{
  "best_metric": 0.9184013322231473,
  "best_model_checkpoint": "bit-50-Glaucoma\\checkpoint-3718",
  "epoch": 44.0,
  "global_step": 3718,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12,
      "learning_rate": 1.1904761904761904e-06,
      "loss": 0.8071,
      "step": 10
    },
    {
      "epoch": 0.24,
      "learning_rate": 2.3809523809523808e-06,
      "loss": 0.7134,
      "step": 20
    },
    {
      "epoch": 0.36,
      "learning_rate": 3.5714285714285714e-06,
      "loss": 0.647,
      "step": 30
    },
    {
      "epoch": 0.47,
      "learning_rate": 4.7619047619047615e-06,
      "loss": 0.547,
      "step": 40
    },
    {
      "epoch": 0.59,
      "learning_rate": 5.9523809523809525e-06,
      "loss": 0.519,
      "step": 50
    },
    {
      "epoch": 0.71,
      "learning_rate": 7.142857142857143e-06,
      "loss": 0.4948,
      "step": 60
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.333333333333334e-06,
      "loss": 0.4692,
      "step": 70
    },
    {
      "epoch": 0.95,
      "learning_rate": 9.523809523809523e-06,
      "loss": 0.4444,
      "step": 80
    },
    {
      "epoch": 0.99,
      "eval_accuracy": 0.813488759367194,
      "eval_loss": 0.4316416382789612,
      "eval_runtime": 16.626,
      "eval_samples_per_second": 72.236,
      "eval_steps_per_second": 2.286,
      "step": 84
    },
    {
      "epoch": 1.07,
      "learning_rate": 1.0714285714285714e-05,
      "loss": 0.4251,
      "step": 90
    },
    {
      "epoch": 1.18,
      "learning_rate": 1.1904761904761905e-05,
      "loss": 0.4638,
      "step": 100
    },
    {
      "epoch": 1.3,
      "learning_rate": 1.3095238095238096e-05,
      "loss": 0.4005,
      "step": 110
    },
    {
      "epoch": 1.42,
      "learning_rate": 1.4285714285714285e-05,
      "loss": 0.4005,
      "step": 120
    },
    {
      "epoch": 1.54,
      "learning_rate": 1.5476190476190476e-05,
      "loss": 0.4023,
      "step": 130
    },
    {
      "epoch": 1.66,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.3837,
      "step": 140
    },
    {
      "epoch": 1.78,
      "learning_rate": 1.785714285714286e-05,
      "loss": 0.3927,
      "step": 150
    },
    {
      "epoch": 1.89,
      "learning_rate": 1.9047619047619046e-05,
      "loss": 0.3947,
      "step": 160
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8592839300582847,
      "eval_loss": 0.31854528188705444,
      "eval_runtime": 15.8188,
      "eval_samples_per_second": 75.922,
      "eval_steps_per_second": 2.402,
      "step": 169
    },
    {
      "epoch": 2.01,
      "learning_rate": 2.023809523809524e-05,
      "loss": 0.3676,
      "step": 170
    },
    {
      "epoch": 2.13,
      "learning_rate": 2.1428571428571428e-05,
      "loss": 0.3767,
      "step": 180
    },
    {
      "epoch": 2.25,
      "learning_rate": 2.261904761904762e-05,
      "loss": 0.385,
      "step": 190
    },
    {
      "epoch": 2.37,
      "learning_rate": 2.380952380952381e-05,
      "loss": 0.3561,
      "step": 200
    },
    {
      "epoch": 2.49,
      "learning_rate": 2.5e-05,
      "loss": 0.356,
      "step": 210
    },
    {
      "epoch": 2.6,
      "learning_rate": 2.6190476190476192e-05,
      "loss": 0.3683,
      "step": 220
    },
    {
      "epoch": 2.72,
      "learning_rate": 2.7380952380952383e-05,
      "loss": 0.3506,
      "step": 230
    },
    {
      "epoch": 2.84,
      "learning_rate": 2.857142857142857e-05,
      "loss": 0.3569,
      "step": 240
    },
    {
      "epoch": 2.96,
      "learning_rate": 2.9761904761904762e-05,
      "loss": 0.3475,
      "step": 250
    },
    {
      "epoch": 2.99,
      "eval_accuracy": 0.8759367194004996,
      "eval_loss": 0.31600221991539,
      "eval_runtime": 16.0329,
      "eval_samples_per_second": 74.908,
      "eval_steps_per_second": 2.37,
      "step": 253
    },
    {
      "epoch": 3.08,
      "learning_rate": 3.095238095238095e-05,
      "loss": 0.3372,
      "step": 260
    },
    {
      "epoch": 3.2,
      "learning_rate": 3.2142857142857144e-05,
      "loss": 0.3394,
      "step": 270
    },
    {
      "epoch": 3.31,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.3077,
      "step": 280
    },
    {
      "epoch": 3.43,
      "learning_rate": 3.4523809523809526e-05,
      "loss": 0.3221,
      "step": 290
    },
    {
      "epoch": 3.55,
      "learning_rate": 3.571428571428572e-05,
      "loss": 0.3581,
      "step": 300
    },
    {
      "epoch": 3.67,
      "learning_rate": 3.690476190476191e-05,
      "loss": 0.3742,
      "step": 310
    },
    {
      "epoch": 3.79,
      "learning_rate": 3.809523809523809e-05,
      "loss": 0.377,
      "step": 320
    },
    {
      "epoch": 3.91,
      "learning_rate": 3.928571428571429e-05,
