rate": 4.694224601514301e-08,
      "loss": 2.7568,
      "step": 6861
    },
    {
      "epoch": 0.888371039259475,
      "grad_norm": 2.1808859934039786,
      "learning_rate": 4.6834899253660195e-08,
      "loss": 2.2068,
      "step": 6862
    },
    {
      "epoch": 0.8885005016668285,
      "grad_norm": 1.9708244448495695,
      "learning_rate": 4.672767059411824e-08,
      "loss": 2.5842,
      "step": 6863
    },
    {
      "epoch": 0.8886299640741819,
      "grad_norm": 1.5567832791817082,
      "learning_rate": 4.662056005842054e-08,
      "loss": 2.5952,
      "step": 6864
    },
    {
      "epoch": 0.8887594264815354,
      "grad_norm": 2.109379645225532,
      "learning_rate": 4.651356766844681e-08,
      "loss": 2.4111,
      "step": 6865
    },
    {
      "epoch": 0.8887594264815354,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.49658203125,
      "eval_runtime": 14.3206,
      "eval_samples_per_second": 3.072,
      "eval_steps_per_second": 0.209,
      "step": 6865
    },
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 1.492008547478116,
      "learning_rate": 4.640669344605206e-08,
      "loss": 2.3145,
      "step": 6866
    },
    {
      "epoch": 0.8890183512962424,
      "grad_norm": 2.4863475947710816,
      "learning_rate": 4.6299937413067794e-08,
      "loss": 2.5022,
      "step": 6867
    },
    {
      "epoch": 0.8891478137035959,
      "grad_norm": 1.8362858630475052,
      "learning_rate": 4.6193299591301076e-08,
      "loss": 2.2554,
      "step": 6868
    },
    {
      "epoch": 0.8892772761109493,
      "grad_norm": 1.7880404505061087,
      "learning_rate": 4.608678000253469e-08,
      "loss": 2.3367,
      "step": 6869
    },
    {
      "epoch": 0.8894067385183028,
      "grad_norm": 1.8306178102149797,
      "learning_rate": 4.598037866852763e-08,
      "loss": 2.5474,
      "step": 6870
    },
    {
      "epoch": 0.8894067385183028,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979358911514282,
      "eval_runtime": 14.57,
      "eval_samples_per_second": 3.02,
      "eval_steps_per_second": 0.206,
      "step": 6870
    },
    {
      "epoch": 0.8895362009256562,
      "grad_norm": 1.5091112829606663,
      "learning_rate": 4.587409561101444e-08,
      "loss": 2.365,
      "step": 6871
    },
    {
      "epoch": 0.8896656633330097,
      "grad_norm": 1.3385445182413755,
      "learning_rate": 4.576793085170547e-08,
      "loss": 2.2312,
      "step": 6872
    },
    {
      "epoch": 0.8897951257403631,
      "grad_norm": 2.5138482603010623,
      "learning_rate": 4.56618844122872e-08,
      "loss": 2.2615,
      "step": 6873
    },
    {
      "epoch": 0.8899245881477166,
      "grad_norm": 1.880052318002391,
      "learning_rate": 4.555595631442158e-08,
      "loss": 2.4336,
      "step": 6874
    },
    {
      "epoch": 0.89005405055507,
      "grad_norm": 1.424451958521749,
      "learning_rate": 4.545014657974682e-08,
      "loss": 2.407,
      "step": 6875
    },
    {
      "epoch": 0.89005405055507,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4944514036178589,
      "eval_runtime": 15.2115,
      "eval_samples_per_second": 2.893,
      "eval_steps_per_second": 0.197,
      "step": 6875
    },
    {
      "epoch": 0.8901835129624235,
      "grad_norm": 1.4429270918344927,
      "learning_rate": 4.534445522987652e-08,
      "loss": 2.3318,
      "step": 6876
    },
    {
      "epoch": 0.890312975369777,
      "grad_norm": 2.26301412988003,
      "learning_rate": 4.5238882286400204e-08,
      "loss": 2.3582,
      "step": 6877
    },
    {
      "epoch": 0.8904424377771305,
      "grad_norm": 1.7604419801924598,
      "learning_rate": 4.513342777088356e-08,
      "loss": 2.4346,
      "step": 6878
    },
    {
      "epoch": 0.890571900184484,
      "grad_norm": 1.5066813985924459,
      "learning_rate": 4.502809170486754e-08,
      "loss": 2.3643,
      "step": 6879
    },
    {
      "epoch": 0.8907013625918374,
      "grad_norm": 2.55366696627522,
      "learning_rate": 4.492287410986937e-08,
      "loss": 2.5508,
      "step": 6880
    },
    {
      "epoch": 0.8907013625918374,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978915452957153,
      "eval_runtime": 13.7504,
      "eval_samples_per_second": 3.2,
      "eval_steps_per_second": 0.218,
      "step": 6880
    },
    {
      "epoch": 0.8908308249991909,
      "grad_norm": 1.5881197527451938,
      "learning_rate": 4.4817775007381795e-08,
      "loss": 2.2556,
      "step": 6881
    },
    {
      "epoch": 0.8909602874065443,
      "grad_norm": 1.4310554373910798,
      "learning_rate": 4.471279441887334e-08,
      "loss": 2.397,
      "step": 6882
    },
    {
      "epoch": 0.8910897498138978,
      "grad_norm": 1.9032917331131038,
      "learning_rate": 4.460793236578859e-08,
      "loss": 2.4518,
      "step": 6883
    },
    {
      "epoch": 0.8912192122212512,
      "grad_norm": 2.1307513203863855,
      "learning_rate": 4.450318886954762e-08,
      "loss": 2.2865,
      "step": 6884
    },
    {
      "epoch": 0.8913486746286047,
      "grad_norm": 2.7305593681769027,
      "learning_rate": 4.439856395154637e-08,
      "loss": 2.2349,
      "step": 6885
    },
    {
      "epoch": 0.8913486746286047,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957386255264282,
      "eval_runtime": 13.6619,
      "eval_samples_per_second": 3.221,
      "eval_steps_per_second": 0.22,
      "step": 6885
    },
    {
      "epoch": 0.8914781370359581,
      "grad_norm": 1.641322534487582,
      "learning_rate": 4.4294057633156864e-08,
      "loss": 2.1875,
      "step": 6886
    },
    {
      "epoch": 0.8916075994433117,
      "grad_norm": 1.350728645644694,
      "learning_rate": 4.4189669935726446e-08,
      "loss": 2.2871,
      "step": 6887
    },
    {
      "epoch": 0.8917370618506651,
      "grad_norm": 1.513077226970648,
      "learning_rate": 4.408540088057842e-08,
      "loss": 2.5142,
      "step": 6888
    },
    {
      "epoch": 0.8918665242580186,
      "grad_norm": 1.6532640452882146,
      "learning_rate": 4.3981250489011776e-08,
      "loss": 2.145,
      "step": 6889
    },
    {
      "epoch": 0.891995986665372,
      "grad_norm": 1.4154330168458797,
      "learning_rate": 4.387721878230139e-08,
      "loss": 2.3655,
      "step": 6890
    },
    {
      "epoch": 0.891995986665372,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4949840307235718,
      "eval_runtime": 14.4213,
      "eval_samples_per_second": 3.051,
      "eval_steps_per_second": 0.208,
      "step": 6890
    },
    {
      "epoch": 0.8921254490727255,
      "grad_norm": 1.806321323102692,
      "learning_rate": 4.3773305781698026e-08,
      "loss": 2.2756,
      "step": 6891
    },
    {
      "epoch": 0.892254911480079,
      "grad_norm": 1.2895354054768229,
      "learning_rate": 4.3669511508427906e-08,
      "loss": 2.4072,
      "step": 6892
    },
    {
      "epoch": 0.8923843738874324,
      "grad_norm": 1.6770658669718612,
      "learning_rate": 4.356583598369296e-08,
      "loss": 2.5588,
      "step": 6893
    },
    {
      "epoch": 0.8925138362947859,
      "grad_norm": 1.1178914600945362,
      "learning_rate": 4.3462279228671243e-08,
      "loss": 2.2297,
      "step": 6894
    },
    {
      "epoch": 0.8926432987021393,
      "grad_norm": 1.658333578972004,
      "learning_rate": 4.335884126451612e-08,
      "loss": 2.3457,
      "step": 6895
    },
    {
      "epoch": 0.8926432987021393,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.7287,
      "eval_samples_per_second": 2.987,
      "eval_steps_per_second": 0.204,
      "step": 6895
    },
    {
      "epoch": 0.8927727611094929,
      "grad_norm": 1.533056140914135,
      "learning_rate": 4.3255522112357134e-08,
      "loss": 2.3196,
      "step": 6896
    },
    {
      "epoch": 0.8929022235168463,
      "grad_norm": 1.5024013722088398,
      "learning_rate": 4.315232179329908e-08,
      "loss": 2.4712,
      "step": 6897
    },
    {
      "epoch": 0.8930316859241998,
      "grad_norm": 1.811341991406139,
      "learning_rate": 4.3049240328422865e-08,
      "loss": 2.1892,
      "step": 6898
    },
    {
      "epoch": 0.8931611483315532,
      "grad_norm": 3.9388349331388235,
      "learning_rate": 4.294627773878499e-08,
      "loss": 2.7144,
      "step": 6899
    },
    {
      "epoch": 0.8932906107389067,
      "grad_norm": 1.403469092732384,
      "learning_rate": 4.284343404541744e-08,
      "loss": 2.24,
      "step": 6900
    },
    {
      "epoch": 0.8932906107389067,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978915452957153,
      "eval_runtime": 14.1089,
      "eval_samples_per_second": 3.119,
      "eval_steps_per_second": 0.213,
      "step": 6900
    },
    {
      "epoch": 0.8934200731462602,
      "grad_norm": 3.7223479338226206,
      "learning_rate": 4.274070926932834e-08,
      "loss": 2.4541,
      "step": 6901
    },
    {
      "epoch": 0.8935495355536136,
      "grad_norm": 1.3644230410435525,
      "learning_rate": 4.2638103431501126e-08,
      "loss": 2.2668,
      "step": 6902
    },
    {
      "epoch": 0.8936789979609671,
      "grad_norm": 1.609142453978087,
      "learning_rate": 4.253561655289527e-08,
      "loss": 2.2925,
      "step": 6903
    },
    {
      "epoch": 0.8938084603683205,
      "grad_norm": 2.4635030186402402,
      "learning_rate": 4.243324865444591e-08,
      "loss": 2.1685,
      "step": 6904
    },
    {
      "epoch": 0.893937922775674,
      "grad_norm": 3.4963444964201673,
      "learning_rate": 4.2330999757063463e-08,
      "loss": 2.6826,
      "step": 6905
    },
    {
      "epoch": 0.893937922775674,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.3624,
      "eval_samples_per_second": 3.064,
      "eval_steps_per_second": 0.209,
      "step": 6905
    },
    {
      "epoch": 0.8940673851830275,
      "grad_norm": 1.8002477917651825,
      "learning_rate": 4.2228869881634496e-08,
      "loss": 2.2915,
      "step": 6906
    },
    {
      "epoch": 0.894196847590381,
      "grad_norm": 3.193912068537218,
      "learning_rate": 4.2126859049021054e-08,
      "loss": 2.4922,
      "step": 6907
    },
    {
      "epoch": 0.8943263099977344,
      "grad_norm": 2.1680841385996215,
      "learning_rate": 4.2024967280060936e-08,
      "loss": 2.4824,
      "step": 6908
    },
    {
      "epoch": 0.8944557724050879,
      "grad_norm": 1.4891514868125184,
      "learning_rate": 4.192319459556775e-08,
      "loss": 2.3562,
      "step": 6909
    },
    {
      "epoch": 0.8945852348124413,
      "grad_norm": 2.4653715109882235,
      "learning_rate": 4.182154101633036e-08,
      "loss": 2.4199,
      "step": 6910
    },
    {
      "epoch": 0.8945852348124413,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964488744735718,
      "eval_runtime": 14.8241,
      "eval_samples_per_second": 2.968,
      "eval_steps_per_second": 0.202,
      "step": 6910
    },
    {
      "epoch": 0.8947146972197948,
      "grad_norm": 2.069641679533272,
      "learning_rate": 4.1720006563113685e-08,
      "loss": 2.4431,
      "step": 6911
    },
    {
      "epoch": 0.8948441596271483,
      "grad_norm": 1.3878531424713867,
      "learning_rate": 4.1618591256658364e-08,
      "loss": 2.3965,
      "step": 6912
    },
    {
      "epoch": 0.8949736220345017,
      "grad_norm": 2.441516934084074,
      "learning_rate": 4.151729511768024e-08,
      "loss": 2.3274,
      "step": 6913
    },
    {
      "epoch": 0.8951030844418552,
      "grad_norm": 1.2444497277156765,
      "learning_rate": 4.141611816687138e-08,
      "loss": 2.1733,
      "step": 6914
    },
    {
      "epoch": 0.8952325468492086,
      "grad_norm": 1.25034331459268,
      "learning_rate": 4.1315060424899006e-08,
      "loss": 2.3225,
      "step": 6915
    },
    {
      "epoch": 0.8952325468492086,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957829713821411,
      "eval_runtime": 13.5391,
      "eval_samples_per_second": 3.25,
      "eval_steps_per_second": 0.222,
      "step": 6915
    },
    {
      "epoch": 0.8953620092565622,
      "grad_norm": 1.4956653083155558,
      "learning_rate": 4.121412191240626e-08,
      "loss": 2.4302,
      "step": 6916
    },
    {
      "epoch": 0.8954914716639156,
      "grad_norm": 3.695456243006164,
      "learning_rate": 4.1113302650012054e-08,
      "loss": 2.373,
      "step": 6917
    },
    {
      "epoch": 0.8956209340712691,
      "grad_norm": 1.5775926229543367,
      "learning_rate": 4.1012602658310476e-08,
      "loss": 2.1853,
      "step": 6918
    },
    {
      "epoch": 0.8957503964786225,
      "grad_norm": 1.755700282843993,
      "learning_rate": 4.0912021957871734e-08,
      "loss": 2.3325,
      "step": 6919
    },
    {
      "epoch": 0.895879858885976,
      "grad_norm": 1.227436237800927,
      "learning_rate": 4.0811560569241485e-08,
      "loss": 2.3989,
      "step": 6920
    },
    {
      "epoch": 0.895879858885976,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4963600635528564,
      "eval_runtime": 14.1878,
      "eval_samples_per_second": 3.101,
      "eval_steps_per_second": 0.211,
      "step": 6920
    },
    {
      "epoch": 0.8960093212933294,
      "grad_norm": 1.688768588163638,
      "learning_rate": 4.071121851294088e-08,
      "loss": 2.3564,
      "step": 6921
    },
    {
      "epoch": 0.8961387837006829,
      "grad_norm": 1.5718793231134163,
      "learning_rate": 4.0610995809466924e-08,
      "loss": 2.3838,
      "step": 6922
    },
    {
      "epoch": 0.8962682461080363,
      "grad_norm": 3.762872273849478,
      "learning_rate": 4.051089247929206e-08,
      "loss": 2.259,
      "step": 6923
    },
    {
      "epoch": 0.8963977085153898,
      "grad_norm": 1.6658417124342157,
      "learning_rate": 4.0410908542864515e-08,
      "loss": 2.3289,
      "step": 6924
    },
    {
      "epoch": 0.8965271709227433,
      "grad_norm": 2.6997947261750572,
      "learning_rate": 4.031104402060795e-08,
      "loss": 2.2056,
      "step": 6925
    },
    {
      "epoch": 0.8965271709227433,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957386255264282,
      "eval_runtime": 15.2393,
      "eval_samples_per_second": 2.887,
      "eval_steps_per_second": 0.197,
      "step": 6925
    },
    {
      "epoch": 0.8966566333300968,
      "grad_norm": 1.6703939407210462,
      "learning_rate": 4.021129893292182e-08,
      "loss": 2.2632,
      "step": 6926
    },
    {
      "epoch": 0.8967860957374503,
      "grad_norm": 2.1294872006307854,
      "learning_rate": 4.0111673300180926e-08,
      "loss": 2.3159,
      "step": 6927
    },
    {
      "epoch": 0.8969155581448037,
      "grad_norm": 1.7415914614987027,
      "learning_rate": 4.0012167142735936e-08,
      "loss": 2.2163,
      "step": 6928
    },
    {
      "epoch": 0.8970450205521572,
      "grad_norm": 1.1887525851054148,
      "learning_rate": 3.991278048091296e-08,
      "loss": 2.1575,
      "step": 6929
    },
    {
      "epoch": 0.8971744829595106,
      "grad_norm": 1.4394557523763398,
      "learning_rate": 3.9813513335013735e-08,
      "loss": 2.1777,
      "step": 6930
    },
    {
      "epoch": 0.8971744829595106,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.500044345855713,
      "eval_runtime": 13.6703,
      "eval_samples_per_second": 3.219,
      "eval_steps_per_second": 0.219,
      "step": 6930
    },
    {
      "epoch": 0.8973039453668641,
      "grad_norm": 1.2593171368973932,
      "learning_rate": 3.971436572531566e-08,
      "loss": 2.2441,
      "step": 6931
    },
    {
      "epoch": 0.8974334077742175,
      "grad_norm": 1.2891194530636572,
      "learning_rate": 3.9615337672071555e-08,
      "loss": 2.296,
      "step": 6932
    },
    {
      "epoch": 0.897562870181571,
      "grad_norm": 1.2017006243417034,
      "learning_rate": 3.95164291955099e-08,
      "loss": 2.3989,
      "step": 6933
    },
    {
      "epoch": 0.8976923325889244,
      "grad_norm": 1.1455584727164558,
      "learning_rate": 3.9417640315834724e-08,
      "loss": 2.2739,
      "step": 6934
    },
    {
      "epoch": 0.897821794996278,
      "grad_norm": 1.1819944504299127,
      "learning_rate": 3.931897105322588e-08,
      "loss": 2.2979,
      "step": 6935
    },
    {
      "epoch": 0.897821794996278,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4935635328292847,
      "eval_runtime": 14.1786,
      "eval_samples_per_second": 3.103,
      "eval_steps_per_second": 0.212,
      "step": 6935
    },
    {
      "epoch": 0.8979512574036315,
      "grad_norm": 2.983504529933777,
      "learning_rate": 3.9220421427838276e-08,
      "loss": 2.3772,
      "step": 6936
    },
    {
      "epoch": 0.8980807198109849,
      "grad_norm": 1.6038312588693155,
      "learning_rate": 3.9121991459803026e-08,
      "loss": 2.3679,
      "step": 6937
    },
    {
      "epoch": 0.8982101822183384,
      "grad_norm": 2.8079810843264,
      "learning_rate": 3.902368116922607e-08,
      "loss": 2.2349,
      "step": 6938
    },
    {
      "epoch": 0.8983396446256918,
      "grad_norm": 1.9090172639941165,
      "learning_rate": 3.89254905761894e-08,
      "loss": 2.4524,
      "step": 6939
    },
    {
      "epoch": 0.8984691070330453,
      "grad_norm": 2.200047488020452,
      "learning_rate": 3.8827419700750574e-08,
      "loss": 2.4019,
      "step": 6940
    },
    {
      "epoch": 0.8984691070330453,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957386255264282,
      "eval_runtime": 14.3267,
      "eval_samples_per_second": 3.071,
      "eval_steps_per_second": 0.209,
      "step": 6940
    },
    {
      "epoch": 0.8985985694403987,
      "grad_norm": 1.5384715793458743,
      "learning_rate": 3.8729468562942325e-08,
      "loss": 2.3403,
      "step": 6941
    },
    {
      "epoch": 0.8987280318477522,
      "grad_norm": 2.1449745506697586,
      "learning_rate": 3.863163718277336e-08,
      "loss": 2.3335,
      "step": 6942
    },
    {
      "epoch": 0.8988574942551056,
      "grad_norm": 4.3114234142983285,
      "learning_rate": 3.853392558022771e-08,
      "loss": 2.5251,
      "step": 6943
    },
    {
      "epoch": 0.8989869566624591,
      "grad_norm": 2.59681552414949,
      "learning_rate": 3.843633377526475e-08,
      "loss": 2.3013,
      "step": 6944
    },
    {
      "epoch": 0.8991164190698127,
      "grad_norm": 1.7149763388676593,
      "learning_rate": 3.833886178781984e-08,
      "loss": 2.4336,
      "step": 6945
    },
    {
      "epoch": 0.8991164190698127,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972034692764282,
      "eval_runtime": 13.9669,
      "eval_samples_per_second": 3.15,
      "eval_steps_per_second": 0.215,
      "step": 6945
    },
    {
      "epoch": 0.8992458814771661,
      "grad_norm": 1.3751559712862687,
      "learning_rate": 3.8241509637803387e-08,
      "loss": 2.4021,
      "step": 6946
    },
    {
      "epoch": 0.8993753438845196,
      "grad_norm": 1.907702766922529,
      "learning_rate": 3.81442773451017e-08,
      "loss": 2.4231,
      "step": 6947
    },
    {
      "epoch": 0.899504806291873,
      "grad_norm": 2.153565100561712,
      "learning_rate": 3.804716492957647e-08,
      "loss": 2.3845,
      "step": 6948
    },
    {
      "epoch": 0.8996342686992265,
      "grad_norm": 1.7111247405700598,
      "learning_rate": 3.7950172411064594e-08,
      "loss": 2.3345,
      "step": 6949
    },
    {
      "epoch": 0.8997637311065799,
      "grad_norm": 1.503382015285647,
      "learning_rate": 3.785329980937912e-08,
      "loss": 2.2749,
      "step": 6950
    },
    {
      "epoch": 0.8997637311065799,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972256422042847,
      "eval_runtime": 14.1392,
      "eval_samples_per_second": 3.112,
      "eval_steps_per_second": 0.212,
      "step": 6950
    },
    {
      "epoch": 0.8998931935139334,
      "grad_norm": 1.6561190580163554,
      "learning_rate": 3.775654714430803e-08,
      "loss": 2.5015,
      "step": 6951
    },
    {
      "epoch": 0.9000226559212868,
      "grad_norm": 1.4122607964748026,
      "learning_rate": 3.765991443561511e-08,
      "loss": 2.4363,
      "step": 6952
    },
    {
      "epoch": 0.9001521183286403,
      "grad_norm": 1.4949728824649122,
      "learning_rate": 3.756340170303942e-08,
      "loss": 2.3779,
      "step": 6953
    },
    {
      "epoch": 0.9002815807359937,
      "grad_norm": 2.0285896518359925,
      "learning_rate": 3.746700896629582e-08,
      "loss": 2.4622,
      "step": 6954
    },
    {
      "epoch": 0.9004110431433473,
      "grad_norm": 1.8488577650294844,
      "learning_rate": 3.73707362450744e-08,
      "loss": 2.4185,
      "step": 6955
    },
    {
      "epoch": 0.9004110431433473,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964488744735718,
      "eval_runtime": 14.9389,
      "eval_samples_per_second": 2.945,
      "eval_steps_per_second": 0.201,
      "step": 6955
    },
    {
      "epoch": 0.9005405055507008,
      "grad_norm": 1.7402328998270247,
      "learning_rate": 3.7274583559040664e-08,
      "loss": 2.4688,
      "step": 6956
    },
    {
      "epoch": 0.9006699679580542,
      "grad_norm": 1.349800529381217,
      "learning_rate": 3.717855092783593e-08,
      "loss": 2.4836,
      "step": 6957
    },
    {
      "epoch": 0.9007994303654077,
      "grad_norm": 1.8948832767582333,
      "learning_rate": 3.7082638371076855e-08,
      "loss": 2.2905,
      "step": 6958
    },
    {
      "epoch": 0.9009288927727611,
      "grad_norm": 3.8180385902227485,
      "learning_rate": 3.698684590835541e-08,
      "loss": 2.5107,
      "step": 6959
    },
    {
      "epoch": 0.9010583551801146,
      "grad_norm": 1.3492656096497837,
      "learning_rate": 3.689117355923921e-08,
      "loss": 2.2882,
      "step": 6960
    },
    {
      "epoch": 0.9010583551801146,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964710474014282,
      "eval_runtime": 14.8863,
      "eval_samples_per_second": 2.956,
      "eval_steps_per_second": 0.202,
      "step": 6960
    },
    {
      "epoch": 0.901187817587468,
      "grad_norm": 1.5259868074002851,
      "learning_rate": 3.679562134327116e-08,
      "loss": 2.2678,
      "step": 6961
    },
    {
      "epoch": 0.9013172799948215,
      "grad_norm": 1.7950690981979358,
      "learning_rate": 3.6700189279969755e-08,
      "loss": 2.3286,
      "step": 6962
    },
    {
      "epoch": 0.9014467424021749,
      "grad_norm": 2.053537008849433,
      "learning_rate": 3.6604877388829104e-08,
      "loss": 2.3889,
      "step": 6963
    },
    {
      "epoch": 0.9015762048095284,
      "grad_norm": 1.3487691471993548,
      "learning_rate": 3.650968568931837e-08,
      "loss": 2.4321,
      "step": 6964
    },
    {
      "epoch": 0.901705667216882,
      "grad_norm": 2.046988425631279,
      "learning_rate": 3.641461420088262e-08,
      "loss": 2.501,
      "step": 6965
    },
    {
      "epoch": 0.901705667216882,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957829713821411,
      "eval_runtime": 14.4573,
      "eval_samples_per_second": 3.043,
      "eval_steps_per_second": 0.208,
      "step": 6965
    },
    {
      "epoch": 0.9018351296242354,
      "grad_norm": 2.0775129527511913,
      "learning_rate": 3.631966294294194e-08,
      "loss": 2.468,
      "step": 6966
    },
    {
      "epoch": 0.9019645920315889,
      "grad_norm": 1.3316849011000063,
      "learning_rate": 3.6224831934892066e-08,
      "loss": 2.261,
      "step": 6967
    },
    {
      "epoch": 0.9020940544389423,
      "grad_norm": 1.1346699448886952,
      "learning_rate": 3.613012119610425e-08,
      "loss": 2.1799,
      "step": 6968
    },
    {
      "epoch": 0.9022235168462958,
      "grad_norm": 1.6321728237788975,
      "learning_rate": 3.603553074592495e-08,
      "loss": 2.4453,
      "step": 6969
    },
    {
      "epoch": 0.9023529792536492,
      "grad_norm": 1.4179594472025552,
      "learning_rate": 3.594106060367635e-08,
