{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.32365601838366187,
  "eval_steps": 5,
  "global_step": 2500,
  "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
    }
  ],
  "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": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2098934106324992.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}