      "loss": 0.3796,
      "step": 330
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8834304746044963,
      "eval_loss": 0.29224836826324463,
      "eval_runtime": 15.8421,
      "eval_samples_per_second": 75.811,
      "eval_steps_per_second": 2.399,
      "step": 338
    },
    {
      "epoch": 4.02,
      "learning_rate": 4.047619047619048e-05,
      "loss": 0.3687,
      "step": 340
    },
    {
      "epoch": 4.14,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.3326,
      "step": 350
    },
    {
      "epoch": 4.26,
      "learning_rate": 4.2857142857142856e-05,
      "loss": 0.3275,
      "step": 360
    },
    {
      "epoch": 4.38,
      "learning_rate": 4.404761904761905e-05,
      "loss": 0.3296,
      "step": 370
    },
    {
      "epoch": 4.5,
      "learning_rate": 4.523809523809524e-05,
      "loss": 0.3257,
      "step": 380
    },
    {
      "epoch": 4.62,
      "learning_rate": 4.642857142857143e-05,
      "loss": 0.3362,
      "step": 390
    },
    {
      "epoch": 4.73,
      "learning_rate": 4.761904761904762e-05,
      "loss": 0.3471,
      "step": 400
    },
    {
      "epoch": 4.85,
      "learning_rate": 4.880952380952381e-05,
      "loss": 0.3511,
      "step": 410
    },
    {
      "epoch": 4.97,
      "learning_rate": 5e-05,
      "loss": 0.3618,
      "step": 420
    },
    {
      "epoch": 4.99,
      "eval_accuracy": 0.8859283930058285,
      "eval_loss": 0.27965524792671204,
      "eval_runtime": 15.8956,
      "eval_samples_per_second": 75.555,
      "eval_steps_per_second": 2.391,
      "step": 422
    },
    {
      "epoch": 5.09,
      "learning_rate": 4.986772486772487e-05,
      "loss": 0.3145,
      "step": 430
    },
    {
      "epoch": 5.21,
      "learning_rate": 4.973544973544973e-05,
      "loss": 0.3305,
      "step": 440
    },
    {
      "epoch": 5.33,
      "learning_rate": 4.960317460317461e-05,
      "loss": 0.3104,
      "step": 450
    },
    {
      "epoch": 5.44,
      "learning_rate": 4.9470899470899475e-05,
      "loss": 0.3015,
      "step": 460
    },
    {
      "epoch": 5.56,
      "learning_rate": 4.933862433862434e-05,
      "loss": 0.3249,
      "step": 470
    },
    {
      "epoch": 5.68,
      "learning_rate": 4.9206349206349204e-05,
      "loss": 0.2987,
      "step": 480
    },
    {
      "epoch": 5.8,
      "learning_rate": 4.9074074074074075e-05,
      "loss": 0.3434,
      "step": 490
    },
    {
      "epoch": 5.92,
      "learning_rate": 4.894179894179895e-05,
      "loss": 0.3205,
      "step": 500
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8409658617818485,
      "eval_loss": 0.3533245027065277,
      "eval_runtime": 15.8631,
      "eval_samples_per_second": 75.71,
      "eval_steps_per_second": 2.395,
      "step": 507
    },
    {
      "epoch": 6.04,
      "learning_rate": 4.880952380952381e-05,
      "loss": 0.3518,
      "step": 510
    },
    {
      "epoch": 6.15,
      "learning_rate": 4.8677248677248676e-05,
      "loss": 0.2728,
      "step": 520
    },
    {
      "epoch": 6.27,
      "learning_rate": 4.854497354497355e-05,
      "loss": 0.3404,
      "step": 530
    },
    {
      "epoch": 6.39,
      "learning_rate": 4.841269841269841e-05,
      "loss": 0.3298,
      "step": 540
    },
    {
      "epoch": 6.51,
      "learning_rate": 4.8280423280423284e-05,
      "loss": 0.3178,
      "step": 550
    },
    {
      "epoch": 6.63,
      "learning_rate": 4.814814814814815e-05,
      "loss": 0.3299,
      "step": 560
    },
    {
      "epoch": 6.75,
      "learning_rate": 4.801587301587302e-05,
      "loss": 0.3369,
      "step": 570
    },
    {
      "epoch": 6.86,
      "learning_rate": 4.7883597883597884e-05,
      "loss": 0.2974,
      "step": 580
    },
    {
      "epoch": 6.98,
      "learning_rate": 4.7751322751322756e-05,
      "loss": 0.2743,
      "step": 590
    },
    {
      "epoch": 6.99,
      "eval_accuracy": 0.8601165695253955,
      "eval_loss": 0.2907889485359192,
      "eval_runtime": 15.8108,
      "eval_samples_per_second": 75.961,
      "eval_steps_per_second": 2.403,
      "step": 591
    },
    {
      "epoch": 7.1,
      "learning_rate": 4.761904761904762e-05,
      "loss": 0.2966,
      "step": 600
    },
    {
      "epoch": 7.22,
      "learning_rate": 4.748677248677249e-05,
      "loss": 0.3083,
      "step": 610
    },
    {
      "epoch": 7.34,
      "learning_rate": 4.7354497354497356e-05,
      "loss": 0.299,
      "step": 620
    },
    {
      "epoch": 7.46,
      "learning_rate": 4.722222222222222e-05,
      "loss": 0.2669,
      "step": 630
    },
    {
      "epoch": 7.57,
      "learning_rate": 4.708994708994709e-05,
      "loss": 0.2996,
      "step": 640
    },
    {