      "loss": 2.251,
      "step": 6970
    },
    {
      "epoch": 0.9023529792536492,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964045286178589,
      "eval_runtime": 13.8209,
      "eval_samples_per_second": 3.184,
      "eval_steps_per_second": 0.217,
      "step": 6970
    },
    {
      "epoch": 0.9024824416610027,
      "grad_norm": 1.3813409505729468,
      "learning_rate": 3.584671078865586e-08,
      "loss": 2.2354,
      "step": 6971
    },
    {
      "epoch": 0.9026119040683561,
      "grad_norm": 1.3871296942601798,
      "learning_rate": 3.5752481320136115e-08,
      "loss": 2.4346,
      "step": 6972
    },
    {
      "epoch": 0.9027413664757096,
      "grad_norm": 1.7645533751442184,
      "learning_rate": 3.5658372217365694e-08,
      "loss": 2.304,
      "step": 6973
    },
    {
      "epoch": 0.902870828883063,
      "grad_norm": 1.5786955515060197,
      "learning_rate": 3.556438349956802e-08,
      "loss": 2.4717,
      "step": 6974
    },
    {
      "epoch": 0.9030002912904166,
      "grad_norm": 1.6999238827098182,
      "learning_rate": 3.547051518594247e-08,
      "loss": 2.3564,
      "step": 6975
    },
    {
      "epoch": 0.9030002912904166,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950950145721436,
      "eval_runtime": 14.0199,
      "eval_samples_per_second": 3.138,
      "eval_steps_per_second": 0.214,
      "step": 6975
    },
    {
      "epoch": 0.90312975369777,
      "grad_norm": 2.3180146528633694,
      "learning_rate": 3.5376767295663416e-08,
      "loss": 2.5898,
      "step": 6976
    },
    {
      "epoch": 0.9032592161051235,
      "grad_norm": 1.2525781138706877,
      "learning_rate": 3.528313984788069e-08,
      "loss": 2.4712,
      "step": 6977
    },
    {
      "epoch": 0.903388678512477,
      "grad_norm": 1.5146753675865754,
      "learning_rate": 3.5189632861719686e-08,
      "loss": 2.4648,
      "step": 6978
    },
    {
      "epoch": 0.9035181409198304,
      "grad_norm": 2.022322291432427,
      "learning_rate": 3.509624635628102e-08,
      "loss": 2.3547,
      "step": 6979
    },
    {
      "epoch": 0.9036476033271839,
      "grad_norm": 2.0150281599560267,
      "learning_rate": 3.5002980350640895e-08,
      "loss": 2.2002,
      "step": 6980
    },
    {
      "epoch": 0.9036476033271839,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.495649814605713,
      "eval_runtime": 14.2311,
      "eval_samples_per_second": 3.092,
      "eval_steps_per_second": 0.211,
      "step": 6980
    },
    {
      "epoch": 0.9037770657345373,
      "grad_norm": 1.4981461399585436,
      "learning_rate": 3.4909834863850746e-08,
      "loss": 2.4382,
      "step": 6981
    },
    {
      "epoch": 0.9039065281418908,
      "grad_norm": 1.4588762847333872,
      "learning_rate": 3.4816809914937423e-08,
      "loss": 2.3455,
      "step": 6982
    },
    {
      "epoch": 0.9040359905492442,
      "grad_norm": 1.3739131419699016,
      "learning_rate": 3.47239055229031e-08,
      "loss": 2.4976,
      "step": 6983
    },
    {
      "epoch": 0.9041654529565978,
      "grad_norm": 1.7316968704340205,
      "learning_rate": 3.463112170672543e-08,
      "loss": 2.3793,
      "step": 6984
    },
    {
      "epoch": 0.9042949153639512,
      "grad_norm": 2.545602064433286,
      "learning_rate": 3.453845848535739e-08,
      "loss": 2.3589,
      "step": 6985
    },
    {
      "epoch": 0.9042949153639512,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4943625926971436,
      "eval_runtime": 13.8958,
      "eval_samples_per_second": 3.166,
      "eval_steps_per_second": 0.216,
      "step": 6985
    },
    {
      "epoch": 0.9044243777713047,
      "grad_norm": 1.3981146941723948,
      "learning_rate": 3.4445915877727375e-08,
      "loss": 2.1975,
      "step": 6986
    },
    {
      "epoch": 0.9045538401786581,
      "grad_norm": 1.7952518086041092,
      "learning_rate": 3.435349390273901e-08,
      "loss": 2.5269,
      "step": 6987
    },
    {
      "epoch": 0.9046833025860116,
      "grad_norm": 1.6254409535581296,
      "learning_rate": 3.426119257927123e-08,
      "loss": 2.3235,
      "step": 6988
    },
    {
      "epoch": 0.904812764993365,
      "grad_norm": 1.5820971972696911,
      "learning_rate": 3.41690119261788e-08,
      "loss": 2.4558,
      "step": 6989
    },
    {
      "epoch": 0.9049422274007185,
      "grad_norm": 3.7090017497336105,
      "learning_rate": 3.407695196229112e-08,
      "loss": 2.1719,
      "step": 6990
    },
    {
      "epoch": 0.9049422274007185,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4948952198028564,
      "eval_runtime": 13.8486,
      "eval_samples_per_second": 3.177,
      "eval_steps_per_second": 0.217,
      "step": 6990
    },
    {
      "epoch": 0.905071689808072,
      "grad_norm": 1.8466436865878584,
      "learning_rate": 3.3985012706413553e-08,
      "loss": 2.2668,
      "step": 6991
    },
    {
      "epoch": 0.9052011522154254,
      "grad_norm": 3.529990179316978,
      "learning_rate": 3.389319417732642e-08,
      "loss": 2.509,
      "step": 6992
    },
    {
      "epoch": 0.9053306146227789,
      "grad_norm": 1.642571111915576,
      "learning_rate": 3.380149639378563e-08,
      "loss": 2.5679,
      "step": 6993
    },
    {
      "epoch": 0.9054600770301324,
      "grad_norm": 1.5897813153183968,
      "learning_rate": 3.370991937452224e-08,
      "loss": 2.5044,
      "step": 6994
    },
    {
      "epoch": 0.9055895394374859,
      "grad_norm": 2.1450771090293044,
      "learning_rate": 3.361846313824267e-08,
      "loss": 2.459,
      "step": 6995
    },
    {
      "epoch": 0.9055895394374859,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.496537685394287,
      "eval_runtime": 14.3434,
      "eval_samples_per_second": 3.068,
      "eval_steps_per_second": 0.209,
      "step": 6995
    },
    {
      "epoch": 0.9057190018448393,
      "grad_norm": 1.90263647944974,
      "learning_rate": 3.3527127703628784e-08,
      "loss": 2.4326,
      "step": 6996
    },
    {
      "epoch": 0.9058484642521928,
      "grad_norm": 1.9344058518450322,
      "learning_rate": 3.343591308933761e-08,
      "loss": 2.5249,
      "step": 6997
    },
    {
      "epoch": 0.9059779266595462,
      "grad_norm": 2.1295819523747905,
      "learning_rate": 3.3344819314001794e-08,
      "loss": 2.4021,
      "step": 6998
    },
    {
      "epoch": 0.9061073890668997,
      "grad_norm": 1.4863903839298132,
      "learning_rate": 3.32538463962289e-08,
      "loss": 2.4209,
      "step": 6999
    },
    {
      "epoch": 0.9062368514742531,
      "grad_norm": 1.3075175734582196,
      "learning_rate": 3.316299435460191e-08,
      "loss": 2.3711,
      "step": 7000
    },
    {
      "epoch": 0.9062368514742531,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4944069385528564,
      "eval_runtime": 14.6662,
      "eval_samples_per_second": 3.0,
      "eval_steps_per_second": 0.205,
      "step": 7000
    },
    {
      "epoch": 0.9063663138816066,
      "grad_norm": 2.148921954739415,
      "learning_rate": 3.307226320767944e-08,
      "loss": 2.375,
      "step": 7001
    },
    {
      "epoch": 0.90649577628896,
      "grad_norm": 1.7078373968909446,
      "learning_rate": 3.2981652973994926e-08,
      "loss": 2.4905,
      "step": 7002
    },
    {
      "epoch": 0.9066252386963135,
      "grad_norm": 1.6210454698802603,
      "learning_rate": 3.289116367205747e-08,
      "loss": 2.4141,
      "step": 7003
    },
    {
      "epoch": 0.9067547011036671,
      "grad_norm": 2.3648603944756434,
      "learning_rate": 3.280079532035151e-08,
      "loss": 2.3235,
      "step": 7004
    },
    {
      "epoch": 0.9068841635110205,
      "grad_norm": 2.8279472592446746,
      "learning_rate": 3.2710547937336234e-08,
      "loss": 2.3628,
      "step": 7005
    },
    {
      "epoch": 0.9068841635110205,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4963600635528564,
      "eval_runtime": 14.4715,
      "eval_samples_per_second": 3.04,
      "eval_steps_per_second": 0.207,
      "step": 7005
    },
    {
      "epoch": 0.907013625918374,
      "grad_norm": 2.333271859951235,
      "learning_rate": 3.262042154144669e-08,
      "loss": 2.373,
      "step": 7006
    },
    {
      "epoch": 0.9071430883257274,
      "grad_norm": 1.2470870792561057,
      "learning_rate": 3.253041615109316e-08,
      "loss": 2.2725,
      "step": 7007
    },
    {
      "epoch": 0.9072725507330809,
      "grad_norm": 2.8137119224734572,
      "learning_rate": 3.244053178466086e-08,
      "loss": 2.7734,
      "step": 7008
    },
    {
      "epoch": 0.9074020131404343,
      "grad_norm": 2.06485292690283,
      "learning_rate": 3.235076846051073e-08,
      "loss": 2.2529,
      "step": 7009
    },
    {
      "epoch": 0.9075314755477878,
      "grad_norm": 1.5484055105075882,
      "learning_rate": 3.226112619697845e-08,
      "loss": 2.3179,
      "step": 7010
    },
    {
      "epoch": 0.9075314755477878,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4949618577957153,
      "eval_runtime": 14.0713,
      "eval_samples_per_second": 3.127,
      "eval_steps_per_second": 0.213,
      "step": 7010
    },
    {
      "epoch": 0.9076609379551412,
      "grad_norm": 1.932646542673694,
      "learning_rate": 3.217160501237542e-08,
      "loss": 2.5288,
      "step": 7011
    },
    {
      "epoch": 0.9077904003624947,
      "grad_norm": 1.669576675402848,
      "learning_rate": 3.2082204924988177e-08,
      "loss": 2.1968,
      "step": 7012
    },
    {
      "epoch": 0.9079198627698482,
      "grad_norm": 2.293024978409148,
      "learning_rate": 3.199292595307844e-08,
      "loss": 2.4526,
      "step": 7013
    },
    {
      "epoch": 0.9080493251772017,
      "grad_norm": 2.0486688931711234,
      "learning_rate": 3.190376811488337e-08,
      "loss": 2.3696,
      "step": 7014
    },
    {
      "epoch": 0.9081787875845552,
      "grad_norm": 1.2180422038269936,
      "learning_rate": 3.1814731428615117e-08,
      "loss": 2.3289,
      "step": 7015
    },
    {
      "epoch": 0.9081787875845552,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4963600635528564,
      "eval_runtime": 13.5512,
      "eval_samples_per_second": 3.247,
      "eval_steps_per_second": 0.221,
      "step": 7015
    },
    {
      "epoch": 0.9083082499919086,
      "grad_norm": 2.410134293015543,
      "learning_rate": 3.1725815912461226e-08,
      "loss": 2.3342,
      "step": 7016
    },
    {
      "epoch": 0.9084377123992621,
      "grad_norm": 3.1286056473308768,
      "learning_rate": 3.163702158458466e-08,
      "loss": 2.2793,
      "step": 7017
    },
    {
      "epoch": 0.9085671748066155,
      "grad_norm": 2.039927821971476,
      "learning_rate": 3.154834846312328e-08,
      "loss": 2.5332,
      "step": 7018
    },
    {
      "epoch": 0.908696637213969,
      "grad_norm": 2.7290131744124366,
      "learning_rate": 3.145979656619051e-08,
      "loss": 2.5381,
      "step": 7019
    },
    {
      "epoch": 0.9088260996213224,
      "grad_norm": 2.135700382421176,
      "learning_rate": 3.13713659118748e-08,
      "loss": 2.1831,
      "step": 7020
    },
    {
      "epoch": 0.9088260996213224,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971812963485718,
      "eval_runtime": 15.1723,
      "eval_samples_per_second": 2.9,
      "eval_steps_per_second": 0.198,
      "step": 7020
    },
    {
      "epoch": 0.9089555620286759,
      "grad_norm": 1.3112041253522417,
      "learning_rate": 3.128305651823982e-08,
      "loss": 2.3525,
      "step": 7021
    },
    {
      "epoch": 0.9090850244360293,
      "grad_norm": 1.3389693680271821,
      "learning_rate": 3.119486840332475e-08,
      "loss": 2.4067,
      "step": 7022
    },
    {
      "epoch": 0.9092144868433829,
      "grad_norm": 3.1209015594634004,
      "learning_rate": 3.110680158514359e-08,
      "loss": 2.5396,
      "step": 7023
    },
    {
      "epoch": 0.9093439492507364,
      "grad_norm": 2.1849977543999484,
      "learning_rate": 3.101885608168598e-08,
      "loss": 2.5208,
      "step": 7024
    },
    {
      "epoch": 0.9094734116580898,
      "grad_norm": 1.8100603513424869,
      "learning_rate": 3.093103191091636e-08,
      "loss": 2.5259,
      "step": 7025
    },
    {
      "epoch": 0.9094734116580898,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.496382236480713,
      "eval_runtime": 13.5595,
      "eval_samples_per_second": 3.245,
      "eval_steps_per_second": 0.221,
      "step": 7025
    },
    {
      "epoch": 0.9096028740654433,
      "grad_norm": 1.3299766474054597,
      "learning_rate": 3.084332909077476e-08,
      "loss": 2.3831,
      "step": 7026
    },
    {
      "epoch": 0.9097323364727967,
      "grad_norm": 4.156505197651604,
      "learning_rate": 3.0755747639176213e-08,
      "loss": 2.5967,
      "step": 7027
    },
    {
      "epoch": 0.9098617988801502,
      "grad_norm": 2.201549212444028,
      "learning_rate": 3.0668287574010944e-08,
      "loss": 2.3037,
      "step": 7028
    },
    {
      "epoch": 0.9099912612875036,
      "grad_norm": 4.604382948344746,
      "learning_rate": 3.0580948913144436e-08,
      "loss": 2.6592,
      "step": 7029
    },
    {
      "epoch": 0.9101207236948571,
      "grad_norm": 2.3195634093031443,
      "learning_rate": 3.0493731674417564e-08,
      "loss": 2.4792,
      "step": 7030
    },
    {
      "epoch": 0.9101207236948571,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957829713821411,
      "eval_runtime": 13.7283,
      "eval_samples_per_second": 3.205,
      "eval_steps_per_second": 0.219,
      "step": 7030
    },
    {
      "epoch": 0.9102501861022105,
      "grad_norm": 1.5210388696036747,
      "learning_rate": 3.0406635875646006e-08,
      "loss": 2.2603,
      "step": 7031
    },
    {
      "epoch": 0.910379648509564,
      "grad_norm": 1.3852925817875432,
      "learning_rate": 3.031966153462108e-08,
      "loss": 2.3931,
      "step": 7032
    },
    {
      "epoch": 0.9105091109169176,
      "grad_norm": 3.3035708809813804,
      "learning_rate": 3.023280866910873e-08,
      "loss": 2.4922,
      "step": 7033
    },
    {
      "epoch": 0.910638573324271,
      "grad_norm": 1.5638238748249917,
      "learning_rate": 3.014607729685065e-08,
      "loss": 2.4146,
      "step": 7034
    },
    {
      "epoch": 0.9107680357316245,
      "grad_norm": 2.4001209242479526,
      "learning_rate": 3.0059467435563444e-08,
      "loss": 2.5466,
      "step": 7035
    },
    {
      "epoch": 0.9107680357316245,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498490810394287,
      "eval_runtime": 14.727,
      "eval_samples_per_second": 2.988,
      "eval_steps_per_second": 0.204,
      "step": 7035
    },
    {
      "epoch": 0.9108974981389779,
      "grad_norm": 1.4807362473242356,
      "learning_rate": 2.997297910293885e-08,
      "loss": 2.3066,
      "step": 7036
    },
    {
      "epoch": 0.9110269605463314,
      "grad_norm": 1.8435060157414962,
      "learning_rate": 2.9886612316644e-08,
      "loss": 2.4077,
      "step": 7037
    },
    {
      "epoch": 0.9111564229536848,
      "grad_norm": 1.4764098208405059,
      "learning_rate": 2.980036709432095e-08,
      "loss": 2.3579,
      "step": 7038
    },
    {
      "epoch": 0.9112858853610383,
      "grad_norm": 1.9988550670817895,
      "learning_rate": 2.9714243453587075e-08,
      "loss": 2.4102,
      "step": 7039
    },
    {
      "epoch": 0.9114153477683917,
      "grad_norm": 1.5734442843561907,
      "learning_rate": 2.9628241412034895e-08,
      "loss": 2.384,
      "step": 7040
    },
    {
      "epoch": 0.9114153477683917,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4963157176971436,
      "eval_runtime": 14.3114,
      "eval_samples_per_second": 3.074,
      "eval_steps_per_second": 0.21,
      "step": 7040
    },
    {
      "epoch": 0.9115448101757452,
      "grad_norm": 1.82226920270065,
      "learning_rate": 2.9542360987231954e-08,
      "loss": 2.4617,
      "step": 7041
    },
    {
      "epoch": 0.9116742725830986,
      "grad_norm": 2.3644801169264933,
      "learning_rate": 2.9456602196721237e-08,
      "loss": 2.3323,
      "step": 7042
    },
    {
      "epoch": 0.9118037349904522,
      "grad_norm": 2.415133112080397,
      "learning_rate": 2.9370965058020603e-08,
      "loss": 2.3757,
      "step": 7043
    },
    {
      "epoch": 0.9119331973978057,
      "grad_norm": 1.996313603038745,
      "learning_rate": 2.9285449588623214e-08,
      "loss": 2.4456,
      "step": 7044
    },
    {
      "epoch": 0.9120626598051591,
      "grad_norm": 1.7716411973054633,
      "learning_rate": 2.9200055805997316e-08,
      "loss": 2.3027,
      "step": 7045
    },
    {
      "epoch": 0.9120626598051591,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972256422042847,
      "eval_runtime": 14.3324,
      "eval_samples_per_second": 3.07,
      "eval_steps_per_second": 0.209,
      "step": 7045
    },
    {
      "epoch": 0.9121921222125126,
      "grad_norm": 3.074760297580371,
      "learning_rate": 2.911478372758632e-08,
      "loss": 2.3528,
      "step": 7046
    },
    {
      "epoch": 0.912321584619866,
      "grad_norm": 1.1516943268842077,
      "learning_rate": 2.9029633370808857e-08,
      "loss": 2.22,
      "step": 7047
    },
    {
      "epoch": 0.9124510470272195,
      "grad_norm": 1.3154973925779796,
      "learning_rate": 2.8944604753058518e-08,
      "loss": 2.4102,
      "step": 7048
    },
    {
      "epoch": 0.9125805094345729,
      "grad_norm": 1.6634282719522047,
      "learning_rate": 2.885969789170412e-08,
      "loss": 2.3018,
      "step": 7049
    },
    {
      "epoch": 0.9127099718419264,
      "grad_norm": 1.823655021872299,
      "learning_rate": 2.87749128040897e-08,
      "loss": 2.5159,
      "step": 7050
    },
    {
      "epoch": 0.9127099718419264,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4955832958221436,
      "eval_runtime": 14.4172,
      "eval_samples_per_second": 3.052,
      "eval_steps_per_second": 0.208,
      "step": 7050
    },
    {
      "epoch": 0.9128394342492798,
      "grad_norm": 1.6234313837207206,
      "learning_rate": 2.8690249507534185e-08,
      "loss": 2.4189,
      "step": 7051
    },
    {
      "epoch": 0.9129688966566333,
      "grad_norm": 1.404942548066484,
      "learning_rate": 2.8605708019331864e-08,
      "loss": 2.3135,
      "step": 7052
    },
    {
      "epoch": 0.9130983590639868,
      "grad_norm": 2.0675388446021565,
      "learning_rate": 2.8521288356752123e-08,
      "loss": 2.2988,
      "step": 7053
    },
    {
      "epoch": 0.9132278214713403,
      "grad_norm": 1.640941086563205,
      "learning_rate": 2.8436990537039285e-08,
      "loss": 2.2163,
      "step": 7054
    },
    {
      "epoch": 0.9133572838786937,
      "grad_norm": 1.806276801027841,
      "learning_rate": 2.8352814577412916e-08,
      "loss": 2.2712,
      "step": 7055
    },
    {
      "epoch": 0.9133572838786937,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4970703125,
      "eval_runtime": 14.4297,
      "eval_samples_per_second": 3.049,
      "eval_steps_per_second": 0.208,
      "step": 7055
    },
    {
      "epoch": 0.9134867462860472,
      "grad_norm": 3.606169832754405,
      "learning_rate": 2.826876049506759e-08,
      "loss": 2.4111,
      "step": 7056
    },
    {
      "epoch": 0.9136162086934007,
      "grad_norm": 1.40169341425273,
      "learning_rate": 2.818482830717312e-08,
      "loss": 2.3347,
      "step": 7057
    },
    {
      "epoch": 0.9137456711007541,
      "grad_norm": 1.4103556528574146,
      "learning_rate": 2.8101018030874464e-08,
      "loss": 2.3987,
      "step": 7058
    },
    {
      "epoch": 0.9138751335081076,
      "grad_norm": 2.160089754763084,
      "learning_rate": 2.8017329683291474e-08,
      "loss": 2.406,
      "step": 7059
    },
    {
      "epoch": 0.914004595915461,
      "grad_norm": 3.3646929673503347,
      "learning_rate": 2.7933763281519018e-08,
      "loss": 2.2146,
      "step": 7060
    },
    {
      "epoch": 0.914004595915461,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498579502105713,
      "eval_runtime": 14.1588,
      "eval_samples_per_second": 3.108,
      "eval_steps_per_second": 0.212,
      "step": 7060
    },
    {
      "epoch": 0.9141340583228145,
      "grad_norm": 5.817382460848479,
      "learning_rate": 2.785031884262754e-08,
      "loss": 2.9331,
      "step": 7061
    },
    {
      "epoch": 0.9142635207301679,
      "grad_norm": 2.408236766999161,
      "learning_rate": 2.776699638366201e-08,
      "loss": 2.3374,
      "step": 7062
    },
    {
      "epoch": 0.9143929831375215,
      "grad_norm": 1.555372059637381,
      "learning_rate": 2.7683795921642983e-08,
      "loss": 2.311,
      "step": 7063
    },
    {
      "epoch": 0.9145224455448749,
      "grad_norm": 2.182891212203474,
      "learning_rate": 2.760071747356548e-08,
      "loss": 2.5698,
      "step": 7064
    },
    {
      "epoch": 0.9146519079522284,
      "grad_norm": 3.1583770042103794,
      "learning_rate": 2.751776105640036e-08,
      "loss": 2.2749,
      "step": 7065
    },
    {
      "epoch": 0.9146519079522284,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4928977489471436,
      "eval_runtime": 13.9422,
      "eval_samples_per_second": 3.156,
      "eval_steps_per_second": 0.215,
      "step": 7065
    },
    {
      "epoch": 0.9147813703595818,
      "grad_norm": 2.0722468234920743,
      "learning_rate": 2.7434926687092897e-08,
      "loss": 2.231,
      "step": 7066
    },
    {
      "epoch": 0.9149108327669353,
      "grad_norm": 1.4149597344797429,
      "learning_rate": 2.735221438256365e-08,
      "loss": 2.2798,
      "step": 7067
    },
    {
      "epoch": 0.9150402951742888,
      "grad_norm": 1.9102510305525295,
      "learning_rate": 2.7269624159708547e-08,
      "loss": 2.2405,
      "step": 7068
    },
    {
      "epoch": 0.9151697575816422,
      "grad_norm": 1.7774178608782931,
      "learning_rate": 2.7187156035397984e-08,
      "loss": 2.2954,
      "step": 7069
    },
    {
      "epoch": 0.9152992199889957,
      "grad_norm": 1.3169836219970414,
      "learning_rate": 2.7104810026478064e-08,
      "loss": 2.3015,
      "step": 7070
    },
    {
      "epoch": 0.9152992199889957,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957386255264282,
      "eval_runtime": 13.9478,
      "eval_samples_per_second": 3.155,
      "eval_steps_per_second": 0.215,
      "step": 7070
    },
    {
      "epoch": 0.9154286823963491,
      "grad_norm": 1.185547105513437,
      "learning_rate": 2.7022586149769438e-08,
      "loss": 2.3291,
      "step": 7071
    },
    {
      "epoch": 0.9155581448037027,
      "grad_norm": 1.480910400180959,
      "learning_rate": 2.6940484422067968e-08,
      "loss": 2.125,
      "step": 7072
    },
    {
      "epoch": 0.9156876072110561,
      "grad_norm": 1.8566020898293176,
      "learning_rate": 2.6858504860144756e-08,
      "loss": 2.3665,
      "step": 7073
    },
    {
      "epoch": 0.9158170696184096,
      "grad_norm": 1.2540132200776988,
      "learning_rate": 2.6776647480745572e-08,
      "loss": 2.4546,
      "step": 7074
    },
    {
      "epoch": 0.915946532025763,
      "grad_norm": 1.907325050926739,
      "learning_rate": 2.6694912300591625e-08,
      "loss": 2.3972,
      "step": 7075
    },
    {
      "epoch": 0.915946532025763,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972034692764282,
      "eval_runtime": 14.6107,
      "eval_samples_per_second": 3.011,
      "eval_steps_per_second": 0.205,
      "step": 7075
    },
    {
      "epoch": 0.9160759944331165,
      "grad_norm": 1.5592317114407166,
      "learning_rate": 2.661329933637914e-08,
      "loss": 2.4392,
      "step": 7076
    },
    {
      "epoch": 0.91620545684047,
      "grad_norm": 2.364951514084217,
      "learning_rate": 2.6531808604778746e-08,
      "loss": 2.5674,
      "step": 7077