      "epoch": 7.69,
      "learning_rate": 4.6957671957671964e-05,
      "loss": 0.2686,
      "step": 650
    },
    {
      "epoch": 7.81,
      "learning_rate": 4.682539682539683e-05,
      "loss": 0.3172,
      "step": 660
    },
    {
      "epoch": 7.93,
      "learning_rate": 4.669312169312169e-05,
      "loss": 0.2765,
      "step": 670
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8601165695253955,
      "eval_loss": 0.2945786714553833,
      "eval_runtime": 16.0316,
      "eval_samples_per_second": 74.914,
      "eval_steps_per_second": 2.37,
      "step": 676
    },
    {
      "epoch": 8.05,
      "learning_rate": 4.656084656084656e-05,
      "loss": 0.2692,
      "step": 680
    },
    {
      "epoch": 8.17,
      "learning_rate": 4.642857142857143e-05,
      "loss": 0.273,
      "step": 690
    },
    {
      "epoch": 8.28,
      "learning_rate": 4.62962962962963e-05,
      "loss": 0.252,
      "step": 700
    },
    {
      "epoch": 8.4,
      "learning_rate": 4.6164021164021165e-05,
      "loss": 0.2719,
      "step": 710
    },
    {
      "epoch": 8.52,
      "learning_rate": 4.603174603174603e-05,
      "loss": 0.273,
      "step": 720
    },
    {
      "epoch": 8.64,
      "learning_rate": 4.58994708994709e-05,
      "loss": 0.2853,
      "step": 730
    },
    {
      "epoch": 8.76,
      "learning_rate": 4.576719576719577e-05,
      "loss": 0.2965,
      "step": 740
    },
    {
      "epoch": 8.88,
      "learning_rate": 4.563492063492064e-05,
      "loss": 0.2842,
      "step": 750
    },
    {
      "epoch": 8.99,
      "learning_rate": 4.55026455026455e-05,
      "loss": 0.2528,
      "step": 760
    },
    {
      "epoch": 8.99,
      "eval_accuracy": 0.8992506244796004,
      "eval_loss": 0.24736784398555756,
      "eval_runtime": 15.802,
      "eval_samples_per_second": 76.003,
      "eval_steps_per_second": 2.405,
      "step": 760
    },
    {
      "epoch": 9.11,
      "learning_rate": 4.5370370370370374e-05,
      "loss": 0.2585,
      "step": 770
    },
    {
      "epoch": 9.23,
      "learning_rate": 4.523809523809524e-05,
      "loss": 0.2567,
      "step": 780
    },
    {
      "epoch": 9.35,
      "learning_rate": 4.510582010582011e-05,
      "loss": 0.2794,
      "step": 790
    },
    {
      "epoch": 9.47,
      "learning_rate": 4.4973544973544974e-05,
      "loss": 0.2698,
      "step": 800
    },
    {
      "epoch": 9.59,
      "learning_rate": 4.4841269841269846e-05,
      "loss": 0.2472,
      "step": 810
    },
    {
      "epoch": 9.7,
      "learning_rate": 4.470899470899471e-05,
      "loss": 0.2537,
      "step": 820
    },
    {
      "epoch": 9.82,
      "learning_rate": 4.4576719576719575e-05,
      "loss": 0.2635,
      "step": 830
    },
    {
      "epoch": 9.94,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 0.2678,
      "step": 840
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9084096586178185,
      "eval_loss": 0.2332490235567093,
      "eval_runtime": 15.7703,
      "eval_samples_per_second": 76.156,
      "eval_steps_per_second": 2.41,
      "step": 845
    },
    {
      "epoch": 10.06,
      "learning_rate": 4.431216931216932e-05,
      "loss": 0.2777,
      "step": 850
    },
    {
      "epoch": 10.18,
      "learning_rate": 4.417989417989418e-05,
      "loss": 0.2543,
      "step": 860
    },
    {
      "epoch": 10.3,
      "learning_rate": 4.404761904761905e-05,
      "loss": 0.2434,
      "step": 870
    },
    {
      "epoch": 10.41,
      "learning_rate": 4.391534391534391e-05,
      "loss": 0.2629,
      "step": 880
    },
    {
      "epoch": 10.53,
      "learning_rate": 4.378306878306879e-05,
      "loss": 0.2453,
      "step": 890
    },
    {
      "epoch": 10.65,
      "learning_rate": 4.3650793650793655e-05,
      "loss": 0.266,
      "step": 900
    },
    {
      "epoch": 10.77,
      "learning_rate": 4.351851851851852e-05,
      "loss": 0.2349,
      "step": 910
    },
    {
      "epoch": 10.89,
      "learning_rate": 4.3386243386243384e-05,
      "loss": 0.2314,
      "step": 920
    },
    {
      "epoch": 10.99,
      "eval_accuracy": 0.8992506244796004,
      "eval_loss": 0.2373119592666626,
      "eval_runtime": 15.6414,
      "eval_samples_per_second": 76.783,
      "eval_steps_per_second": 2.429,
      "step": 929
    },
    {
      "epoch": 11.01,
      "learning_rate": 4.3253968253968256e-05,
      "loss": 0.2436,
      "step": 930
    },
    {
      "epoch": 11.12,
      "learning_rate": 4.312169312169313e-05,
      "loss": 0.2576,
      "step": 940
    },
    {
      "epoch": 11.24,
      "learning_rate": 4.298941798941799e-05,
      "loss": 0.2982,
      "step": 950
    },
    {
      "epoch": 11.36,
      "learning_rate": 4.2857142857142856e-05,
      "loss": 0.2399,
      "step": 960
    },
    {
      "epoch": 11.48,
      "learning_rate": 4.272486772486773e-05,
      "loss": 0.2442,
      "step": 970
    },
    {
      "epoch": 11.6,
      "learning_rate": 4.259259259259259e-05,
      "loss": 0.2629,
      "step": 980
    },
    {
      "epoch": 11.72,
      "learning_rate": 4.2460317460317464e-05,
      "loss": 0.2484,
      "step": 990
    },
    {
      "epoch": 11.83,
      "learning_rate": 4.232804232804233e-05,
      "loss": 0.2462,
      "step": 1000
    },
    {
      "epoch": 11.95,
      "learning_rate": 4.21957671957672e-05,
      "loss": 0.244,
      "step": 1010
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9034138218151541,
      "eval_loss": 0.2284502238035202,
      "eval_runtime": 15.8599,
      "eval_samples_per_second": 75.726,
      "eval_steps_per_second": 2.396,
      "step": 1014
    },
    {
      "epoch": 12.07,
      "learning_rate": 4.2063492063492065e-05,
      "loss": 0.2496,
      "step": 1020
    },
    {
      "epoch": 12.19,
      "learning_rate": 4.193121693121693e-05,
      "loss": 0.2235,
      "step": 1030
    },
    {
      "epoch": 12.31,
      "learning_rate": 4.17989417989418e-05,
      "loss": 0.2194,
      "step": 1040
    },
    {
      "epoch": 12.43,
      "learning_rate": 4.166666666666667e-05,
      "loss": 0.2358,
      "step": 1050
    },
    {
      "epoch": 12.54,
      "learning_rate": 4.153439153439154e-05,
      "loss": 0.2447,
      "step": 1060
    },
    {
      "epoch": 12.66,
      "learning_rate": 4.14021164021164e-05,
      "loss": 0.2537,
      "step": 1070
    },
    {
      "epoch": 12.78,
      "learning_rate": 4.126984126984127e-05,
      "loss": 0.2246,
      "step": 1080
    },
    {
      "epoch": 12.9,
      "learning_rate": 4.1137566137566144e-05,
      "loss": 0.2313,
      "step": 1090
    },
    {
      "epoch": 12.99,
      "eval_accuracy": 0.9017485428809325,
      "eval_loss": 0.2355087250471115,
      "eval_runtime": 15.8512,
      "eval_samples_per_second": 75.767,
      "eval_steps_per_second": 2.397,
      "step": 1098
    },
    {
      "epoch": 13.02,
      "learning_rate": 4.100529100529101e-05,
      "loss": 0.2217,
      "step": 1100
    },
    {
      "epoch": 13.14,
      "learning_rate": 4.0873015873015874e-05,
      "loss": 0.196,
      "step": 1110
    },
    {
      "epoch": 13.25,
      "learning_rate": 4.074074074074074e-05,
      "loss": 0.2298,
      "step": 1120
    },
    {
      "epoch": 13.37,
      "learning_rate": 4.060846560846561e-05,
      "loss": 0.2352,
      "step": 1130
    },
    {
      "epoch": 13.49,
      "learning_rate": 4.047619047619048e-05,
      "loss": 0.2582,
      "step": 1140
    },
    {
      "epoch": 13.61,
      "learning_rate": 4.0343915343915346e-05,
      "loss": 0.2168,
      "step": 1150
    },
    {
      "epoch": 13.73,
      "learning_rate": 4.021164021164021e-05,
      "loss": 0.2278,
      "step": 1160
    },
    {
      "epoch": 13.85,
      "learning_rate": 4.007936507936508e-05,
      "loss": 0.2511,
      "step": 1170
    },
    {
      "epoch": 13.96,
      "learning_rate": 3.9947089947089946e-05,
      "loss": 0.2055,
      "step": 1180
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.8975853455453788,
      "eval_loss": 0.2440473437309265,
      "eval_runtime": 15.5821,
      "eval_samples_per_second": 77.075,
      "eval_steps_per_second": 2.439,
      "step": 1183
    },
    {
      "epoch": 14.08,
      "learning_rate": 3.981481481481482e-05,
      "loss": 0.2228,
      "step": 1190
    },
    {
      "epoch": 14.2,
      "learning_rate": 3.968253968253968e-05,
      "loss": 0.2118,
      "step": 1200
    },
    {
      "epoch": 14.32,
      "learning_rate": 3.9550264550264554e-05,
      "loss": 0.2281,
      "step": 1210
    },
    {
      "epoch": 14.44,
      "learning_rate": 3.941798941798942e-05,
      "loss": 0.2107,
      "step": 1220
    },
    {
      "epoch": 14.56,
      "learning_rate": 3.928571428571429e-05,
      "loss": 0.2133,
      "step": 1230
    },
    {
      "epoch": 14.67,
      "learning_rate": 3.9153439153439155e-05,
      "loss": 0.2271,
      "step": 1240
    },
    {
      "epoch": 14.79,
      "learning_rate": 3.9021164021164026e-05,
      "loss": 0.2252,
      "step": 1250
    },
    {
      "epoch": 14.91,
      "learning_rate": 3.888888888888889e-05,
      "loss": 0.2271,
      "step": 1260
    },
    {
      "epoch": 14.99,
      "eval_accuracy": 0.8884263114071607,
      "eval_loss": 0.2461342066526413,
      "eval_runtime": 15.57,
      "eval_samples_per_second": 77.136,
      "eval_steps_per_second": 2.441,
      "step": 1267
    },
    {
      "epoch": 15.03,
      "learning_rate": 3.8756613756613755e-05,