    },
    {
      "epoch": 0.9163349192478234,
      "grad_norm": 1.3831854310806055,
      "learning_rate": 2.6450440122436915e-08,
      "loss": 2.4541,
      "step": 7078
    },
    {
      "epoch": 0.9164643816551769,
      "grad_norm": 2.234933845138531,
      "learning_rate": 2.6369193905974795e-08,
      "loss": 2.53,
      "step": 7079
    },
    {
      "epoch": 0.9165938440625303,
      "grad_norm": 2.1766923499300077,
      "learning_rate": 2.628806997198842e-08,
      "loss": 2.4595,
      "step": 7080
    },
    {
      "epoch": 0.9165938440625303,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964488744735718,
      "eval_runtime": 14.5093,
      "eval_samples_per_second": 3.033,
      "eval_steps_per_second": 0.207,
      "step": 7080
    },
    {
      "epoch": 0.9167233064698838,
      "grad_norm": 1.7919842517700322,
      "learning_rate": 2.6207068337049183e-08,
      "loss": 2.4656,
      "step": 7081
    },
    {
      "epoch": 0.9168527688772373,
      "grad_norm": 1.4709112462430858,
      "learning_rate": 2.6126189017703223e-08,
      "loss": 2.5549,
      "step": 7082
    },
    {
      "epoch": 0.9169822312845908,
      "grad_norm": 1.7100051678600812,
      "learning_rate": 2.6045432030471704e-08,
      "loss": 2.4683,
      "step": 7083
    },
    {
      "epoch": 0.9171116936919442,
      "grad_norm": 1.9072161131673198,
      "learning_rate": 2.596479739185101e-08,
      "loss": 2.5508,
      "step": 7084
    },
    {
      "epoch": 0.9172411560992977,
      "grad_norm": 1.4712893627163914,
      "learning_rate": 2.5884285118312272e-08,
      "loss": 2.2708,
      "step": 7085
    },
    {
      "epoch": 0.9172411560992977,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.496537685394287,
      "eval_runtime": 13.7652,
      "eval_samples_per_second": 3.196,
      "eval_steps_per_second": 0.218,
      "step": 7085
    },
    {
      "epoch": 0.9173706185066511,
      "grad_norm": 1.2951727690283728,
      "learning_rate": 2.580389522630192e-08,
      "loss": 2.1978,
      "step": 7086
    },
    {
      "epoch": 0.9175000809140046,
      "grad_norm": 1.7036553354717583,
      "learning_rate": 2.5723627732241116e-08,
      "loss": 2.1532,
      "step": 7087
    },
    {
      "epoch": 0.917629543321358,
      "grad_norm": 2.929387473199386,
      "learning_rate": 2.564348265252599e-08,
      "loss": 2.5149,
      "step": 7088
    },
    {
      "epoch": 0.9177590057287115,
      "grad_norm": 3.2133821287494486,
      "learning_rate": 2.5563460003528024e-08,
      "loss": 2.5774,
      "step": 7089
    },
    {
      "epoch": 0.917888468136065,
      "grad_norm": 1.7447901240849042,
      "learning_rate": 2.548355980159338e-08,
      "loss": 2.2886,
      "step": 7090
    },
    {
      "epoch": 0.917888468136065,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.497025966644287,
      "eval_runtime": 13.2259,
      "eval_samples_per_second": 3.327,
      "eval_steps_per_second": 0.227,
      "step": 7090
    },
    {
      "epoch": 0.9180179305434184,
      "grad_norm": 1.2883557299948176,
      "learning_rate": 2.540378206304331e-08,
      "loss": 2.4417,
      "step": 7091
    },
    {
      "epoch": 0.918147392950772,
      "grad_norm": 1.41430104318878,
      "learning_rate": 2.532412680417394e-08,
      "loss": 2.3547,
      "step": 7092
    },
    {
      "epoch": 0.9182768553581254,
      "grad_norm": 3.4193553456719155,
      "learning_rate": 2.5244594041256634e-08,
      "loss": 2.7429,
      "step": 7093
    },
    {
      "epoch": 0.9184063177654789,
      "grad_norm": 1.7762983460110748,
      "learning_rate": 2.5165183790537424e-08,
      "loss": 2.2505,
      "step": 7094
    },
    {
      "epoch": 0.9185357801728323,
      "grad_norm": 2.414024801420144,
      "learning_rate": 2.5085896068237497e-08,
      "loss": 2.0686,
      "step": 7095
    },
    {
      "epoch": 0.9185357801728323,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979580640792847,
      "eval_runtime": 14.2755,
      "eval_samples_per_second": 3.082,
      "eval_steps_per_second": 0.21,
      "step": 7095
    },
    {
      "epoch": 0.9186652425801858,
      "grad_norm": 1.8523981053200522,
      "learning_rate": 2.500673089055314e-08,
      "loss": 2.4197,
      "step": 7096
    },
    {
      "epoch": 0.9187947049875392,
      "grad_norm": 1.7528143058258632,
      "learning_rate": 2.492768827365516e-08,
      "loss": 2.4106,
      "step": 7097
    },
    {
      "epoch": 0.9189241673948927,
      "grad_norm": 1.285329611329612,
      "learning_rate": 2.4848768233689814e-08,
      "loss": 2.3181,
      "step": 7098
    },
    {
      "epoch": 0.9190536298022461,
      "grad_norm": 1.6182332333630947,
      "learning_rate": 2.476997078677809e-08,
      "loss": 2.5557,
      "step": 7099
    },
    {
      "epoch": 0.9191830922095996,
      "grad_norm": 4.127569905044869,
      "learning_rate": 2.469129594901594e-08,
      "loss": 2.7598,
      "step": 7100
    },
    {
      "epoch": 0.9191830922095996,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4943625926971436,
      "eval_runtime": 14.1971,
      "eval_samples_per_second": 3.099,
      "eval_steps_per_second": 0.211,
      "step": 7100
    },
    {
      "epoch": 0.919312554616953,
      "grad_norm": 1.5197704226048916,
      "learning_rate": 2.461274373647418e-08,
      "loss": 2.2683,
      "step": 7101
    },
    {
      "epoch": 0.9194420170243066,
      "grad_norm": 2.1193861160618144,
      "learning_rate": 2.453431416519894e-08,
      "loss": 2.2866,
      "step": 7102
    },
    {
      "epoch": 0.9195714794316601,
      "grad_norm": 1.6487703089372663,
      "learning_rate": 2.4456007251210874e-08,
      "loss": 2.4358,
      "step": 7103
    },
    {
      "epoch": 0.9197009418390135,
      "grad_norm": 3.579570221910329,
      "learning_rate": 2.437782301050587e-08,
      "loss": 2.958,
      "step": 7104
    },
    {
      "epoch": 0.919830404246367,
      "grad_norm": 1.4044541701000823,
      "learning_rate": 2.429976145905448e-08,
      "loss": 2.2527,
      "step": 7105
    },
    {
      "epoch": 0.919830404246367,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5006436109542847,
      "eval_runtime": 13.5183,
      "eval_samples_per_second": 3.255,
      "eval_steps_per_second": 0.222,
      "step": 7105
    },
    {
      "epoch": 0.9199598666537204,
      "grad_norm": 1.4728967037105603,
      "learning_rate": 2.4221822612802426e-08,
      "loss": 2.3408,
      "step": 7106
    },
    {
      "epoch": 0.9200893290610739,
      "grad_norm": 2.601718491161629,
      "learning_rate": 2.4144006487670444e-08,
      "loss": 2.354,
      "step": 7107
    },
    {
      "epoch": 0.9202187914684273,
      "grad_norm": 1.317018352397815,
      "learning_rate": 2.4066313099553803e-08,
      "loss": 2.1799,
      "step": 7108
    },
    {
      "epoch": 0.9203482538757808,
      "grad_norm": 3.0035819449399943,
      "learning_rate": 2.3988742464323213e-08,
      "loss": 2.2373,
      "step": 7109
    },
    {
      "epoch": 0.9204777162831342,
      "grad_norm": 1.360090693885692,
      "learning_rate": 2.3911294597823846e-08,
      "loss": 2.248,
      "step": 7110
    },
    {
      "epoch": 0.9204777162831342,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964267015457153,
      "eval_runtime": 13.4428,
      "eval_samples_per_second": 3.273,
      "eval_steps_per_second": 0.223,
      "step": 7110
    },
    {
      "epoch": 0.9206071786904878,
      "grad_norm": 1.7027887949776572,
      "learning_rate": 2.3833969515876033e-08,
      "loss": 2.3711,
      "step": 7111
    },
    {
      "epoch": 0.9207366410978413,
      "grad_norm": 1.6691317853854986,
      "learning_rate": 2.3756767234275126e-08,
      "loss": 2.2795,
      "step": 7112
    },
    {
      "epoch": 0.9208661035051947,
      "grad_norm": 1.7811320438580154,
      "learning_rate": 2.367968776879101e-08,
      "loss": 2.3435,
      "step": 7113
    },
    {
      "epoch": 0.9209955659125482,
      "grad_norm": 1.273949027832693,
      "learning_rate": 2.3602731135169005e-08,
      "loss": 2.4016,
      "step": 7114
    },
    {
      "epoch": 0.9211250283199016,
      "grad_norm": 2.266894265190741,
      "learning_rate": 2.3525897349128967e-08,
      "loss": 2.2341,
      "step": 7115
    },
    {
      "epoch": 0.9211250283199016,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971591234207153,
      "eval_runtime": 14.5366,
      "eval_samples_per_second": 3.027,
      "eval_steps_per_second": 0.206,
      "step": 7115
    },
    {
      "epoch": 0.9212544907272551,
      "grad_norm": 1.7154309899146465,
      "learning_rate": 2.3449186426365565e-08,
      "loss": 2.4268,
      "step": 7116
    },
    {
      "epoch": 0.9213839531346085,
      "grad_norm": 1.6654087658822747,
      "learning_rate": 2.3372598382548836e-08,
      "loss": 2.395,
      "step": 7117
    },
    {
      "epoch": 0.921513415541962,
      "grad_norm": 2.227577211958741,
      "learning_rate": 2.3296133233323277e-08,
      "loss": 2.2214,
      "step": 7118
    },
    {
      "epoch": 0.9216428779493154,
      "grad_norm": 1.3163057430377807,
      "learning_rate": 2.3219790994308476e-08,
      "loss": 2.2974,
      "step": 7119
    },
    {
      "epoch": 0.9217723403566689,
      "grad_norm": 1.394949476582911,
      "learning_rate": 2.3143571681098903e-08,
      "loss": 2.3654,
      "step": 7120
    },
    {
      "epoch": 0.9217723403566689,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.496537685394287,
      "eval_runtime": 14.3593,
      "eval_samples_per_second": 3.064,
      "eval_steps_per_second": 0.209,
      "step": 7120
    },
    {
      "epoch": 0.9219018027640224,
      "grad_norm": 1.2743034752277569,
      "learning_rate": 2.3067475309263912e-08,
      "loss": 2.3721,
      "step": 7121
    },
    {
      "epoch": 0.9220312651713759,
      "grad_norm": 2.0881291028861493,
      "learning_rate": 2.2991501894347663e-08,
      "loss": 2.3496,
      "step": 7122
    },
    {
      "epoch": 0.9221607275787294,
      "grad_norm": 2.608108249096377,
      "learning_rate": 2.2915651451869346e-08,
      "loss": 2.4844,
      "step": 7123
    },
    {
      "epoch": 0.9222901899860828,
      "grad_norm": 1.9081646854416212,
      "learning_rate": 2.2839923997322807e-08,
      "loss": 2.5474,
      "step": 7124
    },
    {
      "epoch": 0.9224196523934363,
      "grad_norm": 1.8809318730490425,
      "learning_rate": 2.2764319546177138e-08,
      "loss": 2.1777,
      "step": 7125
    },
    {
      "epoch": 0.9224196523934363,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4958717823028564,
      "eval_runtime": 14.205,
      "eval_samples_per_second": 3.098,
      "eval_steps_per_second": 0.211,
      "step": 7125
    },
    {
      "epoch": 0.9225491148007897,
      "grad_norm": 2.0446362618843144,
      "learning_rate": 2.2688838113876023e-08,
      "loss": 2.3235,
      "step": 7126
    },
    {
      "epoch": 0.9226785772081432,
      "grad_norm": 2.261957019239308,
      "learning_rate": 2.2613479715837896e-08,
      "loss": 2.3501,
      "step": 7127
    },
    {
      "epoch": 0.9228080396154966,
      "grad_norm": 2.3239661893106187,
      "learning_rate": 2.2538244367456344e-08,
      "loss": 2.2274,
      "step": 7128
    },
    {
      "epoch": 0.9229375020228501,
      "grad_norm": 2.2076522671478918,
      "learning_rate": 2.2463132084099767e-08,
      "loss": 2.3462,
      "step": 7129
    },
    {
      "epoch": 0.9230669644302035,
      "grad_norm": 1.1950741339495472,
      "learning_rate": 2.238814288111138e-08,
      "loss": 2.2517,
      "step": 7130
    },
    {
      "epoch": 0.9230669644302035,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4949618577957153,
      "eval_runtime": 13.7184,
      "eval_samples_per_second": 3.207,
      "eval_steps_per_second": 0.219,
      "step": 7130
    },
    {
      "epoch": 0.9231964268375571,
      "grad_norm": 2.181796098598965,
      "learning_rate": 2.2313276773809068e-08,
      "loss": 2.4417,
      "step": 7131
    },
    {
      "epoch": 0.9233258892449105,
      "grad_norm": 1.2237474663465835,
      "learning_rate": 2.223853377748601e-08,
      "loss": 2.1438,
      "step": 7132
    },
    {
      "epoch": 0.923455351652264,
      "grad_norm": 1.399037671914145,
      "learning_rate": 2.2163913907409794e-08,
      "loss": 2.3567,
      "step": 7133
    },
    {
      "epoch": 0.9235848140596175,
      "grad_norm": 2.051132177522068,
      "learning_rate": 2.2089417178823012e-08,
      "loss": 2.4241,
      "step": 7134
    },
    {
      "epoch": 0.9237142764669709,
      "grad_norm": 1.364399343498906,
      "learning_rate": 2.2015043606943357e-08,
      "loss": 2.3882,
      "step": 7135
    },
    {
      "epoch": 0.9237142764669709,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4956276416778564,
      "eval_runtime": 14.4444,
      "eval_samples_per_second": 3.046,
      "eval_steps_per_second": 0.208,
      "step": 7135
    },
    {
      "epoch": 0.9238437388743244,
      "grad_norm": 1.5977479507105041,
      "learning_rate": 2.1940793206962907e-08,
      "loss": 2.3955,
      "step": 7136
    },
    {
      "epoch": 0.9239732012816778,
      "grad_norm": 1.4644852585763786,
      "learning_rate": 2.1866665994048908e-08,
      "loss": 2.3579,
      "step": 7137
    },
    {
      "epoch": 0.9241026636890313,
      "grad_norm": 1.2077184194484607,
      "learning_rate": 2.1792661983343415e-08,
      "loss": 2.4602,
      "step": 7138
    },
    {
      "epoch": 0.9242321260963847,
      "grad_norm": 1.4336907303876831,
      "learning_rate": 2.171878118996301e-08,
      "loss": 2.2944,
      "step": 7139
    },
    {
      "epoch": 0.9243615885037382,
      "grad_norm": 1.3119707726064556,
      "learning_rate": 2.164502362899966e-08,
      "loss": 2.2656,
      "step": 7140
    },
    {
      "epoch": 0.9243615885037382,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.494340419769287,
      "eval_runtime": 14.1087,
      "eval_samples_per_second": 3.119,
      "eval_steps_per_second": 0.213,
      "step": 7140
    },
    {
      "epoch": 0.9244910509110917,
      "grad_norm": 3.622724612047138,
      "learning_rate": 2.1571389315519568e-08,
      "loss": 2.5801,
      "step": 7141
    },
    {
      "epoch": 0.9246205133184452,
      "grad_norm": 1.3790066548674327,
      "learning_rate": 2.1497878264564176e-08,
      "loss": 2.1812,
      "step": 7142
    },
    {
      "epoch": 0.9247499757257986,
      "grad_norm": 1.9689126146388305,
      "learning_rate": 2.14244904911496e-08,
      "loss": 2.3926,
      "step": 7143
    },
    {
      "epoch": 0.9248794381331521,
      "grad_norm": 2.0265864140218204,
      "learning_rate": 2.1351226010266764e-08,
      "loss": 2.2722,
      "step": 7144
    },
    {
      "epoch": 0.9250089005405056,
      "grad_norm": 1.4781929181225895,
      "learning_rate": 2.12780848368814e-08,
      "loss": 2.3054,
      "step": 7145
    },
    {
      "epoch": 0.9250089005405056,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4958274364471436,
      "eval_runtime": 14.4997,
      "eval_samples_per_second": 3.035,
      "eval_steps_per_second": 0.207,
      "step": 7145
    },
    {
      "epoch": 0.925138362947859,
      "grad_norm": 1.8876353247831301,
      "learning_rate": 2.1205066985934067e-08,
      "loss": 2.4717,
      "step": 7146
    },
    {
      "epoch": 0.9252678253552125,
      "grad_norm": 2.0383318174500276,
      "learning_rate": 2.1132172472340186e-08,
      "loss": 2.4189,
      "step": 7147
    },
    {
      "epoch": 0.9253972877625659,
      "grad_norm": 1.609758376737458,
      "learning_rate": 2.1059401310990005e-08,
      "loss": 2.5513,
      "step": 7148
    },
    {
      "epoch": 0.9255267501699194,
      "grad_norm": 1.508110485008864,
      "learning_rate": 2.0986753516748297e-08,
      "loss": 2.3855,
      "step": 7149
    },
    {
      "epoch": 0.9256562125772728,
      "grad_norm": 1.5317401254860392,
      "learning_rate": 2.0914229104455062e-08,
      "loss": 2.467,
      "step": 7150
    },
    {
      "epoch": 0.9256562125772728,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.3675,
      "eval_samples_per_second": 3.062,
      "eval_steps_per_second": 0.209,
      "step": 7150
    },
    {
      "epoch": 0.9257856749846264,
      "grad_norm": 1.41752057630314,
      "learning_rate": 2.0841828088924778e-08,
      "loss": 2.1428,
      "step": 7151
    },
    {
      "epoch": 0.9259151373919798,
      "grad_norm": 1.6510502843066648,
      "learning_rate": 2.076955048494672e-08,
      "loss": 2.3738,
      "step": 7152
    },
    {
      "epoch": 0.9260445997993333,
      "grad_norm": 3.0699131274935683,
      "learning_rate": 2.069739630728533e-08,
      "loss": 2.5312,
      "step": 7153
    },
    {
      "epoch": 0.9261740622066867,
      "grad_norm": 1.3023268686076754,
      "learning_rate": 2.0625365570679374e-08,
      "loss": 2.2679,
      "step": 7154
    },
    {
      "epoch": 0.9263035246140402,
      "grad_norm": 1.6198894719857853,
      "learning_rate": 2.0553458289842564e-08,
      "loss": 2.5352,
      "step": 7155
    },
    {
      "epoch": 0.9263035246140402,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4963600635528564,
      "eval_runtime": 13.6428,
      "eval_samples_per_second": 3.225,
      "eval_steps_per_second": 0.22,
      "step": 7155
    },
    {
      "epoch": 0.9264329870213937,
      "grad_norm": 2.175590198320846,
      "learning_rate": 2.0481674479463502e-08,
      "loss": 2.6143,
      "step": 7156
    },
    {
      "epoch": 0.9265624494287471,
      "grad_norm": 1.2319904170801284,
      "learning_rate": 2.041001415420539e-08,
      "loss": 2.3259,
      "step": 7157
    },
    {
      "epoch": 0.9266919118361006,
      "grad_norm": 2.3549156002087552,
      "learning_rate": 2.0338477328706448e-08,
      "loss": 2.3787,
      "step": 7158
    },
    {
      "epoch": 0.926821374243454,
      "grad_norm": 2.7343430493376424,
      "learning_rate": 2.0267064017579358e-08,
      "loss": 2.4712,
      "step": 7159
    },
    {
      "epoch": 0.9269508366508076,
      "grad_norm": 1.0225272335920785,
      "learning_rate": 2.0195774235411898e-08,
      "loss": 2.1694,
      "step": 7160
    },
    {
      "epoch": 0.9269508366508076,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964710474014282,
      "eval_runtime": 15.1303,
      "eval_samples_per_second": 2.908,
      "eval_steps_per_second": 0.198,
      "step": 7160
    },
    {
      "epoch": 0.927080299058161,
      "grad_norm": 1.3246823160714005,
      "learning_rate": 2.0124607996766383e-08,
      "loss": 2.1389,
      "step": 7161
    },
    {
      "epoch": 0.9272097614655145,
      "grad_norm": 1.49051384336523,
      "learning_rate": 2.0053565316179925e-08,
      "loss": 2.4202,
      "step": 7162
    },
    {
      "epoch": 0.9273392238728679,
      "grad_norm": 1.5190813593250359,
      "learning_rate": 1.9982646208164465e-08,
      "loss": 2.1631,
      "step": 7163
    },
    {
      "epoch": 0.9274686862802214,
      "grad_norm": 1.3290962613547905,
      "learning_rate": 1.9911850687206607e-08,
      "loss": 2.356,
      "step": 7164
    },
    {
      "epoch": 0.9275981486875748,
      "grad_norm": 1.587747139532819,
      "learning_rate": 1.9841178767767908e-08,
      "loss": 2.1841,
      "step": 7165
    },
    {
      "epoch": 0.9275981486875748,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957829713821411,
      "eval_runtime": 13.8622,
      "eval_samples_per_second": 3.174,
      "eval_steps_per_second": 0.216,
      "step": 7165
    },
    {
      "epoch": 0.9277276110949283,
      "grad_norm": 1.237932352070521,
      "learning_rate": 1.9770630464284457e-08,
      "loss": 2.3687,
      "step": 7166
    },
    {
      "epoch": 0.9278570735022817,
      "grad_norm": 1.284300661957639,
      "learning_rate": 1.970020579116709e-08,
      "loss": 2.1904,
      "step": 7167
    },
    {
      "epoch": 0.9279865359096352,
      "grad_norm": 1.4032019285160064,
      "learning_rate": 1.9629904762801592e-08,
      "loss": 2.2686,
      "step": 7168
    },
    {
      "epoch": 0.9281159983169887,
      "grad_norm": 1.7381926298643124,
      "learning_rate": 1.955972739354828e-08,
      "loss": 2.3208,
      "step": 7169
    },
    {
      "epoch": 0.9282454607243422,
      "grad_norm": 1.4258728108457215,
      "learning_rate": 1.948967369774242e-08,
      "loss": 2.4304,
      "step": 7170
    },
    {
      "epoch": 0.9282454607243422,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4942737817764282,
      "eval_runtime": 14.1274,
      "eval_samples_per_second": 3.115,
      "eval_steps_per_second": 0.212,
      "step": 7170
    },
    {
      "epoch": 0.9283749231316957,
      "grad_norm": 1.6775506543354486,
      "learning_rate": 1.941974368969396e-08,
      "loss": 2.3286,
      "step": 7171
    },
    {
      "epoch": 0.9285043855390491,
      "grad_norm": 1.3836017005343881,
      "learning_rate": 1.9349937383687235e-08,
      "loss": 2.2793,
      "step": 7172
    },
    {
      "epoch": 0.9286338479464026,
      "grad_norm": 1.9620486970858353,
      "learning_rate": 1.9280254793981808e-08,
      "loss": 2.4409,
      "step": 7173
    },
    {
      "epoch": 0.928763310353756,
      "grad_norm": 1.2811812307679653,
      "learning_rate": 1.9210695934811783e-08,
      "loss": 2.3909,
      "step": 7174
    },
    {
      "epoch": 0.9288927727611095,
      "grad_norm": 1.8706862856075894,
      "learning_rate": 1.9141260820385796e-08,
      "loss": 2.4541,
      "step": 7175
    },
    {
      "epoch": 0.9288927727611095,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979802370071411,
      "eval_runtime": 13.6642,
      "eval_samples_per_second": 3.22,
      "eval_steps_per_second": 0.22,
      "step": 7175
    },
    {
      "epoch": 0.9290222351684629,
      "grad_norm": 1.5540646760332986,
      "learning_rate": 1.9071949464887635e-08,
      "loss": 2.2078,
      "step": 7176
    },
    {
      "epoch": 0.9291516975758164,
      "grad_norm": 1.4997618609996564,
      "learning_rate": 1.9002761882475282e-08,
      "loss": 2.3901,
      "step": 7177
    },
    {
      "epoch": 0.9292811599831698,
      "grad_norm": 2.9361136239944003,
      "learning_rate": 1.8933698087281808e-08,
      "loss": 2.5645,
      "step": 7178
    },
    {
      "epoch": 0.9294106223905233,
      "grad_norm": 2.965376560170296,
      "learning_rate": 1.8864758093414952e-08,
      "loss": 2.415,
      "step": 7179
    },
    {
      "epoch": 0.9295400847978769,
      "grad_norm": 2.276619059412394,
      "learning_rate": 1.8795941914957e-08,
      "loss": 2.7188,
      "step": 7180
    },
    {
      "epoch": 0.9295400847978769,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4943625926971436,
      "eval_runtime": 14.2249,
      "eval_samples_per_second": 3.093,
      "eval_steps_per_second": 0.211,
      "step": 7180
    },
    {
      "epoch": 0.9296695472052303,
      "grad_norm": 2.9092526217954946,
      "learning_rate": 1.8727249565965167e-08,
      "loss": 2.4021,
      "step": 7181
    },
    {
      "epoch": 0.9297990096125838,
      "grad_norm": 1.3953492632765847,
      "learning_rate": 1.8658681060471223e-08,
      "loss": 2.3716,
      "step": 7182
    },
    {
      "epoch": 0.9299284720199372,
      "grad_norm": 1.6801007707681777,
      "learning_rate": 1.8590236412481602e-08,
      "loss": 2.4609,
      "step": 7183
    },
    {
      "epoch": 0.9300579344272907,
      "grad_norm": 1.424212932057574,
      "learning_rate": 1.852191563597762e-08,
      "loss": 2.5208,
      "step": 7184
    },
    {
      "epoch": 0.9301873968346441,
      "grad_norm": 3.194029079307449,
      "learning_rate": 1.8453718744914986e-08,
      "loss": 2.1797,
      "step": 7185
    },
    {
      "epoch": 0.9301873968346441,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 13.5812,