      "loss": 0.223,
      "step": 1270
    },
    {
      "epoch": 15.15,
      "learning_rate": 3.862433862433863e-05,
      "loss": 0.1974,
      "step": 1280
    },
    {
      "epoch": 15.27,
      "learning_rate": 3.84920634920635e-05,
      "loss": 0.2514,
      "step": 1290
    },
    {
      "epoch": 15.38,
      "learning_rate": 3.835978835978836e-05,
      "loss": 0.2005,
      "step": 1300
    },
    {
      "epoch": 15.5,
      "learning_rate": 3.822751322751323e-05,
      "loss": 0.2014,
      "step": 1310
    },
    {
      "epoch": 15.62,
      "learning_rate": 3.809523809523809e-05,
      "loss": 0.1952,
      "step": 1320
    },
    {
      "epoch": 15.74,
      "learning_rate": 3.7962962962962964e-05,
      "loss": 0.2002,
      "step": 1330
    },
    {
      "epoch": 15.86,
      "learning_rate": 3.7830687830687835e-05,
      "loss": 0.2351,
      "step": 1340
    },
    {
      "epoch": 15.98,
      "learning_rate": 3.76984126984127e-05,
      "loss": 0.2106,
      "step": 1350
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.8784346378018318,
      "eval_loss": 0.26777663826942444,
      "eval_runtime": 15.6876,
      "eval_samples_per_second": 76.557,
      "eval_steps_per_second": 2.422,
      "step": 1352
    },
    {
      "epoch": 16.09,
      "learning_rate": 3.7566137566137564e-05,
      "loss": 0.2066,
      "step": 1360
    },
    {
      "epoch": 16.21,
      "learning_rate": 3.7433862433862436e-05,
      "loss": 0.2009,
      "step": 1370
    },
    {
      "epoch": 16.33,
      "learning_rate": 3.730158730158731e-05,
      "loss": 0.2162,
      "step": 1380
    },
    {
      "epoch": 16.45,
      "learning_rate": 3.716931216931217e-05,
      "loss": 0.2155,
      "step": 1390
    },
    {
      "epoch": 16.57,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.1995,
      "step": 1400
    },
    {
      "epoch": 16.69,
      "learning_rate": 3.690476190476191e-05,
      "loss": 0.1992,
      "step": 1410
    },
    {
      "epoch": 16.8,
      "learning_rate": 3.677248677248677e-05,
      "loss": 0.2366,
      "step": 1420
    },
    {
      "epoch": 16.92,
      "learning_rate": 3.6640211640211644e-05,
      "loss": 0.2076,
      "step": 1430
    },
    {
      "epoch": 16.99,
      "eval_accuracy": 0.9150707743547044,
      "eval_loss": 0.22340074181556702,
      "eval_runtime": 16.0204,
      "eval_samples_per_second": 74.967,
      "eval_steps_per_second": 2.372,
      "step": 1436
    },
    {
      "epoch": 17.04,
      "learning_rate": 3.650793650793651e-05,
      "loss": 0.2054,
      "step": 1440
    },
    {
      "epoch": 17.16,
      "learning_rate": 3.637566137566138e-05,
      "loss": 0.1922,
      "step": 1450
    },
    {
      "epoch": 17.28,
      "learning_rate": 3.6243386243386245e-05,
      "loss": 0.1768,
      "step": 1460
    },
    {
      "epoch": 17.4,
      "learning_rate": 3.611111111111111e-05,
      "loss": 0.2221,
      "step": 1470
    },
    {
      "epoch": 17.51,
      "learning_rate": 3.597883597883598e-05,
      "loss": 0.1813,
      "step": 1480
    },
    {
      "epoch": 17.63,
      "learning_rate": 3.584656084656085e-05,
      "loss": 0.2035,
      "step": 1490
    },
    {
      "epoch": 17.75,
      "learning_rate": 3.571428571428572e-05,
      "loss": 0.2015,
      "step": 1500
    },
    {
      "epoch": 17.87,
      "learning_rate": 3.558201058201058e-05,
      "loss": 0.2246,
      "step": 1510
    },
    {
      "epoch": 17.99,
      "learning_rate": 3.5449735449735446e-05,
      "loss": 0.2074,
      "step": 1520
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.8934221482098251,
      "eval_loss": 0.24407817423343658,
      "eval_runtime": 15.7214,
      "eval_samples_per_second": 76.393,
      "eval_steps_per_second": 2.417,
      "step": 1521
    },
    {
      "epoch": 18.11,
      "learning_rate": 3.5317460317460324e-05,
      "loss": 0.208,
      "step": 1530
    },
    {
      "epoch": 18.22,
      "learning_rate": 3.518518518518519e-05,
      "loss": 0.1872,
      "step": 1540
    },
    {
      "epoch": 18.34,
      "learning_rate": 3.5052910052910054e-05,
      "loss": 0.2111,
      "step": 1550
    },
    {
      "epoch": 18.46,
      "learning_rate": 3.492063492063492e-05,
      "loss": 0.1953,
      "step": 1560
    },
    {
      "epoch": 18.58,
      "learning_rate": 3.478835978835979e-05,
      "loss": 0.1802,
      "step": 1570
    },
    {
      "epoch": 18.7,
      "learning_rate": 3.465608465608466e-05,
      "loss": 0.2181,
      "step": 1580
    },
    {
      "epoch": 18.82,
      "learning_rate": 3.4523809523809526e-05,
      "loss": 0.2199,
      "step": 1590
    },
    {
      "epoch": 18.93,
      "learning_rate": 3.439153439153439e-05,