      "eval_samples_per_second": 3.24,
      "eval_steps_per_second": 0.221,
      "step": 7185
    },
    {
      "epoch": 0.9303168592419976,
      "grad_norm": 1.2539054468962592,
      "learning_rate": 1.8385645753224574e-08,
      "loss": 2.2627,
      "step": 7186
    },
    {
      "epoch": 0.930446321649351,
      "grad_norm": 1.5295649609234558,
      "learning_rate": 1.8317696674811507e-08,
      "loss": 2.3096,
      "step": 7187
    },
    {
      "epoch": 0.9305757840567045,
      "grad_norm": 1.711268094693845,
      "learning_rate": 1.824987152355573e-08,
      "loss": 2.4338,
      "step": 7188
    },
    {
      "epoch": 0.930705246464058,
      "grad_norm": 1.4788403065817934,
      "learning_rate": 1.8182170313311986e-08,
      "loss": 2.3828,
      "step": 7189
    },
    {
      "epoch": 0.9308347088714115,
      "grad_norm": 2.6700433395845264,
      "learning_rate": 1.8114593057909564e-08,
      "loss": 2.5293,
      "step": 7190
    },
    {
      "epoch": 0.9308347088714115,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4943625926971436,
      "eval_runtime": 14.4503,
      "eval_samples_per_second": 3.045,
      "eval_steps_per_second": 0.208,
      "step": 7190
    },
    {
      "epoch": 0.930964171278765,
      "grad_norm": 1.783834952534302,
      "learning_rate": 1.8047139771152563e-08,
      "loss": 2.157,
      "step": 7191
    },
    {
      "epoch": 0.9310936336861184,
      "grad_norm": 1.5364898841001369,
      "learning_rate": 1.7979810466819543e-08,
      "loss": 2.5132,
      "step": 7192
    },
    {
      "epoch": 0.9312230960934719,
      "grad_norm": 1.3024390693013825,
      "learning_rate": 1.791260515866408e-08,
      "loss": 2.1809,
      "step": 7193
    },
    {
      "epoch": 0.9313525585008253,
      "grad_norm": 1.6658879528427881,
      "learning_rate": 1.784552386041409e-08,
      "loss": 2.5586,
      "step": 7194
    },
    {
      "epoch": 0.9314820209081788,
      "grad_norm": 1.5238794819936137,
      "learning_rate": 1.777856658577229e-08,
      "loss": 2.4333,
      "step": 7195
    },
    {
      "epoch": 0.9314820209081788,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4963157176971436,
      "eval_runtime": 13.2958,
      "eval_samples_per_second": 3.309,
      "eval_steps_per_second": 0.226,
      "step": 7195
    },
    {
      "epoch": 0.9316114833155322,
      "grad_norm": 2.3791558189713586,
      "learning_rate": 1.771173334841607e-08,
      "loss": 2.5488,
      "step": 7196
    },
    {
      "epoch": 0.9317409457228857,
      "grad_norm": 1.1915978799064315,
      "learning_rate": 1.7645024161997494e-08,
      "loss": 2.3013,
      "step": 7197
    },
    {
      "epoch": 0.9318704081302391,
      "grad_norm": 1.367895020671955,
      "learning_rate": 1.757843904014324e-08,
      "loss": 2.2861,
      "step": 7198
    },
    {
      "epoch": 0.9319998705375927,
      "grad_norm": 1.9638535761932208,
      "learning_rate": 1.7511977996454847e-08,
      "loss": 2.4932,
      "step": 7199
    },
    {
      "epoch": 0.9321293329449462,
      "grad_norm": 1.8446572728837705,
      "learning_rate": 1.7445641044507992e-08,
      "loss": 2.2277,
      "step": 7200
    },
    {
      "epoch": 0.9321293329449462,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978693723678589,
      "eval_runtime": 15.1278,
      "eval_samples_per_second": 2.909,
      "eval_steps_per_second": 0.198,
      "step": 7200
    },
    {
      "epoch": 0.9322587953522996,
      "grad_norm": 1.712432752932684,
      "learning_rate": 1.73794281978535e-08,
      "loss": 2.3877,
      "step": 7201
    },
    {
      "epoch": 0.9323882577596531,
      "grad_norm": 1.097312537550144,
      "learning_rate": 1.7313339470016802e-08,
      "loss": 2.2461,
      "step": 7202
    },
    {
      "epoch": 0.9325177201670065,
      "grad_norm": 1.6428425024104267,
      "learning_rate": 1.724737487449772e-08,
      "loss": 2.3218,
      "step": 7203
    },
    {
      "epoch": 0.93264718257436,
      "grad_norm": 2.334053070514605,
      "learning_rate": 1.7181534424770967e-08,
      "loss": 2.4387,
      "step": 7204
    },
    {
      "epoch": 0.9327766449817134,
      "grad_norm": 1.5783238761177116,
      "learning_rate": 1.7115818134285712e-08,
      "loss": 2.3899,
      "step": 7205
    },
    {
      "epoch": 0.9327766449817134,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4956942796707153,
      "eval_runtime": 14.0207,
      "eval_samples_per_second": 3.138,
      "eval_steps_per_second": 0.214,
      "step": 7205
    },
    {
      "epoch": 0.9329061073890669,
      "grad_norm": 1.1986220101030398,
      "learning_rate": 1.7050226016465804e-08,
      "loss": 2.2395,
      "step": 7206
    },
    {
      "epoch": 0.9330355697964203,
      "grad_norm": 1.364158051465634,
      "learning_rate": 1.6984758084709828e-08,
      "loss": 2.4285,
      "step": 7207
    },
    {
      "epoch": 0.9331650322037738,
      "grad_norm": 1.837306138124314,
      "learning_rate": 1.6919414352390902e-08,
      "loss": 2.5713,
      "step": 7208
    },
    {
      "epoch": 0.9332944946111273,
      "grad_norm": 1.6239673627216982,
      "learning_rate": 1.68541948328569e-08,
      "loss": 2.2537,
      "step": 7209
    },
    {
      "epoch": 0.9334239570184808,
      "grad_norm": 2.3157277060049224,
      "learning_rate": 1.6789099539430144e-08,
      "loss": 2.3594,
      "step": 7210
    },
    {
      "epoch": 0.9334239570184808,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964267015457153,
      "eval_runtime": 14.169,
      "eval_samples_per_second": 3.105,
      "eval_steps_per_second": 0.212,
      "step": 7210
    },
    {
      "epoch": 0.9335534194258343,
      "grad_norm": 3.4770952128168773,
      "learning_rate": 1.672412848540757e-08,
      "loss": 2.6511,
      "step": 7211
    },
    {
      "epoch": 0.9336828818331877,
      "grad_norm": 2.965505816135355,
      "learning_rate": 1.6659281684061056e-08,
      "loss": 2.9141,
      "step": 7212
    },
    {
      "epoch": 0.9338123442405412,
      "grad_norm": 1.2312269413743147,
      "learning_rate": 1.6594559148636682e-08,
      "loss": 2.3218,
      "step": 7213
    },
    {
      "epoch": 0.9339418066478946,
      "grad_norm": 1.5874257597825383,
      "learning_rate": 1.6529960892355464e-08,
      "loss": 2.3645,
      "step": 7214
    },
    {
      "epoch": 0.9340712690552481,
      "grad_norm": 2.4886744739888518,
      "learning_rate": 1.646548692841289e-08,
      "loss": 2.6704,
      "step": 7215
    },
    {
      "epoch": 0.9340712690552481,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.497847080230713,
      "eval_runtime": 15.2907,
      "eval_samples_per_second": 2.878,
      "eval_steps_per_second": 0.196,
      "step": 7215
    },
    {
      "epoch": 0.9342007314626015,
      "grad_norm": 2.2356545294067045,
      "learning_rate": 1.640113726997898e-08,
      "loss": 2.4028,
      "step": 7216
    },
    {
      "epoch": 0.934330193869955,
      "grad_norm": 4.151902055402119,
      "learning_rate": 1.6336911930198567e-08,
      "loss": 2.77,
      "step": 7217
    },
    {
      "epoch": 0.9344596562773084,
      "grad_norm": 2.0940204635500215,
      "learning_rate": 1.627281092219088e-08,
      "loss": 2.3772,
      "step": 7218
    },
    {
      "epoch": 0.934589118684662,
      "grad_norm": 1.406309760622827,
      "learning_rate": 1.6208834259049886e-08,
      "loss": 2.2754,
      "step": 7219
    },
    {
      "epoch": 0.9347185810920154,
      "grad_norm": 1.614074002335195,
      "learning_rate": 1.614498195384423e-08,
      "loss": 2.3096,
      "step": 7220
    },
    {
      "epoch": 0.9347185810920154,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979580640792847,
      "eval_runtime": 14.0668,
      "eval_samples_per_second": 3.128,
      "eval_steps_per_second": 0.213,
      "step": 7220
    },
    {
      "epoch": 0.9348480434993689,
      "grad_norm": 1.4736495612145082,
      "learning_rate": 1.608125401961688e-08,
      "loss": 2.2563,
      "step": 7221
    },
    {
      "epoch": 0.9349775059067224,
      "grad_norm": 1.800849767404705,
      "learning_rate": 1.6017650469385692e-08,
      "loss": 2.3459,
      "step": 7222
    },
    {
      "epoch": 0.9351069683140758,
      "grad_norm": 2.04111783289126,
      "learning_rate": 1.595417131614277e-08,
      "loss": 2.4375,
      "step": 7223
    },
    {
      "epoch": 0.9352364307214293,
      "grad_norm": 1.9879893843451641,
      "learning_rate": 1.589081657285524e-08,
      "loss": 2.4331,
      "step": 7224
    },
    {
      "epoch": 0.9353658931287827,
      "grad_norm": 1.304496121774615,
      "learning_rate": 1.5827586252464485e-08,
      "loss": 2.2385,
      "step": 7225
    },
    {
      "epoch": 0.9353658931287827,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957607984542847,
      "eval_runtime": 14.5501,
      "eval_samples_per_second": 3.024,
      "eval_steps_per_second": 0.206,
      "step": 7225
    },
    {
      "epoch": 0.9354953555361362,
      "grad_norm": 1.1447855321736853,
      "learning_rate": 1.5764480367886637e-08,
      "loss": 2.2419,
      "step": 7226
    },
    {
      "epoch": 0.9356248179434896,
      "grad_norm": 2.0701384285508837,
      "learning_rate": 1.5701498932012212e-08,
      "loss": 2.4126,
      "step": 7227
    },
    {
      "epoch": 0.9357542803508431,
      "grad_norm": 2.295687697472843,
      "learning_rate": 1.563864195770669e-08,
      "loss": 2.3516,
      "step": 7228
    },
    {
      "epoch": 0.9358837427581966,
      "grad_norm": 2.061494305746054,
      "learning_rate": 1.5575909457809583e-08,
      "loss": 2.313,
      "step": 7229
    },
    {
      "epoch": 0.9360132051655501,
      "grad_norm": 1.6929309802550512,
      "learning_rate": 1.5513301445135575e-08,
      "loss": 2.4275,
      "step": 7230
    },
    {
      "epoch": 0.9360132051655501,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979137182235718,
      "eval_runtime": 14.6369,
      "eval_samples_per_second": 3.006,
      "eval_steps_per_second": 0.205,
      "step": 7230
    },
    {
      "epoch": 0.9361426675729035,
      "grad_norm": 1.4897287210815997,
      "learning_rate": 1.5450817932473397e-08,
      "loss": 2.4094,
      "step": 7231
    },
    {
      "epoch": 0.936272129980257,
      "grad_norm": 1.360789228555177,
      "learning_rate": 1.5388458932586655e-08,
      "loss": 2.4407,
      "step": 7232
    },
    {
      "epoch": 0.9364015923876104,
      "grad_norm": 2.040963972659097,
      "learning_rate": 1.5326224458213497e-08,
      "loss": 2.4114,
      "step": 7233
    },
    {
      "epoch": 0.9365310547949639,
      "grad_norm": 4.527746841199033,
      "learning_rate": 1.5264114522066458e-08,
      "loss": 2.4233,
      "step": 7234
    },
    {
      "epoch": 0.9366605172023174,
      "grad_norm": 2.500805781736201,
      "learning_rate": 1.520212913683289e-08,
      "loss": 2.3965,
      "step": 7235
    },
    {
      "epoch": 0.9366605172023174,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964045286178589,
      "eval_runtime": 15.1093,
      "eval_samples_per_second": 2.912,
      "eval_steps_per_second": 0.199,
      "step": 7235
    },
    {
      "epoch": 0.9367899796096708,
      "grad_norm": 1.8022209870815982,
      "learning_rate": 1.51402683151744e-08,
      "loss": 2.3621,
      "step": 7236
    },
    {
      "epoch": 0.9369194420170243,
      "grad_norm": 3.2612554162539626,
      "learning_rate": 1.507853206972741e-08,
      "loss": 3.0359,
      "step": 7237
    },
    {
      "epoch": 0.9370489044243778,
      "grad_norm": 2.4494270523123496,
      "learning_rate": 1.5016920413102874e-08,
      "loss": 2.5046,
      "step": 7238
    },
    {
      "epoch": 0.9371783668317313,
      "grad_norm": 1.2814545802159494,
      "learning_rate": 1.4955433357886005e-08,
      "loss": 2.363,
      "step": 7239
    },
    {
      "epoch": 0.9373078292390847,
      "grad_norm": 2.354240739812939,
      "learning_rate": 1.4894070916637026e-08,
      "loss": 2.6096,
      "step": 7240
    },
    {
      "epoch": 0.9373078292390847,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964488744735718,
      "eval_runtime": 13.9991,
      "eval_samples_per_second": 3.143,
      "eval_steps_per_second": 0.214,
      "step": 7240
    },
    {
      "epoch": 0.9374372916464382,
      "grad_norm": 1.3097829508295689,
      "learning_rate": 1.4832833101890292e-08,
      "loss": 2.2222,
      "step": 7241
    },
    {
      "epoch": 0.9375667540537916,
      "grad_norm": 3.6501386050267715,
      "learning_rate": 1.4771719926154893e-08,
      "loss": 2.1611,
      "step": 7242
    },
    {
      "epoch": 0.9376962164611451,
      "grad_norm": 1.9359320725338194,
      "learning_rate": 1.4710731401914591e-08,
      "loss": 2.4983,
      "step": 7243
    },
    {
      "epoch": 0.9378256788684985,
      "grad_norm": 1.8111453069317403,
      "learning_rate": 1.4649867541627268e-08,
      "loss": 2.4988,
      "step": 7244
    },
    {
      "epoch": 0.937955141275852,
      "grad_norm": 2.5191671569097362,
      "learning_rate": 1.458912835772576e-08,
      "loss": 2.7856,
      "step": 7245
    },
    {
      "epoch": 0.937955141275852,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.494384765625,
      "eval_runtime": 13.1039,
      "eval_samples_per_second": 3.358,
      "eval_steps_per_second": 0.229,
      "step": 7245
    },
    {
      "epoch": 0.9380846036832055,
      "grad_norm": 1.8268612944466591,
      "learning_rate": 1.452851386261729e-08,
      "loss": 2.4724,
      "step": 7246
    },
    {
      "epoch": 0.9382140660905589,
      "grad_norm": 1.4734167866753105,
      "learning_rate": 1.4468024068683487e-08,
      "loss": 2.3552,
      "step": 7247
    },
    {
      "epoch": 0.9383435284979125,
      "grad_norm": 2.675842808448819,
      "learning_rate": 1.4407658988280711e-08,
      "loss": 2.4436,
      "step": 7248
    },
    {
      "epoch": 0.9384729909052659,
      "grad_norm": 1.6734459260739822,
      "learning_rate": 1.43474186337398e-08,
      "loss": 2.342,
      "step": 7249
    },
    {
      "epoch": 0.9386024533126194,
      "grad_norm": 1.589257514410813,
      "learning_rate": 1.4287303017365902e-08,
      "loss": 2.3904,
      "step": 7250
    },
    {
      "epoch": 0.9386024533126194,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.6724,
      "eval_samples_per_second": 2.999,
      "eval_steps_per_second": 0.204,
      "step": 7250
    },
    {
      "epoch": 0.9387319157199728,
      "grad_norm": 1.8545437818859143,
      "learning_rate": 1.4227312151439063e-08,
      "loss": 2.1809,
      "step": 7251
    },
    {
      "epoch": 0.9388613781273263,
      "grad_norm": 3.86337122984894,
      "learning_rate": 1.4167446048213365e-08,
      "loss": 2.4714,
      "step": 7252
    },
    {
      "epoch": 0.9389908405346797,
      "grad_norm": 2.0860696595856747,
      "learning_rate": 1.4107704719917986e-08,
      "loss": 2.3516,
      "step": 7253
    },
    {
      "epoch": 0.9391203029420332,
      "grad_norm": 1.1038209886822736,
      "learning_rate": 1.404808817875608e-08,
      "loss": 2.2783,
      "step": 7254
    },
    {
      "epoch": 0.9392497653493866,
      "grad_norm": 2.4029609772859084,
      "learning_rate": 1.3988596436905478e-08,
      "loss": 2.4556,
      "step": 7255
    },
    {
      "epoch": 0.9392497653493866,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4943181276321411,
      "eval_runtime": 15.7328,
      "eval_samples_per_second": 2.797,
      "eval_steps_per_second": 0.191,
      "step": 7255
    },
    {
      "epoch": 0.9393792277567401,
      "grad_norm": 1.538137985762532,
      "learning_rate": 1.3929229506518816e-08,
      "loss": 2.2795,
      "step": 7256
    },
    {
      "epoch": 0.9395086901640936,
      "grad_norm": 2.3369787079151845,
      "learning_rate": 1.3869987399722854e-08,
      "loss": 2.4534,
      "step": 7257
    },
    {
      "epoch": 0.9396381525714471,
      "grad_norm": 1.4600383518490043,
      "learning_rate": 1.3810870128619022e-08,
      "loss": 2.4375,
      "step": 7258
    },
    {
      "epoch": 0.9397676149788006,
      "grad_norm": 2.1002123536717323,
      "learning_rate": 1.3751877705283079e-08,
      "loss": 2.4001,
      "step": 7259
    },
    {
      "epoch": 0.939897077386154,
      "grad_norm": 1.5048255488105096,
      "learning_rate": 1.3693010141765731e-08,
      "loss": 2.2427,
      "step": 7260
    },
    {
      "epoch": 0.939897077386154,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964710474014282,
      "eval_runtime": 14.0835,
      "eval_samples_per_second": 3.124,
      "eval_steps_per_second": 0.213,
      "step": 7260
    },
    {
      "epoch": 0.9400265397935075,
      "grad_norm": 1.8315493893809824,
      "learning_rate": 1.3634267450091595e-08,
      "loss": 2.3176,
      "step": 7261
    },
    {
      "epoch": 0.9401560022008609,
      "grad_norm": 1.2151798681304828,
      "learning_rate": 1.3575649642260103e-08,
      "loss": 2.3186,
      "step": 7262
    },
    {
      "epoch": 0.9402854646082144,
      "grad_norm": 1.2592064255683664,
      "learning_rate": 1.3517156730245285e-08,
      "loss": 2.219,
      "step": 7263
    },
    {
      "epoch": 0.9404149270155678,
      "grad_norm": 1.8039821055950596,
      "learning_rate": 1.3458788725995293e-08,
      "loss": 2.3335,
      "step": 7264
    },
    {
      "epoch": 0.9405443894229213,
      "grad_norm": 1.8846649694006072,
      "learning_rate": 1.3400545641433157e-08,
      "loss": 2.499,
      "step": 7265
    },
    {
      "epoch": 0.9405443894229213,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4951171875,
      "eval_runtime": 14.1169,
      "eval_samples_per_second": 3.117,
      "eval_steps_per_second": 0.213,
      "step": 7265
    },
    {
      "epoch": 0.9406738518302747,
      "grad_norm": 1.6877894185279019,
      "learning_rate": 1.3342427488456169e-08,
      "loss": 2.3862,
      "step": 7266
    },
    {
      "epoch": 0.9408033142376282,
      "grad_norm": 1.7303874208958079,
      "learning_rate": 1.3284434278936012e-08,
      "loss": 2.3171,
      "step": 7267
    },
    {
      "epoch": 0.9409327766449818,
      "grad_norm": 1.7176868020854008,
      "learning_rate": 1.3226566024719042e-08,
      "loss": 2.6223,
      "step": 7268
    },
    {
      "epoch": 0.9410622390523352,
      "grad_norm": 1.5720220417288304,
      "learning_rate": 1.3168822737626084e-08,
      "loss": 2.5862,
      "step": 7269
    },
    {
      "epoch": 0.9411917014596887,
      "grad_norm": 2.1503685584638084,
      "learning_rate": 1.3111204429452351e-08,
      "loss": 2.4375,
      "step": 7270
    },
    {
      "epoch": 0.9411917014596887,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971591234207153,
      "eval_runtime": 14.4288,
      "eval_samples_per_second": 3.049,
      "eval_steps_per_second": 0.208,
      "step": 7270
    },
    {
      "epoch": 0.9413211638670421,
      "grad_norm": 1.4382514598366116,
      "learning_rate": 1.3053711111967599e-08,
      "loss": 2.3596,
      "step": 7271
    },
    {
      "epoch": 0.9414506262743956,
      "grad_norm": 1.69665867333801,
      "learning_rate": 1.2996342796915833e-08,
      "loss": 2.2056,
      "step": 7272
    },
    {
      "epoch": 0.941580088681749,
      "grad_norm": 3.7559503487136996,
      "learning_rate": 1.2939099496015808e-08,
      "loss": 2.6938,
      "step": 7273
    },
    {
      "epoch": 0.9417095510891025,
      "grad_norm": 1.6909101400781492,
      "learning_rate": 1.2881981220960669e-08,
      "loss": 2.3712,
      "step": 7274
    },
    {
      "epoch": 0.9418390134964559,
      "grad_norm": 2.056743240464644,
      "learning_rate": 1.2824987983417886e-08,
      "loss": 2.2402,
      "step": 7275
    },
    {
      "epoch": 0.9418390134964559,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957164525985718,
      "eval_runtime": 14.2215,
      "eval_samples_per_second": 3.094,
      "eval_steps_per_second": 0.211,
      "step": 7275
    },
    {
      "epoch": 0.9419684759038094,
      "grad_norm": 1.4914811378444897,
      "learning_rate": 1.2768119795029537e-08,
      "loss": 2.3743,
      "step": 7276
    },
    {
      "epoch": 0.9420979383111628,
      "grad_norm": 3.4031284574570906,
      "learning_rate": 1.2711376667412164e-08,
      "loss": 2.2612,
      "step": 7277
    },
    {
      "epoch": 0.9422274007185164,
      "grad_norm": 1.1020569585461424,
      "learning_rate": 1.2654758612156495e-08,
      "loss": 2.197,
      "step": 7278
    },
    {
      "epoch": 0.9423568631258699,
      "grad_norm": 1.281548718578471,
      "learning_rate": 1.2598265640828209e-08,
      "loss": 2.3096,
      "step": 7279
    },
    {
      "epoch": 0.9424863255332233,
      "grad_norm": 1.243489721733227,
      "learning_rate": 1.2541897764966823e-08,
      "loss": 2.1604,
      "step": 7280
    },
    {
      "epoch": 0.9424863255332233,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957607984542847,
      "eval_runtime": 13.5518,
      "eval_samples_per_second": 3.247,
      "eval_steps_per_second": 0.221,
      "step": 7280
    },
    {
      "epoch": 0.9426157879405768,
      "grad_norm": 1.7305596841579918,
      "learning_rate": 1.2485654996086879e-08,
      "loss": 2.4523,
      "step": 7281
    },
    {
      "epoch": 0.9427452503479302,
      "grad_norm": 1.588185853979172,
      "learning_rate": 1.2429537345677033e-08,
      "loss": 2.3833,
      "step": 7282
    },
    {
      "epoch": 0.9428747127552837,
      "grad_norm": 1.978804753073798,
      "learning_rate": 1.237354482520034e-08,
      "loss": 2.5054,
      "step": 7283
    },
    {
      "epoch": 0.9430041751626371,
      "grad_norm": 2.0911675288005442,
      "learning_rate": 1.2317677446094591e-08,
      "loss": 2.2534,
      "step": 7284
    },
    {
      "epoch": 0.9431336375699906,
      "grad_norm": 2.5694186339547396,
      "learning_rate": 1.2261935219771633e-08,
      "loss": 2.408,
      "step": 7285
    },
    {
      "epoch": 0.9431336375699906,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.5,
      "eval_runtime": 13.8672,
      "eval_samples_per_second": 3.173,
      "eval_steps_per_second": 0.216,
      "step": 7285
    },
    {
      "epoch": 0.943263099977344,
      "grad_norm": 2.0447963312070216,
      "learning_rate": 1.2206318157618052e-08,
      "loss": 2.6216,
      "step": 7286
    },
    {
      "epoch": 0.9433925623846976,
      "grad_norm": 1.4553847299666178,
      "learning_rate": 1.215082627099476e-08,
      "loss": 2.429,
      "step": 7287
    },
    {
      "epoch": 0.943522024792051,
      "grad_norm": 2.2504704257219097,
      "learning_rate": 1.2095459571237131e-08,
      "loss": 2.5015,
      "step": 7288
    },
    {
      "epoch": 0.9436514871994045,
      "grad_norm": 1.6413761662322537,
      "learning_rate": 1.2040218069654872e-08,
      "loss": 2.376,
      "step": 7289
    },
    {
      "epoch": 0.943780949606758,
      "grad_norm": 2.3961109207541793,
      "learning_rate": 1.1985101777532149e-08,
      "loss": 2.4802,
      "step": 7290
    },
    {
      "epoch": 0.943780949606758,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986683130264282,
      "eval_runtime": 13.8256,
      "eval_samples_per_second": 3.182,
      "eval_steps_per_second": 0.217,
      "step": 7290
    },
    {
      "epoch": 0.9439104120141114,
      "grad_norm": 1.7954022605824076,
      "learning_rate": 1.1930110706127665e-08,
      "loss": 2.4012,
      "step": 7291
    },
    {
      "epoch": 0.9440398744214649,
      "grad_norm": 1.4679110421983972,
      "learning_rate": 1.1875244866674449e-08,
      "loss": 2.1479,
      "step": 7292
    },
    {
      "epoch": 0.9441693368288183,
      "grad_norm": 3.243872932005203,
      "learning_rate": 1.1820504270379924e-08,
      "loss": 2.6655,
      "step": 7293
    },
    {