      "loss": 0.2043,
      "step": 1600
    },
    {
      "epoch": 18.99,
      "eval_accuracy": 0.9000832639467111,
      "eval_loss": 0.2535579800605774,
      "eval_runtime": 15.8014,
      "eval_samples_per_second": 76.006,
      "eval_steps_per_second": 2.405,
      "step": 1605
    },
    {
      "epoch": 19.05,
      "learning_rate": 3.425925925925926e-05,
      "loss": 0.1837,
      "step": 1610
    },
    {
      "epoch": 19.17,
      "learning_rate": 3.412698412698413e-05,
      "loss": 0.1705,
      "step": 1620
    },
    {
      "epoch": 19.29,
      "learning_rate": 3.3994708994709e-05,
      "loss": 0.1925,
      "step": 1630
    },
    {
      "epoch": 19.41,
      "learning_rate": 3.386243386243386e-05,
      "loss": 0.1857,
      "step": 1640
    },
    {
      "epoch": 19.53,
      "learning_rate": 3.3730158730158734e-05,
      "loss": 0.181,
      "step": 1650
    },
    {
      "epoch": 19.64,
      "learning_rate": 3.35978835978836e-05,
      "loss": 0.2066,
      "step": 1660
    },
    {
      "epoch": 19.76,
      "learning_rate": 3.3465608465608464e-05,
      "loss": 0.2078,
      "step": 1670
    },
    {
      "epoch": 19.88,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.2028,
      "step": 1680
    },
    {
      "epoch": 20.0,
      "learning_rate": 3.3201058201058206e-05,
      "loss": 0.2058,
      "step": 1690
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.8834304746044963,
      "eval_loss": 0.2481454610824585,
      "eval_runtime": 16.3173,
      "eval_samples_per_second": 73.603,
      "eval_steps_per_second": 2.329,
      "step": 1690
    },
    {
      "epoch": 20.12,
      "learning_rate": 3.306878306878307e-05,
      "loss": 0.1715,
      "step": 1700
    },
    {
      "epoch": 20.24,
      "learning_rate": 3.2936507936507936e-05,
      "loss": 0.1735,
      "step": 1710
    },
    {
      "epoch": 20.36,
      "learning_rate": 3.280423280423281e-05,
      "loss": 0.1703,
      "step": 1720
    },
    {
      "epoch": 20.47,
      "learning_rate": 3.267195767195768e-05,
      "loss": 0.1718,
      "step": 1730
    },
    {
      "epoch": 20.59,
      "learning_rate": 3.253968253968254e-05,
      "loss": 0.1817,
      "step": 1740
    },
    {
      "epoch": 20.71,
      "learning_rate": 3.240740740740741e-05,
      "loss": 0.1959,
      "step": 1750
    },
    {
      "epoch": 20.83,
      "learning_rate": 3.227513227513227e-05,
      "loss": 0.1753,
      "step": 1760
    },
    {
      "epoch": 20.95,
      "learning_rate": 3.2142857142857144e-05,
      "loss": 0.1575,
      "step": 1770
    },
    {
      "epoch": 20.99,
      "eval_accuracy": 0.9025811823480433,
      "eval_loss": 0.2368643879890442,
      "eval_runtime": 15.885,
      "eval_samples_per_second": 75.606,
      "eval_steps_per_second": 2.392,
      "step": 1774
    },
    {
      "epoch": 21.07,
      "learning_rate": 3.2010582010582015e-05,
      "loss": 0.1754,
      "step": 1780
    },
    {
      "epoch": 21.18,
      "learning_rate": 3.187830687830688e-05,
      "loss": 0.168,
      "step": 1790
    },
    {
      "epoch": 21.3,
      "learning_rate": 3.1746031746031745e-05,
      "loss": 0.1746,
      "step": 1800
    },
    {
      "epoch": 21.42,
      "learning_rate": 3.1613756613756616e-05,
      "loss": 0.1883,
      "step": 1810
    },
    {
      "epoch": 21.54,
      "learning_rate": 3.148148148148148e-05,
      "loss": 0.1732,
      "step": 1820
    },
    {
      "epoch": 21.66,
      "learning_rate": 3.134920634920635e-05,
      "loss": 0.147,
      "step": 1830
    },
    {
      "epoch": 21.78,
      "learning_rate": 3.121693121693122e-05,
      "loss": 0.1961,
      "step": 1840
    },
    {
      "epoch": 21.89,
      "learning_rate": 3.108465608465609e-05,
      "loss": 0.1889,
      "step": 1850
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.9117402164862615,
      "eval_loss": 0.2264406383037567,
      "eval_runtime": 15.8707,
      "eval_samples_per_second": 75.674,
      "eval_steps_per_second": 2.394,
      "step": 1859
    },
    {
      "epoch": 22.01,
      "learning_rate": 3.095238095238095e-05,
      "loss": 0.1844,
      "step": 1860
    },
    {
      "epoch": 22.13,
      "learning_rate": 3.0820105820105824e-05,
      "loss": 0.175,
      "step": 1870
    },
    {
      "epoch": 22.25,
      "learning_rate": 3.068783068783069e-05,
      "loss": 0.1638,
      "step": 1880
    },
    {
      "epoch": 22.37,
      "learning_rate": 3.055555555555556e-05,
      "loss": 0.1828,
      "step": 1890
    },
    {
      "epoch": 22.49,
      "learning_rate": 3.0423280423280425e-05,
      "loss": 0.1869,
      "step": 1900
    },
    {
      "epoch": 22.6,
      "learning_rate": 3.0291005291005293e-05,
      "loss": 0.1354,
      "step": 1910
    },
    {
      "epoch": 22.72,
      "learning_rate": 3.0158730158730158e-05,
      "loss": 0.1544,
      "step": 1920
    },
    {
      "epoch": 22.84,
      "learning_rate": 3.002645502645503e-05,
      "loss": 0.1607,
      "step": 1930
    },
    {
      "epoch": 22.96,
      "learning_rate": 2.9894179894179897e-05,
      "loss": 0.1797,
      "step": 1940
    },
    {
      "epoch": 22.99,
      "eval_accuracy": 0.9084096586178185,
      "eval_loss": 0.23159563541412354,
      "eval_runtime": 15.8807,
      "eval_samples_per_second": 75.626,
      "eval_steps_per_second": 2.393,
      "step": 1943
    },
    {
      "epoch": 23.08,
      "learning_rate": 2.9761904761904762e-05,
      "loss": 0.164,
      "step": 1950
    },
    {
      "epoch": 23.2,
      "learning_rate": 2.962962962962963e-05,
      "loss": 0.1444,
      "step": 1960
    },
    {
      "epoch": 23.31,
      "learning_rate": 2.94973544973545e-05,
      "loss": 0.1518,
      "step": 1970
    },
    {
      "epoch": 23.43,
      "learning_rate": 2.9365079365079366e-05,
      "loss": 0.1498,
      "step": 1980
    },
    {
      "epoch": 23.55,
      "learning_rate": 2.9232804232804234e-05,
      "loss": 0.1624,
      "step": 1990
    },
    {
      "epoch": 23.67,
      "learning_rate": 2.91005291005291e-05,
      "loss": 0.1367,
      "step": 2000
    },
    {
      "epoch": 23.79,
      "learning_rate": 2.8968253968253974e-05,
      "loss": 0.1645,
      "step": 2010
    },
    {
      "epoch": 23.91,
      "learning_rate": 2.8835978835978838e-05,
      "loss": 0.1729,
      "step": 2020
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.9042464612822648,
      "eval_loss": 0.24072496592998505,
      "eval_runtime": 15.9775,
      "eval_samples_per_second": 75.168,
      "eval_steps_per_second": 2.378,
      "step": 2028
    },
    {
      "epoch": 24.02,
      "learning_rate": 2.8703703703703706e-05,
      "loss": 0.1603,
      "step": 2030
    },
    {
      "epoch": 24.14,
      "learning_rate": 2.857142857142857e-05,
      "loss": 0.1448,
      "step": 2040
    },
    {
      "epoch": 24.26,
      "learning_rate": 2.8439153439153442e-05,
      "loss": 0.1389,
      "step": 2050
    },
    {
      "epoch": 24.38,
      "learning_rate": 2.830687830687831e-05,
      "loss": 0.1612,
      "step": 2060
    },
    {
      "epoch": 24.5,
      "learning_rate": 2.8174603174603175e-05,
      "loss": 0.1429,
      "step": 2070
    },
    {
      "epoch": 24.62,
      "learning_rate": 2.8042328042328043e-05,
      "loss": 0.1498,
      "step": 2080
    },
    {
      "epoch": 24.73,
      "learning_rate": 2.7910052910052914e-05,
      "loss": 0.1699,
      "step": 2090
    },
    {
      "epoch": 24.85,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.1409,
      "step": 2100
    },
    {
      "epoch": 24.97,
      "learning_rate": 2.7645502645502647e-05,
      "loss": 0.163,
      "step": 2110
    },
    {
      "epoch": 24.99,
      "eval_accuracy": 0.9009159034138218,
      "eval_loss": 0.27579963207244873,
      "eval_runtime": 15.7181,
      "eval_samples_per_second": 76.409,
      "eval_steps_per_second": 2.418,
      "step": 2112
    },
    {
      "epoch": 25.09,
      "learning_rate": 2.7513227513227512e-05,
      "loss": 0.1668,
      "step": 2120
    },
    {
      "epoch": 25.21,
      "learning_rate": 2.7380952380952383e-05,
      "loss": 0.1529,
      "step": 2130
    },
    {
      "epoch": 25.33,
      "learning_rate": 2.724867724867725e-05,
      "loss": 0.1581,
      "step": 2140
    },
    {
      "epoch": 25.44,
      "learning_rate": 2.7116402116402116e-05,
      "loss": 0.1408,
      "step": 2150
    },
    {
      "epoch": 25.56,
      "learning_rate": 2.6984126984126984e-05,
      "loss": 0.1463,
      "step": 2160
    },
    {
      "epoch": 25.68,
      "learning_rate": 2.6851851851851855e-05,
      "loss": 0.1431,
      "step": 2170
    },
    {
      "epoch": 25.8,
      "learning_rate": 2.6719576719576723e-05,
      "loss": 0.1397,
      "step": 2180
    },
    {
      "epoch": 25.92,
      "learning_rate": 2.6587301587301588e-05,
      "loss": 0.1609,
      "step": 2190
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.8934221482098251,
      "eval_loss": 0.25153836607933044,
      "eval_runtime": 15.7973,
      "eval_samples_per_second": 76.026,
      "eval_steps_per_second": 2.405,
      "step": 2197
    },
    {
      "epoch": 26.04,
      "learning_rate": 2.6455026455026456e-05,
      "loss": 0.1503,
      "step": 2200
    },
    {
      "epoch": 26.15,
      "learning_rate": 2.6322751322751328e-05,
      "loss": 0.141,
      "step": 2210
    },
    {
      "epoch": 26.27,