      "epoch": 0.9442987992361718,
      "grad_norm": 1.2887108874804751,
      "learning_rate": 1.1765888928426047e-08,
      "loss": 2.2993,
      "step": 7294
    },
    {
      "epoch": 0.9444282616435252,
      "grad_norm": 1.4228039966431218,
      "learning_rate": 1.1711398851968967e-08,
      "loss": 2.408,
      "step": 7295
    },
    {
      "epoch": 0.9444282616435252,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.49560546875,
      "eval_runtime": 14.1426,
      "eval_samples_per_second": 3.111,
      "eval_steps_per_second": 0.212,
      "step": 7295
    },
    {
      "epoch": 0.9445577240508787,
      "grad_norm": 3.88450674771092,
      "learning_rate": 1.165703405213943e-08,
      "loss": 2.7812,
      "step": 7296
    },
    {
      "epoch": 0.9446871864582322,
      "grad_norm": 1.8927213303239423,
      "learning_rate": 1.1602794540042651e-08,
      "loss": 2.4941,
      "step": 7297
    },
    {
      "epoch": 0.9448166488655857,
      "grad_norm": 2.575082662827145,
      "learning_rate": 1.154868032675803e-08,
      "loss": 2.3613,
      "step": 7298
    },
    {
      "epoch": 0.9449461112729391,
      "grad_norm": 1.434058251434029,
      "learning_rate": 1.1494691423339643e-08,
      "loss": 2.1841,
      "step": 7299
    },
    {
      "epoch": 0.9450755736802926,
      "grad_norm": 1.3001621989805072,
      "learning_rate": 1.1440827840815683e-08,
      "loss": 2.4939,
      "step": 7300
    },
    {
      "epoch": 0.9450755736802926,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4958717823028564,
      "eval_runtime": 14.5442,
      "eval_samples_per_second": 3.025,
      "eval_steps_per_second": 0.206,
      "step": 7300
    },
    {
      "epoch": 0.9452050360876461,
      "grad_norm": 1.5466346549261907,
      "learning_rate": 1.1387089590188874e-08,
      "loss": 2.1616,
      "step": 7301
    },
    {
      "epoch": 0.9453344984949995,
      "grad_norm": 2.395263966581637,
      "learning_rate": 1.1333476682436409e-08,
      "loss": 2.2578,
      "step": 7302
    },
    {
      "epoch": 0.945463960902353,
      "grad_norm": 2.0596771618054115,
      "learning_rate": 1.1279989128509735e-08,
      "loss": 2.4712,
      "step": 7303
    },
    {
      "epoch": 0.9455934233097064,
      "grad_norm": 2.1018028387666368,
      "learning_rate": 1.1226626939334905e-08,
      "loss": 2.3896,
      "step": 7304
    },
    {
      "epoch": 0.9457228857170599,
      "grad_norm": 1.6881665606409486,
      "learning_rate": 1.1173390125812158e-08,
      "loss": 2.6094,
      "step": 7305
    },
    {
      "epoch": 0.9457228857170599,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972478151321411,
      "eval_runtime": 13.475,
      "eval_samples_per_second": 3.265,
      "eval_steps_per_second": 0.223,
      "step": 7305
    },
    {
      "epoch": 0.9458523481244133,
      "grad_norm": 2.6122881772824984,
      "learning_rate": 1.1120278698816059e-08,
      "loss": 2.2791,
      "step": 7306
    },
    {
      "epoch": 0.9459818105317669,
      "grad_norm": 2.1562276031343464,
      "learning_rate": 1.1067292669195915e-08,
      "loss": 2.3938,
      "step": 7307
    },
    {
      "epoch": 0.9461112729391203,
      "grad_norm": 3.5752633437378947,
      "learning_rate": 1.1014432047775011e-08,
      "loss": 2.1299,
      "step": 7308
    },
    {
      "epoch": 0.9462407353464738,
      "grad_norm": 2.330924484528379,
      "learning_rate": 1.096169684535131e-08,
      "loss": 2.5522,
      "step": 7309
    },
    {
      "epoch": 0.9463701977538272,
      "grad_norm": 2.180375180429506,
      "learning_rate": 1.0909087072697097e-08,
      "loss": 2.373,
      "step": 7310
    },
    {
      "epoch": 0.9463701977538272,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4942959547042847,
      "eval_runtime": 14.1808,
      "eval_samples_per_second": 3.103,
      "eval_steps_per_second": 0.212,
      "step": 7310
    },
    {
      "epoch": 0.9464996601611807,
      "grad_norm": 1.6505233346470747,
      "learning_rate": 1.0856602740558775e-08,
      "loss": 2.5283,
      "step": 7311
    },
    {
      "epoch": 0.9466291225685342,
      "grad_norm": 2.1844685099038528,
      "learning_rate": 1.080424385965756e-08,
      "loss": 2.5142,
      "step": 7312
    },
    {
      "epoch": 0.9467585849758876,
      "grad_norm": 1.2951390275557058,
      "learning_rate": 1.0752010440688646e-08,
      "loss": 2.1941,
      "step": 7313
    },
    {
      "epoch": 0.9468880473832411,
      "grad_norm": 3.2822655345435665,
      "learning_rate": 1.0699902494321832e-08,
      "loss": 2.3247,
      "step": 7314
    },
    {
      "epoch": 0.9470175097905945,
      "grad_norm": 2.133097875897563,
      "learning_rate": 1.0647920031201312e-08,
      "loss": 2.3745,
      "step": 7315
    },
    {
      "epoch": 0.9470175097905945,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950505495071411,
      "eval_runtime": 14.1203,
      "eval_samples_per_second": 3.116,
      "eval_steps_per_second": 0.212,
      "step": 7315
    },
    {
      "epoch": 0.947146972197948,
      "grad_norm": 1.3479860455786605,
      "learning_rate": 1.0596063061945466e-08,
      "loss": 2.22,
      "step": 7316
    },
    {
      "epoch": 0.9472764346053015,
      "grad_norm": 1.941777100398384,
      "learning_rate": 1.0544331597147072e-08,
      "loss": 2.4612,
      "step": 7317
    },
    {
      "epoch": 0.947405897012655,
      "grad_norm": 3.6319490911739414,
      "learning_rate": 1.0492725647373441e-08,
      "loss": 2.0918,
      "step": 7318
    },
    {
      "epoch": 0.9475353594200084,
      "grad_norm": 1.4410784602179658,
      "learning_rate": 1.0441245223166071e-08,
      "loss": 2.3333,
      "step": 7319
    },
    {
      "epoch": 0.9476648218273619,
      "grad_norm": 2.173902107438115,
      "learning_rate": 1.0389890335040925e-08,
      "loss": 2.3936,
      "step": 7320
    },
    {
      "epoch": 0.9476648218273619,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.492919921875,
      "eval_runtime": 16.1278,
      "eval_samples_per_second": 2.728,
      "eval_steps_per_second": 0.186,
      "step": 7320
    },
    {
      "epoch": 0.9477942842347153,
      "grad_norm": 1.4147388185005652,
      "learning_rate": 1.0338660993488292e-08,
      "loss": 2.3953,
      "step": 7321
    },
    {
      "epoch": 0.9479237466420688,
      "grad_norm": 1.3625984598317105,
      "learning_rate": 1.0287557208972648e-08,
      "loss": 2.3193,
      "step": 7322
    },
    {
      "epoch": 0.9480532090494223,
      "grad_norm": 1.5637963379552904,
      "learning_rate": 1.0236578991933213e-08,
      "loss": 2.3391,
      "step": 7323
    },
    {
      "epoch": 0.9481826714567757,
      "grad_norm": 1.3188596313646903,
      "learning_rate": 1.0185726352783116e-08,
      "loss": 2.3853,
      "step": 7324
    },
    {
      "epoch": 0.9483121338641292,
      "grad_norm": 1.4629474978314922,
      "learning_rate": 1.0134999301910226e-08,
      "loss": 2.4214,
      "step": 7325
    },
    {
      "epoch": 0.9483121338641292,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4936301708221436,
      "eval_runtime": 13.9688,
      "eval_samples_per_second": 3.15,
      "eval_steps_per_second": 0.215,
      "step": 7325
    },
    {
      "epoch": 0.9484415962714827,
      "grad_norm": 2.832314534420589,
      "learning_rate": 1.0084397849676395e-08,
      "loss": 2.5669,
      "step": 7326
    },
    {
      "epoch": 0.9485710586788362,
      "grad_norm": 2.606187797682997,
      "learning_rate": 1.0033922006418145e-08,
      "loss": 2.2905,
      "step": 7327
    },
    {
      "epoch": 0.9487005210861896,
      "grad_norm": 0.9888701230945862,
      "learning_rate": 9.98357178244612e-09,
      "loss": 2.2888,
      "step": 7328
    },
    {
      "epoch": 0.9488299834935431,
      "grad_norm": 1.667527783015043,
      "learning_rate": 9.933347188045283e-09,
      "loss": 2.4023,
      "step": 7329
    },
    {
      "epoch": 0.9489594459008965,
      "grad_norm": 2.3367484663827747,
      "learning_rate": 9.88324823347521e-09,
      "loss": 2.1078,
      "step": 7330
    },
    {
      "epoch": 0.9489594459008965,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950950145721436,
      "eval_runtime": 15.054,
      "eval_samples_per_second": 2.923,
      "eval_steps_per_second": 0.199,
      "step": 7330
    },
    {
      "epoch": 0.94908890830825,
      "grad_norm": 2.2762944094936435,
      "learning_rate": 9.833274928969446e-09,
      "loss": 2.5332,
      "step": 7331
    },
    {
      "epoch": 0.9492183707156034,
      "grad_norm": 1.309417144735979,
      "learning_rate": 9.78342728473622e-09,
      "loss": 2.3677,
      "step": 7332
    },
    {
      "epoch": 0.9493478331229569,
      "grad_norm": 1.4407181980524342,
      "learning_rate": 9.733705310957797e-09,
      "loss": 2.3867,
      "step": 7333
    },
    {
      "epoch": 0.9494772955303104,
      "grad_norm": 1.4792144822822209,
      "learning_rate": 9.684109017790988e-09,
      "loss": 2.2162,
      "step": 7334
    },
    {
      "epoch": 0.9496067579376638,
      "grad_norm": 1.5951997593903797,
      "learning_rate": 9.634638415366785e-09,
      "loss": 2.2603,
      "step": 7335
    },
    {
      "epoch": 0.9496067579376638,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4956721067428589,
      "eval_runtime": 20.932,
      "eval_samples_per_second": 2.102,
      "eval_steps_per_second": 0.143,
      "step": 7335
    },
    {
      "epoch": 0.9497362203450174,
      "grad_norm": 1.843351952072748,
      "learning_rate": 9.585293513790506e-09,
      "loss": 2.3887,
      "step": 7336
    },
    {
      "epoch": 0.9498656827523708,
      "grad_norm": 2.7029986207250163,
      "learning_rate": 9.536074323141937e-09,
      "loss": 2.3555,
      "step": 7337
    },
    {
      "epoch": 0.9499951451597243,
      "grad_norm": 1.4480094228327978,
      "learning_rate": 9.486980853475186e-09,
      "loss": 2.3281,
      "step": 7338
    },
    {
      "epoch": 0.9501246075670777,
      "grad_norm": 1.6471152645029068,
      "learning_rate": 9.438013114818346e-09,
      "loss": 2.3877,
      "step": 7339
    },
    {
      "epoch": 0.9502540699744312,
      "grad_norm": 1.5877721985286477,
      "learning_rate": 9.389171117174245e-09,
      "loss": 2.4888,
      "step": 7340
    },
    {
      "epoch": 0.9502540699744312,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957386255264282,
      "eval_runtime": 14.6441,
      "eval_samples_per_second": 3.005,
      "eval_steps_per_second": 0.205,
      "step": 7340
    },
    {
      "epoch": 0.9503835323817846,
      "grad_norm": 1.9562294333358496,
      "learning_rate": 9.340454870519767e-09,
      "loss": 2.3325,
      "step": 7341
    },
    {
      "epoch": 0.9505129947891381,
      "grad_norm": 2.6576745196322142,
      "learning_rate": 9.291864384806253e-09,
      "loss": 2.5793,
      "step": 7342
    },
    {
      "epoch": 0.9506424571964915,
      "grad_norm": 1.2614410037030606,
      "learning_rate": 9.24339966995938e-09,
      "loss": 2.2195,
      "step": 7343
    },
    {
      "epoch": 0.950771919603845,
      "grad_norm": 1.8564653598604044,
      "learning_rate": 9.195060735878794e-09,
      "loss": 2.5464,
      "step": 7344
    },
    {
      "epoch": 0.9509013820111984,
      "grad_norm": 1.7508264041034656,
      "learning_rate": 9.14684759243889e-09,
      "loss": 2.4492,
      "step": 7345
    },
    {
      "epoch": 0.9509013820111984,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.496537685394287,
      "eval_runtime": 13.8925,
      "eval_samples_per_second": 3.167,
      "eval_steps_per_second": 0.216,
      "step": 7345
    },
    {
      "epoch": 0.951030844418552,
      "grad_norm": 1.6623442134018056,
      "learning_rate": 9.098760249488178e-09,
      "loss": 2.454,
      "step": 7346
    },
    {
      "epoch": 0.9511603068259055,
      "grad_norm": 1.5296077539592157,
      "learning_rate": 9.050798716849427e-09,
      "loss": 2.2803,
      "step": 7347
    },
    {
      "epoch": 0.9512897692332589,
      "grad_norm": 2.7290107666186083,
      "learning_rate": 9.002963004319728e-09,
      "loss": 2.3677,
      "step": 7348
    },
    {
      "epoch": 0.9514192316406124,
      "grad_norm": 1.1660048225786328,
      "learning_rate": 8.955253121670573e-09,
      "loss": 2.3154,
      "step": 7349
    },
    {
      "epoch": 0.9515486940479658,
      "grad_norm": 1.3474657693293968,
      "learning_rate": 8.907669078647568e-09,
      "loss": 2.2114,
      "step": 7350
    },
    {
      "epoch": 0.9515486940479658,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964045286178589,
      "eval_runtime": 14.0514,
      "eval_samples_per_second": 3.131,
      "eval_steps_per_second": 0.214,
      "step": 7350
    },
    {
      "epoch": 0.9516781564553193,
      "grad_norm": 1.2518515538968766,
      "learning_rate": 8.860210884970855e-09,
      "loss": 2.2188,
      "step": 7351
    },
    {
      "epoch": 0.9518076188626727,
      "grad_norm": 1.5542544330508121,
      "learning_rate": 8.812878550334553e-09,
      "loss": 2.4314,
      "step": 7352
    },
    {
      "epoch": 0.9519370812700262,
      "grad_norm": 1.5253678710962868,
      "learning_rate": 8.765672084407455e-09,
      "loss": 2.3535,
      "step": 7353
    },
    {
      "epoch": 0.9520665436773796,
      "grad_norm": 1.8207910289766538,
      "learning_rate": 8.718591496832337e-09,
      "loss": 2.1721,
      "step": 7354
    },
    {
      "epoch": 0.9521960060847331,
      "grad_norm": 1.485413547109573,
      "learning_rate": 8.671636797226227e-09,
      "loss": 2.3162,
      "step": 7355
    },
    {
      "epoch": 0.9521960060847331,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4928532838821411,
      "eval_runtime": 14.1456,
      "eval_samples_per_second": 3.111,
      "eval_steps_per_second": 0.212,
      "step": 7355
    },
    {
      "epoch": 0.9523254684920867,
      "grad_norm": 1.5397106453179055,
      "learning_rate": 8.624807995180825e-09,
      "loss": 2.2751,
      "step": 7356
    },
    {
      "epoch": 0.9524549308994401,
      "grad_norm": 1.3819130490074076,
      "learning_rate": 8.578105100261676e-09,
      "loss": 2.2366,
      "step": 7357
    },
    {
      "epoch": 0.9525843933067936,
      "grad_norm": 2.6660214328035488,
      "learning_rate": 8.531528122008929e-09,
      "loss": 2.49,
      "step": 7358
    },
    {
      "epoch": 0.952713855714147,
      "grad_norm": 1.0891506131120157,
      "learning_rate": 8.485077069936773e-09,
      "loss": 2.3694,
      "step": 7359
    },
    {
      "epoch": 0.9528433181215005,
      "grad_norm": 1.5732031233792254,
      "learning_rate": 8.438751953533803e-09,
      "loss": 2.4065,
      "step": 7360
    },
    {
      "epoch": 0.9528433181215005,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.495805263519287,
      "eval_runtime": 15.1968,
      "eval_samples_per_second": 2.895,
      "eval_steps_per_second": 0.197,
      "step": 7360
    },
    {
      "epoch": 0.9529727805288539,
      "grad_norm": 1.1946812534155529,
      "learning_rate": 8.392552782262934e-09,
      "loss": 2.1963,
      "step": 7361
    },
    {
      "epoch": 0.9531022429362074,
      "grad_norm": 1.8128307529180057,
      "learning_rate": 8.34647956556113e-09,
      "loss": 2.2092,
      "step": 7362
    },
    {
      "epoch": 0.9532317053435608,
      "grad_norm": 3.097269006809245,
      "learning_rate": 8.300532312839894e-09,
      "loss": 2.1963,
      "step": 7363
    },
    {
      "epoch": 0.9533611677509143,
      "grad_norm": 1.3939014814617985,
      "learning_rate": 8.254711033484907e-09,
      "loss": 2.2734,
      "step": 7364
    },
    {
      "epoch": 0.9534906301582677,
      "grad_norm": 1.8002878653893486,
      "learning_rate": 8.209015736856049e-09,
      "loss": 2.3633,
      "step": 7365
    },
    {
      "epoch": 0.9534906301582677,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971591234207153,
      "eval_runtime": 14.3697,
      "eval_samples_per_second": 3.062,
      "eval_steps_per_second": 0.209,
      "step": 7365
    },
    {
      "epoch": 0.9536200925656213,
      "grad_norm": 2.7013738275137995,
      "learning_rate": 8.163446432287516e-09,
      "loss": 2.364,
      "step": 7366
    },
    {
      "epoch": 0.9537495549729748,
      "grad_norm": 1.2682560711328297,
      "learning_rate": 8.118003129087698e-09,
      "loss": 2.0994,
      "step": 7367
    },
    {
      "epoch": 0.9538790173803282,
      "grad_norm": 1.7449854370050677,
      "learning_rate": 8.072685836539376e-09,
      "loss": 2.2263,
      "step": 7368
    },
    {
      "epoch": 0.9540084797876817,
      "grad_norm": 1.2615681162859909,
      "learning_rate": 8.02749456389959e-09,
      "loss": 2.3801,
      "step": 7369
    },
    {
      "epoch": 0.9541379421950351,
      "grad_norm": 2.5175617986323386,
      "learning_rate": 7.98242932039943e-09,
      "loss": 2.4434,
      "step": 7370
    },
    {
      "epoch": 0.9541379421950351,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972034692764282,
      "eval_runtime": 14.2087,
      "eval_samples_per_second": 3.097,
      "eval_steps_per_second": 0.211,
      "step": 7370
    },
    {
      "epoch": 0.9542674046023886,
      "grad_norm": 1.3913174879237014,
      "learning_rate": 7.937490115244514e-09,
      "loss": 2.4275,
      "step": 7371
    },
    {
      "epoch": 0.954396867009742,
      "grad_norm": 1.6307154292148878,
      "learning_rate": 7.892676957614517e-09,
      "loss": 2.4678,
      "step": 7372
    },
    {
      "epoch": 0.9545263294170955,
      "grad_norm": 1.387128514893371,
      "learning_rate": 7.847989856663436e-09,
      "loss": 2.4358,
      "step": 7373
    },
    {
      "epoch": 0.9546557918244489,
      "grad_norm": 1.8225805585392338,
      "learning_rate": 7.80342882151952e-09,
      "loss": 2.3372,
      "step": 7374
    },
    {
      "epoch": 0.9547852542318025,
      "grad_norm": 1.591653032360651,
      "learning_rate": 7.758993861285216e-09,
      "loss": 2.5059,
      "step": 7375
    },
    {
      "epoch": 0.9547852542318025,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4959162473678589,
      "eval_runtime": 13.7938,
      "eval_samples_per_second": 3.19,
      "eval_steps_per_second": 0.217,
      "step": 7375
    },
    {
      "epoch": 0.954914716639156,
      "grad_norm": 2.3955796559237976,
      "learning_rate": 7.714684985037426e-09,
      "loss": 2.1912,
      "step": 7376
    },
    {
      "epoch": 0.9550441790465094,
      "grad_norm": 1.931298410374395,
      "learning_rate": 7.670502201827038e-09,
      "loss": 2.2544,
      "step": 7377
    },
    {
      "epoch": 0.9551736414538629,
      "grad_norm": 2.271405774693737,
      "learning_rate": 7.62644552067919e-09,
      "loss": 2.3115,
      "step": 7378
    },
    {
      "epoch": 0.9553031038612163,
      "grad_norm": 1.8591363206488856,
      "learning_rate": 7.582514950593491e-09,
      "loss": 2.3755,
      "step": 7379
    },
    {
      "epoch": 0.9554325662685698,
      "grad_norm": 1.7811899413870635,
      "learning_rate": 7.538710500543528e-09,
      "loss": 2.5339,
      "step": 7380
    },
    {
      "epoch": 0.9554325662685698,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.3729,
      "eval_samples_per_second": 3.061,
      "eval_steps_per_second": 0.209,
      "step": 7380
    },
    {
      "epoch": 0.9555620286759232,
      "grad_norm": 1.5921141649860202,
      "learning_rate": 7.495032179477349e-09,
      "loss": 2.2957,
      "step": 7381
    },
    {
      "epoch": 0.9556914910832767,
      "grad_norm": 3.8904089006426767,
      "learning_rate": 7.451479996317123e-09,
      "loss": 2.3994,
      "step": 7382
    },
    {
      "epoch": 0.9558209534906301,
      "grad_norm": 2.158294674467274,
      "learning_rate": 7.408053959959277e-09,
      "loss": 2.3506,
      "step": 7383
    },
    {
      "epoch": 0.9559504158979836,
      "grad_norm": 3.405678172720716,
      "learning_rate": 7.364754079274352e-09,
      "loss": 2.8477,
      "step": 7384
    },
    {
      "epoch": 0.9560798783053371,
      "grad_norm": 4.029904704501532,
      "learning_rate": 7.321580363107359e-09,
      "loss": 2.4832,
      "step": 7385
    },
    {
      "epoch": 0.9560798783053371,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950283765792847,
      "eval_runtime": 14.5052,
      "eval_samples_per_second": 3.033,
      "eval_steps_per_second": 0.207,
      "step": 7385
    },
    {
      "epoch": 0.9562093407126906,
      "grad_norm": 2.7943147027713455,
      "learning_rate": 7.2785328202772815e-09,
      "loss": 2.3892,
      "step": 7386
    },
    {
      "epoch": 0.956338803120044,
      "grad_norm": 2.83647018930182,
      "learning_rate": 7.235611459577574e-09,
      "loss": 2.4927,
      "step": 7387
    },
    {
      "epoch": 0.9564682655273975,
      "grad_norm": 2.5348447177149214,
      "learning_rate": 7.192816289775736e-09,
      "loss": 2.2959,
      "step": 7388
    },
    {
      "epoch": 0.956597727934751,
      "grad_norm": 1.4694890842365598,
      "learning_rate": 7.150147319613596e-09,
      "loss": 2.2485,
      "step": 7389
    },
    {
      "epoch": 0.9567271903421044,
      "grad_norm": 1.2651517160774781,
      "learning_rate": 7.107604557807027e-09,
      "loss": 2.1941,
      "step": 7390
    },
    {
      "epoch": 0.9567271903421044,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4943181276321411,
      "eval_runtime": 14.1163,
      "eval_samples_per_second": 3.117,
      "eval_steps_per_second": 0.213,
      "step": 7390
    },
    {
      "epoch": 0.9568566527494579,
      "grad_norm": 2.060285527833526,
      "learning_rate": 7.06518801304637e-09,
      "loss": 2.4399,
      "step": 7391
    },
    {
      "epoch": 0.9569861151568113,
      "grad_norm": 1.8839575616650006,
      "learning_rate": 7.0228976939960145e-09,
      "loss": 2.4902,
      "step": 7392
    },
    {
      "epoch": 0.9571155775641648,
      "grad_norm": 1.2234292033772605,
      "learning_rate": 6.980733609294674e-09,
      "loss": 2.2197,
      "step": 7393
    },
    {
      "epoch": 0.9572450399715182,
      "grad_norm": 2.182477749736551,
      "learning_rate": 6.938695767555112e-09,
      "loss": 2.5063,
      "step": 7394
    },
    {
      "epoch": 0.9573745023788718,
      "grad_norm": 1.2196145707447217,
      "learning_rate": 6.896784177364557e-09,
      "loss": 2.3311,
      "step": 7395
    },
    {
      "epoch": 0.9573745023788718,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.496382236480713,
      "eval_runtime": 14.1676,
      "eval_samples_per_second": 3.106,
      "eval_steps_per_second": 0.212,
      "step": 7395
    },
    {
      "epoch": 0.9575039647862252,
      "grad_norm": 1.4896137056307857,
      "learning_rate": 6.854998847284147e-09,
      "loss": 2.4617,
      "step": 7396
    },
    {
      "epoch": 0.9576334271935787,
      "grad_norm": 1.7910939064656823,
      "learning_rate": 6.8133397858494844e-09,
      "loss": 2.3257,
      "step": 7397
    },
    {
      "epoch": 0.9577628896009321,
      "grad_norm": 2.3105936271762704,
      "learning_rate": 6.771807001570221e-09,
      "loss": 2.5579,
      "step": 7398
    },
    {
      "epoch": 0.9578923520082856,
      "grad_norm": 2.24662579700072,
      "learning_rate": 6.730400502930265e-09,
      "loss": 2.4114,
      "step": 7399
    },
    {
      "epoch": 0.958021814415639,
      "grad_norm": 1.642161690725086,
      "learning_rate": 6.68912029838785e-09,
      "loss": 2.3645,
      "step": 7400
    },
    {
      "epoch": 0.958021814415639,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4935857057571411,
      "eval_runtime": 13.9396,
      "eval_samples_per_second": 3.156,
      "eval_steps_per_second": 0.215,
      "step": 7400
    },
    {
      "epoch": 0.9581512768229925,
      "grad_norm": 2.8475943290860273,
      "learning_rate": 6.647966396375119e-09,
      "loss": 2.1982,
      "step": 7401
    },
    {
      "epoch": 0.958280739230346,
      "grad_norm": 2.5673135935940006,
      "learning_rate": 6.606938805298682e-09,
      "loss": 2.3123,
      "step": 7402
    },
    {
      "epoch": 0.9584102016376994,
      "grad_norm": 1.7357702681693234,
      "learning_rate": 6.566037533539266e-09,
      "loss": 2.2329,
      "step": 7403
    },
    {
      "epoch": 0.9585396640450529,
      "grad_norm": 2.218895473209508,
      "learning_rate": 6.525262589451784e-09,
      "loss": 2.4163,
      "step": 7404
    },
    {
      "epoch": 0.9586691264524064,
      "grad_norm": 1.2753904555028839,
      "learning_rate": 6.484613981365337e-09,
      "loss": 2.1873,
      "step": 7405
    },
    {
      "epoch": 0.9586691264524064,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.497270107269287,
      "eval_runtime": 15.0803,
      "eval_samples_per_second": 2.918,
      "eval_steps_per_second": 0.199,
      "step": 7405
    },
    {
      "epoch": 0.9587985888597599,
      "grad_norm": 1.6352691522700775,
      "learning_rate": 6.4440917175832144e-09,
      "loss": 2.3765,
      "step": 7406
    },
    {
      "epoch": 0.9589280512671133,
      "grad_norm": 1.4377703134113937,
      "learning_rate": 6.4036958063828925e-09,
      "loss": 2.2217,
      "step": 7407
    },
    {
      "epoch": 0.9590575136744668,
      "grad_norm": 2.380025249900769,
      "learning_rate": 6.3634262560161035e-09,
      "loss": 2.4846,
      "step": 7408
    },
    {
      "epoch": 0.9591869760818202,
      "grad_norm": 1.9385130878747947,
      "learning_rate": 6.3232830747086275e-09,
      "loss": 2.3342,
      "step": 7409
    },
    {
      "epoch": 0.9593164384891737,
      "grad_norm": 1.8136432467278916,
      "learning_rate": 6.283266270660712e-09,
      "loss": 2.4487,
      "step": 7410
    },
    {
      "epoch": 0.9593164384891737,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4942959547042847,
      "eval_runtime": 16.5124,
      "eval_samples_per_second": 2.665,
      "eval_steps_per_second": 0.182,
      "step": 7410
    },
    {
      "epoch": 0.9594459008965271,
      "grad_norm": 1.7327231883119312,
      "learning_rate": 6.243375852046374e-09,
      "loss": 2.283,
      "step": 7411
    },
    {
      "epoch": 0.9595753633038806,
      "grad_norm": 1.7306148387042344,
      "learning_rate": 6.203611827014233e-09,
      "loss": 2.4419,
      "step": 7412
    },
    {
      "epoch": 0.959704825711234,
      "grad_norm": 2.5763046055640126,
      "learning_rate": 6.163974203686751e-09,
      "loss": 2.3936,
      "step": 7413
    },
    {
      "epoch": 0.9598342881185876,
      "grad_norm": 1.2926971777747578,
      "learning_rate": 6.124462990160715e-09,
      "loss": 2.376,
      "step": 7414
    },
    {
      "epoch": 0.9599637505259411,
      "grad_norm": 1.6452507064872957,
      "learning_rate": 6.0850781945072375e-09,
      "loss": 2.2906,
      "step": 7415
    },
    {
      "epoch": 0.9599637505259411,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972034692764282,
      "eval_runtime": 14.2594,
      "eval_samples_per_second": 3.086,
      "eval_steps_per_second": 0.21,
      "step": 7415
    },
    {
      "epoch": 0.9600932129332945,
      "grad_norm": 1.648440044225304,
      "learning_rate": 6.0458198247713425e-09,
      "loss": 2.4849,
      "step": 7416
    },
    {
      "epoch": 0.960222675340648,
      "grad_norm": 1.7107653427103073,
      "learning_rate": 6.006687888972379e-09,
      "loss": 2.439,
      "step": 7417
    },
    {
      "epoch": 0.9603521377480014,
      "grad_norm": 1.3902533686193985,
      "learning_rate": 5.967682395103883e-09,
      "loss": 2.4209,
      "step": 7418
    },
    {
      "epoch": 0.9604816001553549,
      "grad_norm": 1.2322505205847083,
      "learning_rate": 5.92880335113337e-09,
      "loss": 2.2783,
      "step": 7419
    },
    {
      "epoch": 0.9606110625627083,
      "grad_norm": 1.3683754312834826,
      "learning_rate": 5.890050765002819e-09,
      "loss": 2.3694,
      "step": 7420
    },
    {
      "epoch": 0.9606110625627083,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.497114658355713,
      "eval_runtime": 13.6502,
      "eval_samples_per_second": 3.223,
      "eval_steps_per_second": 0.22,
      "step": 7420
    },
    {
      "epoch": 0.9607405249700618,
      "grad_norm": 1.4658658248581233,
      "learning_rate": 5.851424644628262e-09,
      "loss": 2.5957,
      "step": 7421
    },
    {
      "epoch": 0.9608699873774152,
      "grad_norm": 2.9531600162078013,
      "learning_rate": 5.812924997899705e-09,
      "loss": 2.3623,
      "step": 7422
    },
    {
      "epoch": 0.9609994497847687,
      "grad_norm": 1.869778510641594,
      "learning_rate": 5.774551832681621e-09,
      "loss": 2.302,
      "step": 7423
    },
    {
      "epoch": 0.9611289121921223,
      "grad_norm": 1.6932333173121359,
      "learning_rate": 5.736305156812394e-09,
      "loss": 2.4238,
      "step": 7424
    },
    {
      "epoch": 0.9612583745994757,
      "grad_norm": 1.2830919415889201,
      "learning_rate": 5.698184978104731e-09,
      "loss": 2.2888,
      "step": 7425
    },
    {
      "epoch": 0.9612583745994757,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4970924854278564,
      "eval_runtime": 14.0412,
      "eval_samples_per_second": 3.134,
      "eval_steps_per_second": 0.214,
      "step": 7425
    },
    {
      "epoch": 0.9613878370068292,
      "grad_norm": 1.6564187720301609,
      "learning_rate": 5.66019130434553e-09,
      "loss": 2.3262,
      "step": 7426
    },
    {
      "epoch": 0.9615172994141826,
      "grad_norm": 2.1426401544146647,
      "learning_rate": 5.622324143295665e-09,
      "loss": 2.4475,
      "step": 7427
    },
    {
      "epoch": 0.9616467618215361,
      "grad_norm": 1.8369538278154183,
      "learning_rate": 5.584583502690338e-09,
      "loss": 2.488,
      "step": 7428
    },
    {
      "epoch": 0.9617762242288895,
      "grad_norm": 2.5430452042533878,
      "learning_rate": 5.54696939023873e-09,
      "loss": 2.3572,
      "step": 7429
    },
    {
      "epoch": 0.961905686636243,
      "grad_norm": 1.909907000933575,
      "learning_rate": 5.509481813624484e-09,
      "loss": 2.2617,
      "step": 7430
    },
    {
      "epoch": 0.961905686636243,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.0165,
      "eval_samples_per_second": 3.139,
      "eval_steps_per_second": 0.214,
      "step": 7430
    },
    {
      "epoch": 0.9620351490435964,
      "grad_norm": 1.4639130920563277,
      "learning_rate": 5.47212078050495e-09,
      "loss": 2.3118,
      "step": 7431
    },
    {
      "epoch": 0.9621646114509499,
      "grad_norm": 2.1443348856738926,
      "learning_rate": 5.434886298512146e-09,
      "loss": 2.4409,
      "step": 7432
    },
    {
      "epoch": 0.9622940738583033,
      "grad_norm": 1.1973969370564006,
      "learning_rate": 5.397778375251725e-09,
      "loss": 2.4333,
      "step": 7433
    },
    {
      "epoch": 0.9624235362656569,
      "grad_norm": 1.3432746464111354,
      "learning_rate": 5.360797018303873e-09,
      "loss": 2.3792,
      "step": 7434
    },
    {
      "epoch": 0.9625529986730104,
      "grad_norm": 4.4104965382615555,
      "learning_rate": 5.323942235222757e-09,
      "loss": 2.5217,
      "step": 7435
    },
    {
      "epoch": 0.9625529986730104,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957164525985718,
      "eval_runtime": 14.6662,
      "eval_samples_per_second": 3.0,
      "eval_steps_per_second": 0.205,
      "step": 7435
    },
    {
      "epoch": 0.9626824610803638,
      "grad_norm": 1.2670258732032245,
      "learning_rate": 5.287214033536658e-09,
      "loss": 2.3494,
      "step": 7436
    },
    {
      "epoch": 0.9628119234877173,
      "grad_norm": 4.482408149177954,
      "learning_rate": 5.2506124207480496e-09,
      "loss": 2.8271,
      "step": 7437
    },
    {
      "epoch": 0.9629413858950707,
      "grad_norm": 1.4480607877057987,
      "learning_rate": 5.214137404333727e-09,
      "loss": 2.2925,
      "step": 7438
    },
    {
      "epoch": 0.9630708483024242,
      "grad_norm": 1.8314618494620087,
      "learning_rate": 5.177788991744259e-09,
      "loss": 2.3604,
      "step": 7439
    },
    {
      "epoch": 0.9632003107097776,
      "grad_norm": 1.7526169349094323,
      "learning_rate": 5.141567190404609e-09,
      "loss": 2.2532,
      "step": 7440
    },
    {
      "epoch": 0.9632003107097776,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957607984542847,
      "eval_runtime": 15.0718,
      "eval_samples_per_second": 2.919,
      "eval_steps_per_second": 0.199,
      "step": 7440
    },
    {
      "epoch": 0.9633297731171311,
      "grad_norm": 2.4784323523124034,
      "learning_rate": 5.105472007713858e-09,
      "loss": 2.4558,
      "step": 7441
    },
    {
      "epoch": 0.9634592355244845,
      "grad_norm": 2.435086425448772,
      "learning_rate": 5.0695034510450656e-09,
      "loss": 2.4861,
      "step": 7442
    },
    {
      "epoch": 0.963588697931838,
      "grad_norm": 1.9545920770822254,
      "learning_rate": 5.033661527745759e-09,
      "loss": 2.3247,
      "step": 7443
    },
    {
      "epoch": 0.9637181603391916,
      "grad_norm": 4.510582731836052,
      "learning_rate": 4.997946245137095e-09,
      "loss": 2.6438,
      "step": 7444
    },
    {
      "epoch": 0.963847622746545,
      "grad_norm": 1.2979363658566292,
      "learning_rate": 4.9623576105148344e-09,
      "loss": 2.2865,
      "step": 7445
    },
    {
      "epoch": 0.963847622746545,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4942294359207153,
      "eval_runtime": 14.6086,
      "eval_samples_per_second": 3.012,
      "eval_steps_per_second": 0.205,
      "step": 7445
    },
    {
      "epoch": 0.9639770851538985,
      "grad_norm": 1.7241561865100121,
      "learning_rate": 4.9268956311485785e-09,
      "loss": 2.2761,
      "step": 7446
    },
    {
      "epoch": 0.9641065475612519,
      "grad_norm": 1.5281300437446959,
      "learning_rate": 4.891560314282187e-09,
      "loss": 2.2876,
      "step": 7447
    },
    {
      "epoch": 0.9642360099686054,
      "grad_norm": 1.4660663313724847,
      "learning_rate": 4.856351667133705e-09,
      "loss": 2.3298,
      "step": 7448
    },
    {
      "epoch": 0.9643654723759588,
      "grad_norm": 1.6712522836740253,
      "learning_rate": 4.821269696895086e-09,
      "loss": 2.2236,
      "step": 7449
    },
    {
      "epoch": 0.9644949347833123,
      "grad_norm": 1.4911135784857747,
      "learning_rate": 4.786314410732615e-09,
      "loss": 2.3521,
      "step": 7450
    },
    {
      "epoch": 0.9644949347833123,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957386255264282,
      "eval_runtime": 13.9619,
      "eval_samples_per_second": 3.151,
      "eval_steps_per_second": 0.215,
      "step": 7450
    },
    {
      "epoch": 0.9646243971906657,
      "grad_norm": 2.0692026567267363,
      "learning_rate": 4.751485815786549e-09,
      "loss": 2.4583,
      "step": 7451
    },
    {
      "epoch": 0.9647538595980192,
      "grad_norm": 1.3609082294235575,
      "learning_rate": 4.716783919171408e-09,
      "loss": 2.1257,
      "step": 7452
    },
    {
      "epoch": 0.9648833220053726,
      "grad_norm": 2.2904256937471663,
      "learning_rate": 4.682208727975687e-09,
      "loss": 2.3555,
      "step": 7453
    },
    {
      "epoch": 0.9650127844127262,
      "grad_norm": 1.152316635432278,
      "learning_rate": 4.647760249262071e-09,
      "loss": 2.2781,
      "step": 7454
    },
    {
      "epoch": 0.9651422468200797,
      "grad_norm": 1.6201693137425381,
      "learning_rate": 4.613438490067429e-09,
      "loss": 2.2991,
      "step": 7455
    },
    {
      "epoch": 0.9651422468200797,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 13.781,
      "eval_samples_per_second": 3.193,
      "eval_steps_per_second": 0.218,
      "step": 7455
    },
    {
      "epoch": 0.9652717092274331,
      "grad_norm": 1.6774732698187402,
      "learning_rate": 4.579243457402682e-09,
      "loss": 2.4812,
      "step": 7456
    },
    {
      "epoch": 0.9654011716347866,
      "grad_norm": 3.3740605434319613,
      "learning_rate": 4.54517515825273e-09,
      "loss": 2.2764,
      "step": 7457
    },
    {
      "epoch": 0.96553063404214,
      "grad_norm": 2.928283095137224,
      "learning_rate": 4.511233599576864e-09,
      "loss": 2.4565,
      "step": 7458
    },
    {
      "epoch": 0.9656600964494935,
      "grad_norm": 1.4831480097936056,
      "learning_rate": 4.477418788308222e-09,
      "loss": 2.3706,
      "step": 7459
    },
    {
      "epoch": 0.9657895588568469,
      "grad_norm": 2.19929370895813,
      "learning_rate": 4.443730731354262e-09,
      "loss": 2.3383,
      "step": 7460
    },
    {
      "epoch": 0.9657895588568469,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950062036514282,
      "eval_runtime": 14.123,
      "eval_samples_per_second": 3.115,
      "eval_steps_per_second": 0.212,
      "step": 7460
    },
    {
      "epoch": 0.9659190212642004,
      "grad_norm": 1.900076421541362,
      "learning_rate": 4.410169435596357e-09,
      "loss": 2.4509,
      "step": 7461
    },
    {
      "epoch": 0.9660484836715538,
      "grad_norm": 2.326648516727546,
      "learning_rate": 4.376734907890132e-09,
      "loss": 2.2881,
      "step": 7462
    },
    {
      "epoch": 0.9661779460789074,
      "grad_norm": 1.4274927574589986,
      "learning_rate": 4.343427155065264e-09,
      "loss": 2.2949,
      "step": 7463
    },
    {
      "epoch": 0.9663074084862608,
      "grad_norm": 2.5119325293108514,
      "learning_rate": 4.310246183925615e-09,
      "loss": 2.3308,
      "step": 7464
    },
    {
      "epoch": 0.9664368708936143,
      "grad_norm": 1.3129529473112151,
      "learning_rate": 4.27719200124889e-09,
      "loss": 2.3574,
      "step": 7465
    },
    {
      "epoch": 0.9664368708936143,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950283765792847,
      "eval_runtime": 15.8373,
      "eval_samples_per_second": 2.778,
      "eval_steps_per_second": 0.189,
      "step": 7465
    },
    {
      "epoch": 0.9665663333009678,
      "grad_norm": 2.387842532371235,
      "learning_rate": 4.244264613787186e-09,
      "loss": 2.6504,
      "step": 7466
    },
    {
      "epoch": 0.9666957957083212,
      "grad_norm": 1.7636110616149214,
      "learning_rate": 4.211464028266651e-09,
      "loss": 2.395,
      "step": 7467
    },
    {
      "epoch": 0.9668252581156747,
      "grad_norm": 1.9109816563681847,
      "learning_rate": 4.178790251387344e-09,
      "loss": 2.5222,
      "step": 7468
    },
    {
      "epoch": 0.9669547205230281,
      "grad_norm": 1.5215191486770077,
      "learning_rate": 4.146243289823576e-09,
      "loss": 2.2246,
      "step": 7469
    },
    {
      "epoch": 0.9670841829303816,
      "grad_norm": 1.5342723301472245,
      "learning_rate": 4.113823150223781e-09,
      "loss": 2.3511,
      "step": 7470
    },
    {
      "epoch": 0.9670841829303816,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978693723678589,
      "eval_runtime": 15.8859,
      "eval_samples_per_second": 2.77,
      "eval_steps_per_second": 0.189,
      "step": 7470
    },
    {
      "epoch": 0.967213645337735,
      "grad_norm": 1.6056039831797502,
      "learning_rate": 4.081529839210438e-09,
      "loss": 2.4873,
      "step": 7471
    },
    {
      "epoch": 0.9673431077450885,
      "grad_norm": 1.3977792438402266,
      "learning_rate": 4.0493633633800075e-09,
      "loss": 2.321,
      "step": 7472
    },
    {
      "epoch": 0.967472570152442,
      "grad_norm": 1.5323062202084872,
      "learning_rate": 4.017323729303204e-09,
      "loss": 2.2068,
      "step": 7473
    },
    {
      "epoch": 0.9676020325597955,
      "grad_norm": 1.3573902372216242,
      "learning_rate": 3.985410943524864e-09,
      "loss": 2.2437,
      "step": 7474
    },
    {
      "epoch": 0.9677314949671489,
      "grad_norm": 2.1150732118529554,
      "learning_rate": 3.953625012563592e-09,
      "loss": 2.3384,
      "step": 7475
    },
    {
      "epoch": 0.9677314949671489,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971591234207153,
      "eval_runtime": 14.4355,
      "eval_samples_per_second": 3.048,
      "eval_steps_per_second": 0.208,
      "step": 7475
    },
    {
      "epoch": 0.9678609573745024,
      "grad_norm": 1.6399905266065677,
      "learning_rate": 3.921965942912595e-09,
      "loss": 2.5869,
      "step": 7476
    },
    {
      "epoch": 0.9679904197818558,
      "grad_norm": 1.2362791833419275,
      "learning_rate": 3.890433741038646e-09,
      "loss": 2.1628,
      "step": 7477
    },
    {
      "epoch": 0.9681198821892093,
      "grad_norm": 1.2730674305757086,
      "learning_rate": 3.859028413382912e-09,
      "loss": 2.2163,
      "step": 7478
    },
    {
      "epoch": 0.9682493445965628,
      "grad_norm": 2.3378260534166344,
      "learning_rate": 3.827749966360609e-09,
      "loss": 2.864,
      "step": 7479
    },
    {
      "epoch": 0.9683788070039162,
      "grad_norm": 1.8525815863145987,
      "learning_rate": 3.79659840636093e-09,
      "loss": 2.7358,
      "step": 7480
    },
    {
      "epoch": 0.9683788070039162,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4943181276321411,
      "eval_runtime": 14.2482,
      "eval_samples_per_second": 3.088,
      "eval_steps_per_second": 0.211,
      "step": 7480
    },
    {
      "epoch": 0.9685082694112697,
      "grad_norm": 4.512358638693356,
      "learning_rate": 3.765573739747258e-09,
      "loss": 2.5627,
      "step": 7481
    },
    {
      "epoch": 0.9686377318186231,
      "grad_norm": 1.3330150408292283,
      "learning_rate": 3.734675972857093e-09,
      "loss": 2.4629,
      "step": 7482
    },
    {
      "epoch": 0.9687671942259767,
      "grad_norm": 2.155860407525293,
      "learning_rate": 3.7039051120017074e-09,
      "loss": 2.459,
      "step": 7483
    },
    {
      "epoch": 0.9688966566333301,
      "grad_norm": 2.5748492806283045,
      "learning_rate": 3.6732611634668347e-09,
      "loss": 3.0095,
      "step": 7484
    },
    {
      "epoch": 0.9690261190406836,
      "grad_norm": 2.4978403677276537,
      "learning_rate": 3.6427441335121218e-09,
      "loss": 2.3953,
      "step": 7485
    },
    {
      "epoch": 0.9690261190406836,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964488744735718,
      "eval_runtime": 14.3139,
      "eval_samples_per_second": 3.074,
      "eval_steps_per_second": 0.21,
      "step": 7485
    },
    {
      "epoch": 0.969155581448037,
      "grad_norm": 3.0965877420364394,
      "learning_rate": 3.6123540283711927e-09,
      "loss": 2.4995,
      "step": 7486
    },
    {
      "epoch": 0.9692850438553905,
      "grad_norm": 1.2981925691305594,
      "learning_rate": 3.582090854251999e-09,
      "loss": 2.3113,
      "step": 7487
    },
    {
      "epoch": 0.969414506262744,
      "grad_norm": 2.3897403198169758,
      "learning_rate": 3.551954617336262e-09,
      "loss": 2.0898,
      "step": 7488
    },
    {
      "epoch": 0.9695439686700974,
      "grad_norm": 2.242910065031112,
      "learning_rate": 3.52194532377996e-09,
      "loss": 2.2312,
      "step": 7489
    },
    {
      "epoch": 0.9696734310774509,
      "grad_norm": 1.5433396722517625,
      "learning_rate": 3.4920629797131206e-09,
      "loss": 2.4941,
      "step": 7490
    },
    {
      "epoch": 0.9696734310774509,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4951171875,
      "eval_runtime": 13.7626,
      "eval_samples_per_second": 3.197,
      "eval_steps_per_second": 0.218,
      "step": 7490
    },
    {
      "epoch": 0.9698028934848043,
      "grad_norm": 1.9289326942626048,
      "learning_rate": 3.4623075912398193e-09,
      "loss": 2.4224,
      "step": 7491
    },
    {
      "epoch": 0.9699323558921578,
      "grad_norm": 3.511050789255123,
      "learning_rate": 3.4326791644381794e-09,
      "loss": 2.5933,
      "step": 7492
    },
    {
      "epoch": 0.9700618182995113,
      "grad_norm": 1.8380120276262875,
      "learning_rate": 3.403177705360444e-09,
      "loss": 2.3831,
      "step": 7493
    },
    {
      "epoch": 0.9701912807068648,
      "grad_norm": 1.7896693346471626,
      "learning_rate": 3.373803220032834e-09,
      "loss": 2.4495,
      "step": 7494
    },
    {
      "epoch": 0.9703207431142182,
      "grad_norm": 1.2570198969761672,
      "learning_rate": 3.344555714455688e-09,
      "loss": 2.4331,
      "step": 7495
    },
    {
      "epoch": 0.9703207431142182,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986461400985718,
      "eval_runtime": 14.7597,
      "eval_samples_per_second": 2.981,
      "eval_steps_per_second": 0.203,
      "step": 7495
    },
    {
      "epoch": 0.9704502055215717,
      "grad_norm": 1.3574737833534676,
      "learning_rate": 3.315435194603464e-09,
      "loss": 2.3708,
      "step": 7496
    },
    {
      "epoch": 0.9705796679289251,
      "grad_norm": 2.683654598506334,
      "learning_rate": 3.286441666424528e-09,
      "loss": 2.3386,
      "step": 7497
    },
    {
      "epoch": 0.9707091303362786,
      "grad_norm": 2.429968705412448,
      "learning_rate": 3.2575751358414344e-09,
      "loss": 2.2017,
      "step": 7498
    },
    {
      "epoch": 0.970838592743632,
      "grad_norm": 2.379328142632694,
      "learning_rate": 3.228835608750855e-09,
      "loss": 2.6353,
      "step": 7499
    },
    {
      "epoch": 0.9709680551509855,
      "grad_norm": 1.1689158769248162,
      "learning_rate": 3.200223091023233e-09,
      "loss": 2.2852,
      "step": 7500
    },
    {
      "epoch": 0.9709680551509855,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.497114658355713,
      "eval_runtime": 13.8657,
      "eval_samples_per_second": 3.173,
      "eval_steps_per_second": 0.216,
      "step": 7500
    },
    {
      "epoch": 0.971097517558339,
      "grad_norm": 2.078256325407057,
      "learning_rate": 3.171737588503407e-09,
      "loss": 2.4272,
      "step": 7501
    },
    {
      "epoch": 0.9712269799656925,
      "grad_norm": 1.366999089040965,
      "learning_rate": 3.1433791070100555e-09,
      "loss": 2.229,
      "step": 7502
    },
    {
      "epoch": 0.971356442373046,
      "grad_norm": 1.8412842996905656,
      "learning_rate": 3.1151476523359064e-09,
      "loss": 2.374,
      "step": 7503
    },
    {
      "epoch": 0.9714859047803994,
      "grad_norm": 2.4036313021269984,
      "learning_rate": 3.0870432302479435e-09,
      "loss": 2.552,
      "step": 7504
    },
    {
      "epoch": 0.9716153671877529,
      "grad_norm": 1.3319207705837204,
      "learning_rate": 3.059065846487061e-09,
      "loss": 2.4368,
      "step": 7505
    },
    {
      "epoch": 0.9716153671877529,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.552,
      "eval_samples_per_second": 3.024,
      "eval_steps_per_second": 0.206,
      "step": 7505
    },
    {
      "epoch": 0.9717448295951063,
      "grad_norm": 2.604357541325846,
      "learning_rate": 3.031215506768062e-09,
      "loss": 2.4111,
      "step": 7506
    },
    {
      "epoch": 0.9718742920024598,
      "grad_norm": 1.7566724147974662,
      "learning_rate": 3.003492216780077e-09,
      "loss": 2.4978,
      "step": 7507
    },
    {
      "epoch": 0.9720037544098132,
      "grad_norm": 1.8860731801998685,
      "learning_rate": 2.975895982186075e-09,
      "loss": 2.1368,
      "step": 7508
    },
    {
      "epoch": 0.9721332168171667,
      "grad_norm": 1.336728000590379,
      "learning_rate": 2.948426808623145e-09,
      "loss": 2.3132,
      "step": 7509
    },
    {
      "epoch": 0.9722626792245201,
      "grad_norm": 1.6261016992599011,
      "learning_rate": 2.9210847017024922e-09,
      "loss": 2.2693,
      "step": 7510
    },
    {
      "epoch": 0.9722626792245201,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.495072841644287,
      "eval_runtime": 14.6337,
      "eval_samples_per_second": 3.007,
      "eval_steps_per_second": 0.205,
      "step": 7510
    },
    {
      "epoch": 0.9723921416318736,
      "grad_norm": 1.7337031603610027,
      "learning_rate": 2.893869667009233e-09,
      "loss": 2.573,
      "step": 7511
    },
    {
      "epoch": 0.9725216040392272,
      "grad_norm": 1.516881266320023,
      "learning_rate": 2.866781710102601e-09,
      "loss": 2.4393,
      "step": 7512
    },
    {
      "epoch": 0.9726510664465806,
      "grad_norm": 1.778019440023385,
      "learning_rate": 2.839820836515947e-09,
      "loss": 2.4497,
      "step": 7513
    },
    {
      "epoch": 0.9727805288539341,
      "grad_norm": 1.649562667349022,
      "learning_rate": 2.812987051756394e-09,
      "loss": 2.3696,
      "step": 7514
    },
    {
      "epoch": 0.9729099912612875,
      "grad_norm": 1.6709675287516073,
      "learning_rate": 2.78628036130546e-09,
      "loss": 2.7583,
      "step": 7515
    },
    {
      "epoch": 0.9729099912612875,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964045286178589,
      "eval_runtime": 14.1338,
      "eval_samples_per_second": 3.113,
      "eval_steps_per_second": 0.212,
      "step": 7515
    },
    {
      "epoch": 0.973039453668641,
      "grad_norm": 1.208623897904988,
      "learning_rate": 2.7597007706184344e-09,
      "loss": 2.2378,
      "step": 7516
    },
    {
      "epoch": 0.9731689160759944,
      "grad_norm": 1.3760642333243103,
      "learning_rate": 2.7332482851247926e-09,
      "loss": 2.4258,
      "step": 7517
    },
    {
      "epoch": 0.9732983784833479,
      "grad_norm": 3.037334589617975,
      "learning_rate": 2.7069229102279217e-09,
      "loss": 2.3164,
      "step": 7518
    },
    {
      "epoch": 0.9734278408907013,
      "grad_norm": 1.5660663366500944,
      "learning_rate": 2.680724651305325e-09,
      "loss": 2.1655,
      "step": 7519
    },
    {
      "epoch": 0.9735573032980548,
      "grad_norm": 2.3157849481641852,
      "learning_rate": 2.6546535137086244e-09,
      "loss": 2.5334,
      "step": 7520
    },
    {
      "epoch": 0.9735573032980548,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.4152,
      "eval_samples_per_second": 3.052,
      "eval_steps_per_second": 0.208,
      "step": 7520
    },
    {
      "epoch": 0.9736867657054082,
      "grad_norm": 1.6032188100029758,
      "learning_rate": 2.6287095027632824e-09,
      "loss": 2.3384,
      "step": 7521
    },
    {
      "epoch": 0.9738162281127618,
      "grad_norm": 1.6009070614900751,
      "learning_rate": 2.6028926237689488e-09,
      "loss": 2.4028,
      "step": 7522
    },
    {
      "epoch": 0.9739456905201153,
      "grad_norm": 1.935103516636611,
      "learning_rate": 2.577202881999183e-09,
      "loss": 2.2874,
      "step": 7523
    },
    {
      "epoch": 0.9740751529274687,
      "grad_norm": 1.8155538775672537,
      "learning_rate": 2.5516402827016617e-09,
      "loss": 2.3831,
      "step": 7524
    },
    {
      "epoch": 0.9742046153348222,
      "grad_norm": 1.0483947077967555,
      "learning_rate": 2.5262048310980418e-09,
      "loss": 2.1833,
      "step": 7525
    },
    {
      "epoch": 0.9742046153348222,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.495805263519287,
      "eval_runtime": 14.1041,
      "eval_samples_per_second": 3.12,
      "eval_steps_per_second": 0.213,
      "step": 7525
    },
    {
      "epoch": 0.9743340777421756,
      "grad_norm": 1.552266890430507,
      "learning_rate": 2.500896532384098e-09,
      "loss": 2.1396,
      "step": 7526
    },
    {
      "epoch": 0.9744635401495291,
      "grad_norm": 4.162519199111511,
      "learning_rate": 2.4757153917295136e-09,
      "loss": 3.0793,
      "step": 7527
    },
    {
      "epoch": 0.9745930025568825,
      "grad_norm": 1.4020772771060606,
      "learning_rate": 2.4506614142780216e-09,
      "loss": 2.2017,
      "step": 7528
    },
    {
      "epoch": 0.974722464964236,
      "grad_norm": 2.257257302308101,
      "learning_rate": 2.4257346051474034e-09,
      "loss": 2.3549,
      "step": 7529
    },
    {
      "epoch": 0.9748519273715894,
      "grad_norm": 2.304261799764049,
      "learning_rate": 2.400934969429558e-09,
      "loss": 2.4275,
      "step": 7530
    },
    {
      "epoch": 0.9748519273715894,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.495072841644287,
      "eval_runtime": 13.7187,
      "eval_samples_per_second": 3.207,
      "eval_steps_per_second": 0.219,
      "step": 7530
    },
    {
      "epoch": 0.9749813897789429,
      "grad_norm": 1.358154053410943,
      "learning_rate": 2.376262512190225e-09,
      "loss": 2.2069,
      "step": 7531
    },
    {
      "epoch": 0.9751108521862965,
      "grad_norm": 1.9448873246460772,
      "learning_rate": 2.351717238469331e-09,
      "loss": 2.4498,
      "step": 7532
    },
    {
      "epoch": 0.9752403145936499,
      "grad_norm": 2.323208439137572,
      "learning_rate": 2.3272991532807126e-09,
      "loss": 2.5518,
      "step": 7533
    },
    {
      "epoch": 0.9753697770010034,
      "grad_norm": 1.9000478401818919,
      "learning_rate": 2.3030082616121858e-09,
      "loss": 2.2683,
      "step": 7534
    },
    {
      "epoch": 0.9754992394083568,
      "grad_norm": 2.0009142246658937,
      "learning_rate": 2.2788445684256836e-09,
      "loss": 2.2664,
      "step": 7535
    },
    {
      "epoch": 0.9754992394083568,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978693723678589,
      "eval_runtime": 14.3015,
      "eval_samples_per_second": 3.077,
      "eval_steps_per_second": 0.21,
      "step": 7535
    },
    {
      "epoch": 0.9756287018157103,
      "grad_norm": 1.3166239883775896,
      "learning_rate": 2.254808078657189e-09,
      "loss": 2.3547,
      "step": 7536
    },
    {
      "epoch": 0.9757581642230637,
      "grad_norm": 3.8384585910489326,
      "learning_rate": 2.230898797216663e-09,
      "loss": 2.3691,
      "step": 7537
    },
    {
      "epoch": 0.9758876266304172,
      "grad_norm": 1.5529322232040648,
      "learning_rate": 2.207116728987976e-09,
      "loss": 2.4724,
      "step": 7538
    },
    {
      "epoch": 0.9760170890377706,
      "grad_norm": 1.2127774713992632,
      "learning_rate": 2.1834618788291183e-09,
      "loss": 2.2393,
      "step": 7539
    },
    {
      "epoch": 0.9761465514451241,
      "grad_norm": 2.1174450047835274,
      "learning_rate": 2.1599342515720573e-09,
      "loss": 2.335,
      "step": 7540
    },
    {
      "epoch": 0.9761465514451241,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.493696689605713,
      "eval_runtime": 14.3755,
      "eval_samples_per_second": 3.061,
      "eval_steps_per_second": 0.209,
      "step": 7540
    },
    {
      "epoch": 0.9762760138524776,
      "grad_norm": 2.483136951823826,
      "learning_rate": 2.136533852022879e-09,
      "loss": 2.4524,
      "step": 7541
    },
    {
      "epoch": 0.9764054762598311,
      "grad_norm": 1.5511048057136474,
      "learning_rate": 2.11326068496151e-09,
      "loss": 2.4346,
      "step": 7542
    },
    {
      "epoch": 0.9765349386671845,
      "grad_norm": 1.8425095487094147,
      "learning_rate": 2.0901147551419948e-09,
      "loss": 2.3542,
      "step": 7543
    },
    {
      "epoch": 0.976664401074538,
      "grad_norm": 1.591283586442896,
      "learning_rate": 2.0670960672923567e-09,
      "loss": 2.2773,
      "step": 7544
    },
    {
      "epoch": 0.9767938634818915,
      "grad_norm": 3.219894890994954,
      "learning_rate": 2.044204626114599e-09,
      "loss": 2.686,
      "step": 7545
    },
    {
      "epoch": 0.9767938634818915,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4943181276321411,
      "eval_runtime": 14.3109,
      "eval_samples_per_second": 3.075,
      "eval_steps_per_second": 0.21,
      "step": 7545
    },
    {
      "epoch": 0.9769233258892449,
      "grad_norm": 1.6948323938636074,
      "learning_rate": 2.021440436284841e-09,
      "loss": 2.5454,
      "step": 7546
    },
    {
      "epoch": 0.9770527882965984,
      "grad_norm": 2.0809224454371016,
      "learning_rate": 1.9988035024529756e-09,
      "loss": 2.2063,
      "step": 7547
    },
    {
      "epoch": 0.9771822507039518,
      "grad_norm": 1.8854420394243903,
      "learning_rate": 1.97629382924322e-09,
      "loss": 2.2332,
      "step": 7548
    },
    {
      "epoch": 0.9773117131113053,
      "grad_norm": 2.5643896611272123,
      "learning_rate": 1.9539114212535625e-09,
      "loss": 2.293,
      "step": 7549
    },
    {
      "epoch": 0.9774411755186587,
      "grad_norm": 1.1925655779849842,
      "learning_rate": 1.931656283056041e-09,
      "loss": 2.1436,
      "step": 7550
    },
    {
      "epoch": 0.9774411755186587,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.495649814605713,
      "eval_runtime": 14.3042,
      "eval_samples_per_second": 3.076,
      "eval_steps_per_second": 0.21,
      "step": 7550
    },
    {
      "epoch": 0.9775706379260123,
      "grad_norm": 6.1310483327674925,
      "learning_rate": 1.90952841919681e-09,
      "loss": 2.8223,
      "step": 7551
    },
    {
      "epoch": 0.9777001003333657,
      "grad_norm": 1.8490057934229254,
      "learning_rate": 1.887527834195796e-09,
      "loss": 2.2754,
      "step": 7552
    },
    {
      "epoch": 0.9778295627407192,
      "grad_norm": 1.9545709254230654,
      "learning_rate": 1.8656545325471122e-09,
      "loss": 2.3342,
      "step": 7553
    },
    {
      "epoch": 0.9779590251480726,
      "grad_norm": 1.7527908462995896,
      "learning_rate": 1.8439085187189205e-09,
      "loss": 2.324,
      "step": 7554
    },
    {
      "epoch": 0.9780884875554261,
      "grad_norm": 1.8819808845260853,
      "learning_rate": 1.822289797153154e-09,
      "loss": 2.3325,
      "step": 7555
    },
    {
      "epoch": 0.9780884875554261,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 13.6691,
      "eval_samples_per_second": 3.219,
      "eval_steps_per_second": 0.219,
      "step": 7555
    },
    {
      "epoch": 0.9782179499627796,
      "grad_norm": 1.9981064278922835,
      "learning_rate": 1.8007983722660021e-09,
      "loss": 2.519,
      "step": 7556
    },
    {
      "epoch": 0.978347412370133,
      "grad_norm": 2.209991202418474,
      "learning_rate": 1.7794342484474259e-09,
      "loss": 2.4434,
      "step": 7557
    },
    {
      "epoch": 0.9784768747774865,
      "grad_norm": 2.282220337288573,
      "learning_rate": 1.758197430061434e-09,
      "loss": 2.5867,
      "step": 7558
    },
    {
      "epoch": 0.9786063371848399,
      "grad_norm": 2.386537577649605,
      "learning_rate": 1.7370879214462232e-09,
      "loss": 2.3914,
      "step": 7559
    },
    {
      "epoch": 0.9787357995921934,
      "grad_norm": 1.4211663268329082,
      "learning_rate": 1.7161057269138303e-09,
      "loss": 2.387,
      "step": 7560
    },
    {
      "epoch": 0.9787357995921934,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.494384765625,
      "eval_runtime": 14.0268,
      "eval_samples_per_second": 3.137,
      "eval_steps_per_second": 0.214,
      "step": 7560
    },
    {
      "epoch": 0.9788652619995469,
      "grad_norm": 2.1706050298282404,
      "learning_rate": 1.6952508507501324e-09,
      "loss": 2.1416,
      "step": 7561
    },
    {
      "epoch": 0.9789947244069004,
      "grad_norm": 1.4697119638840281,
      "learning_rate": 1.6745232972152637e-09,
      "loss": 2.3052,
      "step": 7562
    },
    {
      "epoch": 0.9791241868142538,
      "grad_norm": 1.9887583015605543,
      "learning_rate": 1.6539230705432675e-09,
      "loss": 2.7405,
      "step": 7563
    },
    {
      "epoch": 0.9792536492216073,
      "grad_norm": 1.3002785682904727,
      "learning_rate": 1.6334501749421667e-09,
      "loss": 2.3696,
      "step": 7564
    },
    {
      "epoch": 0.9793831116289607,
      "grad_norm": 4.01701817403821,
      "learning_rate": 1.613104614593894e-09,
      "loss": 2.3389,
      "step": 7565
    },
    {
      "epoch": 0.9793831116289607,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498579502105713,
      "eval_runtime": 15.3295,
      "eval_samples_per_second": 2.87,
      "eval_steps_per_second": 0.196,
      "step": 7565
    },
    {
      "epoch": 0.9795125740363142,
      "grad_norm": 2.82389856272306,
      "learning_rate": 1.5928863936544998e-09,
      "loss": 2.4026,
      "step": 7566
    },
    {
      "epoch": 0.9796420364436677,
      "grad_norm": 1.9015822937119669,
      "learning_rate": 1.5727955162539444e-09,
      "loss": 2.387,
      "step": 7567
    },
    {
      "epoch": 0.9797714988510211,
      "grad_norm": 1.1406419283583769,
      "learning_rate": 1.5528319864961672e-09,
      "loss": 2.1597,
      "step": 7568
    },
    {
      "epoch": 0.9799009612583746,
      "grad_norm": 2.1355981721634114,
      "learning_rate": 1.5329958084592255e-09,
      "loss": 2.3271,
      "step": 7569
    },
    {
      "epoch": 0.980030423665728,
      "grad_norm": 1.4628505282297621,
      "learning_rate": 1.5132869861948785e-09,
      "loss": 2.4343,
      "step": 7570
    },
    {
      "epoch": 0.980030423665728,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957829713821411,
      "eval_runtime": 13.8423,
      "eval_samples_per_second": 3.179,
      "eval_steps_per_second": 0.217,
      "step": 7570
    },
    {
      "epoch": 0.9801598860730816,
      "grad_norm": 1.5826016277513821,
      "learning_rate": 1.4937055237292802e-09,
      "loss": 2.3635,
      "step": 7571
    },
    {
      "epoch": 0.980289348480435,
      "grad_norm": 1.2967653618589454,
      "learning_rate": 1.4742514250621481e-09,
      "loss": 2.2004,
      "step": 7572
    },
    {
      "epoch": 0.9804188108877885,
      "grad_norm": 1.348762418289272,
      "learning_rate": 1.454924694167456e-09,
      "loss": 2.439,
      "step": 7573
    },
    {
      "epoch": 0.9805482732951419,
      "grad_norm": 2.0059993316066187,
      "learning_rate": 1.4357253349931571e-09,
      "loss": 2.3491,
      "step": 7574
    },
    {
      "epoch": 0.9806777357024954,
      "grad_norm": 1.0696062980009842,
      "learning_rate": 1.4166533514609754e-09,
      "loss": 2.189,
      "step": 7575
    },
    {
      "epoch": 0.9806777357024954,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971369504928589,
      "eval_runtime": 14.5855,
      "eval_samples_per_second": 3.017,
      "eval_steps_per_second": 0.206,
      "step": 7575
    },
    {
      "epoch": 0.9808071981098488,
      "grad_norm": 1.0643332651729815,
      "learning_rate": 1.3977087474667527e-09,
      "loss": 2.2942,
      "step": 7576
    },
    {
      "epoch": 0.9809366605172023,
      "grad_norm": 1.3466758091724915,
      "learning_rate": 1.3788915268804492e-09,
      "loss": 2.3136,
      "step": 7577
    },
    {
      "epoch": 0.9810661229245558,
      "grad_norm": 2.204496975992815,
      "learning_rate": 1.3602016935457262e-09,
      "loss": 2.4861,
      "step": 7578
    },
    {
      "epoch": 0.9811955853319092,
      "grad_norm": 1.414782482719819,
      "learning_rate": 1.3416392512804327e-09,
      "loss": 2.283,
      "step": 7579
    },
    {
      "epoch": 0.9813250477392627,
      "grad_norm": 2.985776601234628,
      "learning_rate": 1.3232042038762577e-09,
      "loss": 2.5293,
      "step": 7580
    },
    {
      "epoch": 0.9813250477392627,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950283765792847,
      "eval_runtime": 14.9875,
      "eval_samples_per_second": 2.936,
      "eval_steps_per_second": 0.2,
      "step": 7580
    },
    {
      "epoch": 0.9814545101466162,
      "grad_norm": 2.130158472795342,
      "learning_rate": 1.3048965550990083e-09,
      "loss": 2.3813,
      "step": 7581
    },
    {
      "epoch": 0.9815839725539697,
      "grad_norm": 1.623329970728458,
      "learning_rate": 1.2867163086884015e-09,
      "loss": 2.3645,
      "step": 7582
    },
    {
      "epoch": 0.9817134349613231,
      "grad_norm": 1.9459265474325154,
      "learning_rate": 1.268663468358064e-09,
      "loss": 2.3744,
      "step": 7583
    },
    {
      "epoch": 0.9818428973686766,
      "grad_norm": 2.1195756554460385,
      "learning_rate": 1.2507380377956012e-09,
      "loss": 2.4797,
      "step": 7584
    },
    {
      "epoch": 0.98197235977603,
      "grad_norm": 1.7386627470623568,
      "learning_rate": 1.2329400206628068e-09,
      "loss": 2.4109,
      "step": 7585
    },
    {
      "epoch": 0.98197235977603,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4935635328292847,
      "eval_runtime": 13.6963,
      "eval_samples_per_second": 3.213,
      "eval_steps_per_second": 0.219,
      "step": 7585
    },
    {
      "epoch": 0.9821018221833835,
      "grad_norm": 1.1887944261227992,
      "learning_rate": 1.2152694205951054e-09,
      "loss": 2.2488,
      "step": 7586
    },
    {
      "epoch": 0.9822312845907369,
      "grad_norm": 2.586722476127736,
      "learning_rate": 1.1977262412022489e-09,
      "loss": 2.4194,
      "step": 7587
    },
    {
      "epoch": 0.9823607469980904,
      "grad_norm": 2.4786217471919127,
      "learning_rate": 1.18031048606769e-09,
      "loss": 2.8074,
      "step": 7588
    },
    {
      "epoch": 0.9824902094054438,
      "grad_norm": 3.7031030228112605,
      "learning_rate": 1.1630221587489304e-09,
      "loss": 2.2295,
      "step": 7589
    },
    {
      "epoch": 0.9826196718127974,
      "grad_norm": 2.26698201879527,
      "learning_rate": 1.1458612627775201e-09,
      "loss": 2.2422,
      "step": 7590
    },
    {
      "epoch": 0.9826196718127974,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950283765792847,
      "eval_runtime": 14.7278,
      "eval_samples_per_second": 2.988,
      "eval_steps_per_second": 0.204,
      "step": 7590
    },
    {
      "epoch": 0.9827491342201509,
      "grad_norm": 1.8635644563124136,
      "learning_rate": 1.1288278016588493e-09,
      "loss": 2.2876,
      "step": 7591
    },
    {
      "epoch": 0.9828785966275043,
      "grad_norm": 3.6206585566509584,
      "learning_rate": 1.1119217788724268e-09,
      "loss": 2.5964,
      "step": 7592
    },
    {
      "epoch": 0.9830080590348578,
      "grad_norm": 1.868015361402047,
      "learning_rate": 1.095143197871601e-09,
      "loss": 2.5551,
      "step": 7593
    },
    {
      "epoch": 0.9831375214422112,
      "grad_norm": 1.266705226144988,
      "learning_rate": 1.078492062083769e-09,
      "loss": 2.3367,
      "step": 7594
    },
    {
      "epoch": 0.9832669838495647,
      "grad_norm": 1.8780513790793911,
      "learning_rate": 1.0619683749102384e-09,
      "loss": 2.2266,
      "step": 7595
    },
    {
      "epoch": 0.9832669838495647,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4957386255264282,
      "eval_runtime": 14.1659,
      "eval_samples_per_second": 3.106,
      "eval_steps_per_second": 0.212,
      "step": 7595
    },
    {
      "epoch": 0.9833964462569181,
      "grad_norm": 2.314347414973597,
      "learning_rate": 1.045572139726364e-09,
      "loss": 2.3955,
      "step": 7596
    },
    {
      "epoch": 0.9835259086642716,
      "grad_norm": 2.6351190170336465,
      "learning_rate": 1.029303359881273e-09,
      "loss": 2.2913,
      "step": 7597
    },
    {
      "epoch": 0.983655371071625,
      "grad_norm": 1.7546872803255078,
      "learning_rate": 1.0131620386983482e-09,
      "loss": 2.4309,
      "step": 7598
    },
    {
      "epoch": 0.9837848334789785,
      "grad_norm": 2.3215434065106426,
      "learning_rate": 9.971481794746745e-10,
      "loss": 2.1787,
      "step": 7599
    },
    {
      "epoch": 0.9839142958863321,
      "grad_norm": 1.2198424416141735,
      "learning_rate": 9.812617854814549e-10,
      "loss": 2.3247,
      "step": 7600
    },
    {
      "epoch": 0.9839142958863321,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950283765792847,
      "eval_runtime": 14.5217,
      "eval_samples_per_second": 3.03,
      "eval_steps_per_second": 0.207,
      "step": 7600
    },
    {
      "epoch": 0.9840437582936855,
      "grad_norm": 1.519794772751305,
      "learning_rate": 9.655028599638017e-10,
      "loss": 2.5454,
      "step": 7601
    },
    {
      "epoch": 0.984173220701039,
      "grad_norm": 1.2600965157647652,
      "learning_rate": 9.498714061408065e-10,
      "loss": 2.21,
      "step": 7602
    },
    {
      "epoch": 0.9843026831083924,
      "grad_norm": 2.838040649791305,
      "learning_rate": 9.343674272054709e-10,
      "loss": 2.4016,
      "step": 7603
    },
    {
      "epoch": 0.9844321455157459,
      "grad_norm": 1.6408515614253465,
      "learning_rate": 9.189909263248448e-10,
      "loss": 2.3752,
      "step": 7604
    },
    {
      "epoch": 0.9845616079230993,
      "grad_norm": 2.1381858784410843,
      "learning_rate": 9.037419066398878e-10,
      "loss": 2.5369,
      "step": 7605
    },
    {
      "epoch": 0.9845616079230993,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4979802370071411,
      "eval_runtime": 13.3075,
      "eval_samples_per_second": 3.306,
      "eval_steps_per_second": 0.225,
      "step": 7605
    },
    {
      "epoch": 0.9846910703304528,
      "grad_norm": 1.5820801596214173,
      "learning_rate": 8.886203712654695e-10,
      "loss": 2.5869,
      "step": 7606
    },
    {
      "epoch": 0.9848205327378062,
      "grad_norm": 3.27107927874882,
      "learning_rate": 8.736263232905079e-10,
      "loss": 2.4082,
      "step": 7607
    },
    {
      "epoch": 0.9849499951451597,
      "grad_norm": 1.6374668708983278,
      "learning_rate": 8.587597657778307e-10,
      "loss": 2.4065,
      "step": 7608
    },
    {
      "epoch": 0.9850794575525131,
      "grad_norm": 1.727336438934298,
      "learning_rate": 8.440207017641755e-10,
      "loss": 2.5813,
      "step": 7609
    },
    {
      "epoch": 0.9852089199598667,
      "grad_norm": 1.6646088684556826,
      "learning_rate": 8.294091342604671e-10,
      "loss": 2.4685,
      "step": 7610
    },
    {
      "epoch": 0.9852089199598667,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.4666,
      "eval_samples_per_second": 3.041,
      "eval_steps_per_second": 0.207,
      "step": 7610
    },
    {
      "epoch": 0.9853383823672202,
      "grad_norm": 1.6183130409645539,
      "learning_rate": 8.149250662512626e-10,
      "loss": 2.2349,
      "step": 7611
    },
    {
      "epoch": 0.9854678447745736,
      "grad_norm": 1.9724083536469152,
      "learning_rate": 8.005685006952368e-10,
      "loss": 2.4695,
      "step": 7612
    },
    {
      "epoch": 0.9855973071819271,
      "grad_norm": 1.3922509704379329,
      "learning_rate": 7.863394405251135e-10,
      "loss": 2.374,
      "step": 7613
    },
    {
      "epoch": 0.9857267695892805,
      "grad_norm": 2.1223017834335094,
      "learning_rate": 7.722378886473176e-10,
      "loss": 2.1777,
      "step": 7614
    },
    {
      "epoch": 0.985856231996634,
      "grad_norm": 2.718200331729673,
      "learning_rate": 7.582638479424614e-10,
      "loss": 2.4619,
      "step": 7615
    },
    {
      "epoch": 0.985856231996634,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.493607997894287,
      "eval_runtime": 15.0409,
      "eval_samples_per_second": 2.925,
      "eval_steps_per_second": 0.199,
      "step": 7615
    },
    {
      "epoch": 0.9859856944039874,
      "grad_norm": 1.8789229257617968,
      "learning_rate": 7.444173212650673e-10,
      "loss": 2.3359,
      "step": 7616
    },
    {
      "epoch": 0.9861151568113409,
      "grad_norm": 2.58106500301192,
      "learning_rate": 7.306983114434979e-10,
      "loss": 2.5012,
      "step": 7617
    },
    {
      "epoch": 0.9862446192186943,
      "grad_norm": 1.7565557791717235,
      "learning_rate": 7.171068212801641e-10,
      "loss": 2.3738,
      "step": 7618
    },
    {
      "epoch": 0.9863740816260478,
      "grad_norm": 2.554639240543607,
      "learning_rate": 7.036428535513867e-10,
      "loss": 2.5596,
      "step": 7619
    },
    {
      "epoch": 0.9865035440334013,
      "grad_norm": 2.2603757514179184,
      "learning_rate": 6.903064110074659e-10,
      "loss": 2.4717,
      "step": 7620
    },
    {
      "epoch": 0.9865035440334013,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978693723678589,
      "eval_runtime": 14.3112,
      "eval_samples_per_second": 3.075,
      "eval_steps_per_second": 0.21,
      "step": 7620
    },
    {
      "epoch": 0.9866330064407548,
      "grad_norm": 1.848935015750131,
      "learning_rate": 6.770974963726807e-10,
      "loss": 2.4548,
      "step": 7621
    },
    {
      "epoch": 0.9867624688481083,
      "grad_norm": 1.4956095265091809,
      "learning_rate": 6.640161123451505e-10,
      "loss": 2.1816,
      "step": 7622
    },
    {
      "epoch": 0.9868919312554617,
      "grad_norm": 1.630094864183574,
      "learning_rate": 6.510622615970436e-10,
      "loss": 2.3577,
      "step": 7623
    },
    {
      "epoch": 0.9870213936628152,
      "grad_norm": 3.1326216222556083,
      "learning_rate": 6.382359467744375e-10,
      "loss": 2.3452,
      "step": 7624
    },
    {
      "epoch": 0.9871508560701686,
      "grad_norm": 2.2543400031695144,
      "learning_rate": 6.255371704973895e-10,
      "loss": 2.4229,
      "step": 7625
    },
    {
      "epoch": 0.9871508560701686,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950283765792847,
      "eval_runtime": 15.7043,
      "eval_samples_per_second": 2.802,
      "eval_steps_per_second": 0.191,
      "step": 7625
    },
    {
      "epoch": 0.9872803184775221,
      "grad_norm": 1.4943666856649598,
      "learning_rate": 6.129659353597967e-10,
      "loss": 2.377,
      "step": 7626
    },
    {
      "epoch": 0.9874097808848755,
      "grad_norm": 1.6970512279219263,
      "learning_rate": 6.005222439297437e-10,
      "loss": 2.458,
      "step": 7627
    },
    {
      "epoch": 0.987539243292229,
      "grad_norm": 1.360629472544963,
      "learning_rate": 5.882060987490168e-10,
      "loss": 2.2747,
      "step": 7628
    },
    {
      "epoch": 0.9876687056995825,
      "grad_norm": 1.947585735123565,
      "learning_rate": 5.760175023334507e-10,
      "loss": 2.4478,
      "step": 7629
    },
    {
      "epoch": 0.987798168106936,
      "grad_norm": 1.506017763428392,
      "learning_rate": 5.639564571727902e-10,
      "loss": 2.4949,
      "step": 7630
    },
    {
      "epoch": 0.987798168106936,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4949840307235718,
      "eval_runtime": 14.2747,
      "eval_samples_per_second": 3.082,
      "eval_steps_per_second": 0.21,
      "step": 7630
    },
    {
      "epoch": 0.9879276305142894,
      "grad_norm": 2.42281583678435,
      "learning_rate": 5.520229657307585e-10,
      "loss": 2.3638,
      "step": 7631
    },
    {
      "epoch": 0.9880570929216429,
      "grad_norm": 1.5200091251742374,
      "learning_rate": 5.402170304449894e-10,
      "loss": 2.3445,
      "step": 7632
    },
    {
      "epoch": 0.9881865553289964,
      "grad_norm": 3.56315214653513,
      "learning_rate": 5.285386537271647e-10,
      "loss": 2.4712,
      "step": 7633
    },
    {
      "epoch": 0.9883160177363498,
      "grad_norm": 1.4993946555982482,
      "learning_rate": 5.169878379628069e-10,
      "loss": 2.3345,
      "step": 7634
    },
    {
      "epoch": 0.9884454801437033,
      "grad_norm": 1.3766723430853567,
      "learning_rate": 5.055645855112784e-10,
      "loss": 2.2922,
      "step": 7635
    },
    {
      "epoch": 0.9884454801437033,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4942737817764282,
      "eval_runtime": 13.9902,
      "eval_samples_per_second": 3.145,
      "eval_steps_per_second": 0.214,
      "step": 7635
    },
    {
      "epoch": 0.9885749425510567,
      "grad_norm": 2.765105962883385,
      "learning_rate": 4.942688987061989e-10,
      "loss": 2.948,
      "step": 7636
    },
    {
      "epoch": 0.9887044049584102,
      "grad_norm": 1.7884727220792906,
      "learning_rate": 4.831007798548199e-10,
      "loss": 2.4609,
      "step": 7637
    },
    {
      "epoch": 0.9888338673657636,
      "grad_norm": 2.585771279521546,
      "learning_rate": 4.720602312385109e-10,
      "loss": 2.498,
      "step": 7638
    },
    {
      "epoch": 0.9889633297731172,
      "grad_norm": 1.6577037565047226,
      "learning_rate": 4.611472551124818e-10,
      "loss": 2.323,
      "step": 7639
    },
    {
      "epoch": 0.9890927921804706,
      "grad_norm": 2.742440349149222,
      "learning_rate": 4.5036185370592176e-10,
      "loss": 2.7051,
      "step": 7640
    },
    {
      "epoch": 0.9890927921804706,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 14.2938,
      "eval_samples_per_second": 3.078,
      "eval_steps_per_second": 0.21,
      "step": 7640
    },
    {
      "epoch": 0.9892222545878241,
      "grad_norm": 1.899570315318355,
      "learning_rate": 4.3970402922199895e-10,
      "loss": 2.3279,
      "step": 7641
    },
    {
      "epoch": 0.9893517169951775,
      "grad_norm": 1.429668343151256,
      "learning_rate": 4.291737838378607e-10,
      "loss": 2.2539,
      "step": 7642
    },
    {
      "epoch": 0.989481179402531,
      "grad_norm": 1.666726346466479,
      "learning_rate": 4.187711197043559e-10,
      "loss": 2.2537,
      "step": 7643
    },
    {
      "epoch": 0.9896106418098845,
      "grad_norm": 1.7583614698230214,
      "learning_rate": 4.0849603894659027e-10,
      "loss": 2.3398,
      "step": 7644
    },
    {
      "epoch": 0.9897401042172379,
      "grad_norm": 1.218907723659834,
      "learning_rate": 3.9834854366344044e-10,
      "loss": 2.1571,
      "step": 7645
    },
    {
      "epoch": 0.9897401042172379,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964932203292847,
      "eval_runtime": 13.6451,
      "eval_samples_per_second": 3.225,
      "eval_steps_per_second": 0.22,
      "step": 7645
    },
    {
      "epoch": 0.9898695666245914,
      "grad_norm": 2.16683363153041,
      "learning_rate": 3.8832863592762336e-10,
      "loss": 2.4543,
      "step": 7646
    },
    {
      "epoch": 0.9899990290319448,
      "grad_norm": 1.8880776975893787,
      "learning_rate": 3.7843631778604337e-10,
      "loss": 2.314,
      "step": 7647
    },
    {
      "epoch": 0.9901284914392983,
      "grad_norm": 1.913928145061539,
      "learning_rate": 3.686715912593064e-10,
      "loss": 2.385,
      "step": 7648
    },
    {
      "epoch": 0.9902579538466518,
      "grad_norm": 2.598615507596037,
      "learning_rate": 3.5903445834213625e-10,
      "loss": 2.4478,
      "step": 7649
    },
    {
      "epoch": 0.9903874162540053,
      "grad_norm": 2.781411612118827,
      "learning_rate": 3.4952492100302773e-10,
      "loss": 2.5112,
      "step": 7650
    },
    {
      "epoch": 0.9903874162540053,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4964488744735718,
      "eval_runtime": 14.049,
      "eval_samples_per_second": 3.132,
      "eval_steps_per_second": 0.214,
      "step": 7650
    },
    {
      "epoch": 0.9905168786613587,
      "grad_norm": 4.920852837984811,
      "learning_rate": 3.4014298118452415e-10,
      "loss": 2.8179,
      "step": 7651
    },
    {
      "epoch": 0.9906463410687122,
      "grad_norm": 1.602683531452493,
      "learning_rate": 3.308886408031481e-10,
      "loss": 2.4128,
      "step": 7652
    },
    {
      "epoch": 0.9907758034760656,
      "grad_norm": 2.1060923661609112,
      "learning_rate": 3.217619017492624e-10,
      "loss": 2.374,
      "step": 7653
    },
    {
      "epoch": 0.9909052658834191,
      "grad_norm": 1.6204569374519608,
      "learning_rate": 3.127627658870702e-10,
      "loss": 2.3461,
      "step": 7654
    },
    {
      "epoch": 0.9910347282907725,
      "grad_norm": 1.1008634986922998,
      "learning_rate": 3.038912350549622e-10,
      "loss": 2.2966,
      "step": 7655
    },
    {
      "epoch": 0.9910347282907725,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950505495071411,
      "eval_runtime": 13.6045,
      "eval_samples_per_second": 3.234,
      "eval_steps_per_second": 0.221,
      "step": 7655
    },
    {
      "epoch": 0.991164190698126,
      "grad_norm": 2.1093204697180243,
      "learning_rate": 2.9514731106509985e-10,
      "loss": 2.3687,
      "step": 7656
    },
    {
      "epoch": 0.9912936531054795,
      "grad_norm": 1.9171259210801308,
      "learning_rate": 2.8653099570355447e-10,
      "loss": 2.4336,
      "step": 7657
    },
    {
      "epoch": 0.9914231155128329,
      "grad_norm": 2.3993710563557116,
      "learning_rate": 2.7804229073044584e-10,
      "loss": 2.3958,
      "step": 7658
    },
    {
      "epoch": 0.9915525779201865,
      "grad_norm": 1.2834745849996845,
      "learning_rate": 2.696811978797342e-10,
      "loss": 2.4238,
      "step": 7659
    },
    {
      "epoch": 0.9916820403275399,
      "grad_norm": 2.040530792128796,
      "learning_rate": 2.614477188593589e-10,
      "loss": 2.2217,
      "step": 7660
    },
    {
      "epoch": 0.9916820403275399,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.498579502105713,
      "eval_runtime": 13.8118,
      "eval_samples_per_second": 3.186,
      "eval_steps_per_second": 0.217,
      "step": 7660
    },
    {
      "epoch": 0.9918115027348934,
      "grad_norm": 1.146395381273546,
      "learning_rate": 2.5334185535109967e-10,
      "loss": 2.1646,
      "step": 7661
    },
    {
      "epoch": 0.9919409651422468,
      "grad_norm": 1.1207392650082522,
      "learning_rate": 2.453636090108541e-10,
      "loss": 2.3267,
      "step": 7662
    },
    {
      "epoch": 0.9920704275496003,
      "grad_norm": 1.749398980875084,
      "learning_rate": 2.375129814682908e-10,
      "loss": 2.3125,
      "step": 7663
    },
    {
      "epoch": 0.9921998899569537,
      "grad_norm": 1.4514255061424015,
      "learning_rate": 2.2978997432705764e-10,
      "loss": 2.2827,
      "step": 7664
    },
    {
      "epoch": 0.9923293523643072,
      "grad_norm": 1.3744947085311505,
      "learning_rate": 2.2219458916471205e-10,
      "loss": 2.1985,
      "step": 7665
    },
    {
      "epoch": 0.9923293523643072,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971369504928589,
      "eval_runtime": 13.8034,
      "eval_samples_per_second": 3.188,
      "eval_steps_per_second": 0.217,
      "step": 7665
    },
    {
      "epoch": 0.9924588147716606,
      "grad_norm": 1.9129455404788904,
      "learning_rate": 2.1472682753279073e-10,
      "loss": 2.4282,
      "step": 7666
    },
    {
      "epoch": 0.9925882771790141,
      "grad_norm": 1.6119870892933774,
      "learning_rate": 2.0738669095680952e-10,
      "loss": 2.4197,
      "step": 7667
    },
    {
      "epoch": 0.9927177395863676,
      "grad_norm": 2.025417941338564,
      "learning_rate": 2.0017418093598583e-10,
      "loss": 2.5481,
      "step": 7668
    },
    {
      "epoch": 0.9928472019937211,
      "grad_norm": 1.7189407225275821,
      "learning_rate": 1.9308929894372441e-10,
      "loss": 2.249,
      "step": 7669
    },
    {
      "epoch": 0.9929766644010746,
      "grad_norm": 2.6274370671024756,
      "learning_rate": 1.8613204642720096e-10,
      "loss": 2.5369,
      "step": 7670
    },
    {
      "epoch": 0.9929766644010746,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986461400985718,
      "eval_runtime": 14.0005,
      "eval_samples_per_second": 3.143,
      "eval_steps_per_second": 0.214,
      "step": 7670
    },
    {
      "epoch": 0.993106126808428,
      "grad_norm": 1.733854755515088,
      "learning_rate": 1.793024248076397e-10,
      "loss": 2.4133,
      "step": 7671
    },
    {
      "epoch": 0.9932355892157815,
      "grad_norm": 3.5557057390425304,
      "learning_rate": 1.726004354801053e-10,
      "loss": 2.5322,
      "step": 7672
    },
    {
      "epoch": 0.9933650516231349,
      "grad_norm": 2.5351565687700592,
      "learning_rate": 1.660260798136415e-10,
      "loss": 2.5747,
      "step": 7673
    },
    {
      "epoch": 0.9934945140304884,
      "grad_norm": 2.155793966215189,
      "learning_rate": 1.5957935915113242e-10,
      "loss": 2.3379,
      "step": 7674
    },
    {
      "epoch": 0.9936239764378418,
      "grad_norm": 2.561628615143274,
      "learning_rate": 1.532602748094414e-10,
      "loss": 2.3025,
      "step": 7675
    },
    {
      "epoch": 0.9936239764378418,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4943625926971436,
      "eval_runtime": 14.2777,
      "eval_samples_per_second": 3.082,
      "eval_steps_per_second": 0.21,
      "step": 7675
    },
    {
      "epoch": 0.9937534388451953,
      "grad_norm": 1.4586702756577856,
      "learning_rate": 1.470688280794108e-10,
      "loss": 2.4033,
      "step": 7676
    },
    {
      "epoch": 0.9938829012525487,
      "grad_norm": 1.7218032271015382,
      "learning_rate": 1.410050202257929e-10,
      "loss": 2.3477,
      "step": 7677
    },
    {
      "epoch": 0.9940123636599023,
      "grad_norm": 2.066099568411916,
      "learning_rate": 1.3506885248718016e-10,
      "loss": 2.2959,
      "step": 7678
    },
    {
      "epoch": 0.9941418260672558,
      "grad_norm": 2.098283026570332,
      "learning_rate": 1.2926032607621375e-10,
      "loss": 2.5974,
      "step": 7679
    },
    {
      "epoch": 0.9942712884746092,
      "grad_norm": 1.6535605307389196,
      "learning_rate": 1.2357944217937513e-10,
      "loss": 2.1594,
      "step": 7680
    },
    {
      "epoch": 0.9942712884746092,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.49853515625,
      "eval_runtime": 13.6992,
      "eval_samples_per_second": 3.212,
      "eval_steps_per_second": 0.219,
      "step": 7680
    },
    {
      "epoch": 0.9944007508819627,
      "grad_norm": 1.6459518108342641,
      "learning_rate": 1.1802620195712487e-10,
      "loss": 2.2615,
      "step": 7681
    },
    {
      "epoch": 0.9945302132893161,
      "grad_norm": 2.1204057298785104,
      "learning_rate": 1.1260060654376403e-10,
      "loss": 2.3254,
      "step": 7682
    },
    {
      "epoch": 0.9946596756966696,
      "grad_norm": 1.4357120737136115,
      "learning_rate": 1.0730265704764208e-10,
      "loss": 2.2004,
      "step": 7683
    },
    {
      "epoch": 0.994789138104023,
      "grad_norm": 1.3728528906654596,
      "learning_rate": 1.0213235455094895e-10,
      "loss": 2.1951,
      "step": 7684
    },
    {
      "epoch": 0.9949186005113765,
      "grad_norm": 3.21900535857529,
      "learning_rate": 9.708970010978424e-11,
      "loss": 2.3359,
      "step": 7685
    },
    {
      "epoch": 0.9949186005113765,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4950062036514282,
      "eval_runtime": 14.3312,
      "eval_samples_per_second": 3.07,
      "eval_steps_per_second": 0.209,
      "step": 7685
    },
    {
      "epoch": 0.9950480629187299,
      "grad_norm": 1.6686792950383713,
      "learning_rate": 9.217469475429619e-11,
      "loss": 2.2742,
      "step": 7686
    },
    {
      "epoch": 0.9951775253260834,
      "grad_norm": 2.2413072500408857,
      "learning_rate": 8.738733948840395e-11,
      "loss": 2.1741,
      "step": 7687
    },
    {
      "epoch": 0.995306987733437,
      "grad_norm": 2.9388080891161135,
      "learning_rate": 8.272763529007521e-11,
      "loss": 2.5015,
      "step": 7688
    },
    {
      "epoch": 0.9954364501407904,
      "grad_norm": 1.2448282419855086,
      "learning_rate": 7.819558311111808e-11,
      "loss": 2.158,
      "step": 7689
    },
    {
      "epoch": 0.9955659125481439,
      "grad_norm": 1.5770981587173023,
      "learning_rate": 7.379118387731977e-11,
      "loss": 2.605,
      "step": 7690
    },
    {
      "epoch": 0.9955659125481439,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4972034692764282,
      "eval_runtime": 14.7955,
      "eval_samples_per_second": 2.974,
      "eval_steps_per_second": 0.203,
      "step": 7690
    },
    {
      "epoch": 0.9956953749554973,
      "grad_norm": 2.2313159470985133,
      "learning_rate": 6.951443848830785e-11,
      "loss": 2.2695,
      "step": 7691
    },
    {
      "epoch": 0.9958248373628508,
      "grad_norm": 1.583978575410211,
      "learning_rate": 6.53653478177585e-11,
      "loss": 2.5193,
      "step": 7692
    },
    {
      "epoch": 0.9959542997702042,
      "grad_norm": 2.6329864704078796,
      "learning_rate": 6.134391271318818e-11,
      "loss": 2.1763,
      "step": 7693
    },
    {
      "epoch": 0.9960837621775577,
      "grad_norm": 1.9048592700851108,
      "learning_rate": 5.745013399609256e-11,
      "loss": 2.2927,
      "step": 7694
    },
    {
      "epoch": 0.9962132245849111,
      "grad_norm": 1.2076390899334741,
      "learning_rate": 5.3684012461738265e-11,
      "loss": 2.2676,
      "step": 7695
    },
    {
      "epoch": 0.9962132245849111,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4951171875,
      "eval_runtime": 15.159,
      "eval_samples_per_second": 2.903,
      "eval_steps_per_second": 0.198,
      "step": 7695
    },
    {
      "epoch": 0.9963426869922646,
      "grad_norm": 1.6253165998282486,
      "learning_rate": 5.004554887957924e-11,
      "loss": 2.2419,
      "step": 7696
    },
    {
      "epoch": 0.996472149399618,
      "grad_norm": 2.151738149711087,
      "learning_rate": 4.6534743992701635e-11,
      "loss": 2.6377,
      "step": 7697
    },
    {
      "epoch": 0.9966016118069716,
      "grad_norm": 1.9143121114234054,
      "learning_rate": 4.315159851837891e-11,
      "loss": 2.3135,
      "step": 7698
    },
    {
      "epoch": 0.996731074214325,
      "grad_norm": 1.5023196597769006,
      "learning_rate": 3.989611314765551e-11,
      "loss": 2.3765,
      "step": 7699
    },
    {
      "epoch": 0.9968605366216785,
      "grad_norm": 1.9325526285591823,
      "learning_rate": 3.6768288545485644e-11,
      "loss": 2.4561,
      "step": 7700
    },
    {
      "epoch": 0.9968605366216785,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4986239671707153,
      "eval_runtime": 14.5617,
      "eval_samples_per_second": 3.022,
      "eval_steps_per_second": 0.206,
      "step": 7700
    },
    {
      "epoch": 0.996989999029032,
      "grad_norm": 1.4172984725227213,
      "learning_rate": 3.376812535080265e-11,
      "loss": 2.373,
      "step": 7701
    },
    {
      "epoch": 0.9971194614363854,
      "grad_norm": 1.3930326525241075,
      "learning_rate": 3.0895624176519036e-11,
      "loss": 2.3347,
      "step": 7702
    },
    {
      "epoch": 0.9972489238437389,
      "grad_norm": 1.3253972058225258,
      "learning_rate": 2.8150785609318276e-11,
      "loss": 2.4019,
      "step": 7703
    },
    {
      "epoch": 0.9973783862510923,
      "grad_norm": 1.5408348561147462,
      "learning_rate": 2.5533610209932392e-11,
      "loss": 2.3169,
      "step": 7704
    },
    {
      "epoch": 0.9975078486584458,
      "grad_norm": 2.3013034684937415,
      "learning_rate": 2.3044098512933766e-11,
      "loss": 2.5208,
      "step": 7705
    },
    {
      "epoch": 0.9975078486584458,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4978249073028564,
      "eval_runtime": 13.1607,
      "eval_samples_per_second": 3.343,
      "eval_steps_per_second": 0.228,
      "step": 7705
    },
    {
      "epoch": 0.9976373110657992,
      "grad_norm": 2.259931800473903,
      "learning_rate": 2.068225102694332e-11,
      "loss": 2.1882,
      "step": 7706
    },
    {
      "epoch": 0.9977667734731527,
      "grad_norm": 1.7890624569047229,
      "learning_rate": 1.8448068234283578e-11,
      "loss": 2.4595,
      "step": 7707
    },
    {
      "epoch": 0.9978962358805062,
      "grad_norm": 1.2401958555293557,
      "learning_rate": 1.634155059146436e-11,
      "loss": 2.2754,
      "step": 7708
    },
    {
      "epoch": 0.9980256982878597,
      "grad_norm": 1.0822891791340588,
      "learning_rate": 1.4362698528697094e-11,
      "loss": 2.4175,
      "step": 7709
    },
    {
      "epoch": 0.9981551606952132,
      "grad_norm": 1.6127689542819217,
      "learning_rate": 1.2511512450241737e-11,
      "loss": 2.4133,
      "step": 7710
    },
    {
      "epoch": 0.9981551606952132,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.497114658355713,
      "eval_runtime": 14.9295,
      "eval_samples_per_second": 2.947,
      "eval_steps_per_second": 0.201,
      "step": 7710
    },
    {
      "epoch": 0.9982846231025666,
      "grad_norm": 1.7945901352090199,
      "learning_rate": 1.0787992734198616e-11,
      "loss": 2.4961,
      "step": 7711
    },
    {
      "epoch": 0.9984140855099201,
      "grad_norm": 1.3928437748878686,
      "learning_rate": 9.192139732716598e-12,
      "loss": 2.4438,
      "step": 7712
    },
    {
      "epoch": 0.9985435479172735,
      "grad_norm": 2.9852499683419422,
      "learning_rate": 7.723953771715531e-12,
      "loss": 2.3274,
      "step": 7713
    },
    {
      "epoch": 0.998673010324627,
      "grad_norm": 2.150419645821146,
      "learning_rate": 6.383435151094408e-12,
      "loss": 2.4382,
      "step": 7714
    },
    {
      "epoch": 0.9988024727319804,
      "grad_norm": 1.657145315275181,
      "learning_rate": 5.170584144661983e-12,
      "loss": 2.5605,
      "step": 7715
    },
    {
      "epoch": 0.9988024727319804,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971591234207153,
      "eval_runtime": 15.0228,
      "eval_samples_per_second": 2.929,
      "eval_steps_per_second": 0.2,
      "step": 7715
    },
    {
      "epoch": 0.9989319351393339,
      "grad_norm": 1.4614906798288771,
      "learning_rate": 4.0854010002755466e-12,
      "loss": 2.2092,
      "step": 7716
    },
    {
      "epoch": 0.9990613975466874,
      "grad_norm": 2.668030707241662,
      "learning_rate": 3.12788593949398e-12,
      "loss": 2.4106,
      "step": 7717
    },
    {
      "epoch": 0.9991908599540409,
      "grad_norm": 1.4888698890193617,
      "learning_rate": 2.298039157994092e-12,
      "loss": 2.3706,
      "step": 7718
    },
    {
      "epoch": 0.9993203223613943,
      "grad_norm": 1.558004894188135,
      "learning_rate": 1.5958608251542828e-12,
      "loss": 2.3206,
      "step": 7719
    },
    {
      "epoch": 0.9994497847687478,
      "grad_norm": 2.4306575905372134,
      "learning_rate": 1.0213510846096564e-12,
      "loss": 2.4265,
      "step": 7720
    },
    {
      "epoch": 0.9994497847687478,
      "eval_PRM Accuracy": 0.2358490566037736,
      "eval_PRM F1": 0.047058823529411764,
      "eval_PRM F1 AUC": 0.5120481927710844,
      "eval_PRM F1 Neg": 0.36220472440944884,
      "eval_PRM NPV": 0.22115384615384615,
      "eval_PRM Precision": 1.0,
      "eval_PRM Recall": 0.024096385542168676,
      "eval_PRM Specificty": 1.0,
      "eval_loss": 1.4971591234207153,
      "eval_runtime": 14.2046,
      "eval_samples_per_second": 3.098,
      "eval_steps_per_second": 0.211,
      "step": 7720
    },
    {
      "epoch": 0.9995792471761012,
      "grad_norm": 2.458064036563779,
      "learning_rate": 5.745100535581305e-13,
      "loss": 2.3542,
      "step": 7721
    },
    {
      "epoch": 0.9997087095834547,
      "grad_norm": 2.295762242918334,
      "learning_rate": 2.5533782331554903e-13,
      "loss": 2.5386,
      "step": 7722
    },
    {
      "epoch": 0.9998381719908082,
      "grad_norm": 2.5005497572826556,
      "learning_rate": 6.383445910751463e-14,
      "loss": 2.3679,
      "step": 7723
    },
    {
      "epoch": 0.9999676343981616,
      "grad_norm": 2.2599512852855326,
      "learning_rate": 0.0,
      "loss": 2.4456,
      "step": 7724
    }
  ],
  "logging_steps": 1,
  "max_steps": 7724,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 6493605880496128.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}