      "learning_rate": 2.6190476190476192e-05,
      "loss": 0.1384,
      "step": 2220
    },
    {
      "epoch": 26.39,
      "learning_rate": 2.605820105820106e-05,
      "loss": 0.1505,
      "step": 2230
    },
    {
      "epoch": 26.51,
      "learning_rate": 2.5925925925925925e-05,
      "loss": 0.153,
      "step": 2240
    },
    {
      "epoch": 26.63,
      "learning_rate": 2.5793650793650796e-05,
      "loss": 0.1367,
      "step": 2250
    },
    {
      "epoch": 26.75,
      "learning_rate": 2.5661375661375664e-05,
      "loss": 0.1587,
      "step": 2260
    },
    {
      "epoch": 26.86,
      "learning_rate": 2.552910052910053e-05,
      "loss": 0.1493,
      "step": 2270
    },
    {
      "epoch": 26.98,
      "learning_rate": 2.5396825396825397e-05,
      "loss": 0.1345,
      "step": 2280
    },
    {
      "epoch": 26.99,
      "eval_accuracy": 0.8950874271440467,
      "eval_loss": 0.2521910071372986,
      "eval_runtime": 15.9043,
      "eval_samples_per_second": 75.514,
      "eval_steps_per_second": 2.389,
      "step": 2281
    },
    {
      "epoch": 27.1,
      "learning_rate": 2.526455026455027e-05,
      "loss": 0.1349,
      "step": 2290
    },
    {
      "epoch": 27.22,
      "learning_rate": 2.5132275132275137e-05,
      "loss": 0.1301,
      "step": 2300
    },
    {
      "epoch": 27.34,
      "learning_rate": 2.5e-05,
      "loss": 0.1382,
      "step": 2310
    },
    {
      "epoch": 27.46,
      "learning_rate": 2.4867724867724866e-05,
      "loss": 0.1336,
      "step": 2320
    },
    {
      "epoch": 27.57,
      "learning_rate": 2.4735449735449737e-05,
      "loss": 0.1308,
      "step": 2330
    },
    {
      "epoch": 27.69,
      "learning_rate": 2.4603174603174602e-05,
      "loss": 0.1285,
      "step": 2340
    },
    {
      "epoch": 27.81,
      "learning_rate": 2.4470899470899473e-05,
      "loss": 0.1424,
      "step": 2350
    },
    {
      "epoch": 27.93,
      "learning_rate": 2.4338624338624338e-05,
      "loss": 0.1567,
      "step": 2360
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.9084096586178185,
      "eval_loss": 0.24210087954998016,
      "eval_runtime": 15.8265,
      "eval_samples_per_second": 75.885,
      "eval_steps_per_second": 2.401,
      "step": 2366
    },
    {
      "epoch": 28.05,
      "learning_rate": 2.4206349206349206e-05,
      "loss": 0.135,
      "step": 2370
    },
    {
      "epoch": 28.17,
      "learning_rate": 2.4074074074074074e-05,
      "loss": 0.1305,
      "step": 2380
    },
    {
      "epoch": 28.28,
      "learning_rate": 2.3941798941798942e-05,
      "loss": 0.1351,
      "step": 2390
    },
    {
      "epoch": 28.4,
      "learning_rate": 2.380952380952381e-05,
      "loss": 0.1351,
      "step": 2400
    },
    {
      "epoch": 28.52,
      "learning_rate": 2.3677248677248678e-05,
      "loss": 0.1272,
      "step": 2410
    },
    {
      "epoch": 28.64,
      "learning_rate": 2.3544973544973546e-05,
      "loss": 0.1522,
      "step": 2420
    },
    {
      "epoch": 28.76,
      "learning_rate": 2.3412698412698414e-05,
      "loss": 0.1295,
      "step": 2430
    },
    {
      "epoch": 28.88,
      "learning_rate": 2.328042328042328e-05,
      "loss": 0.141,
      "step": 2440
    },
    {
      "epoch": 28.99,
      "learning_rate": 2.314814814814815e-05,
      "loss": 0.1441,
      "step": 2450
    },
    {
      "epoch": 28.99,
      "eval_accuracy": 0.9167360532889259,
      "eval_loss": 0.2525836229324341,
      "eval_runtime": 16.1677,
      "eval_samples_per_second": 74.284,
      "eval_steps_per_second": 2.35,
      "step": 2450
    },
    {
      "epoch": 29.11,
      "learning_rate": 2.3015873015873015e-05,
      "loss": 0.1166,
      "step": 2460
    },
    {
      "epoch": 29.23,
      "learning_rate": 2.2883597883597886e-05,
      "loss": 0.1318,
      "step": 2470
    },
    {
      "epoch": 29.35,
      "learning_rate": 2.275132275132275e-05,
      "loss": 0.1171,
      "step": 2480
    },
    {
      "epoch": 29.47,
      "learning_rate": 2.261904761904762e-05,
      "loss": 0.1276,
      "step": 2490
    },
    {
      "epoch": 29.59,
      "learning_rate": 2.2486772486772487e-05,
      "loss": 0.1442,
      "step": 2500
    },
    {
      "epoch": 29.7,
      "learning_rate": 2.2354497354497355e-05,
      "loss": 0.1522,
      "step": 2510
    },
    {
      "epoch": 29.82,
      "learning_rate": 2.2222222222222223e-05,
      "loss": 0.1357,
      "step": 2520
    },
    {
      "epoch": 29.94,
      "learning_rate": 2.208994708994709e-05,
      "loss": 0.1219,
      "step": 2530
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.9109075770191507,
      "eval_loss": 0.2636389136314392,
      "eval_runtime": 15.8974,
      "eval_samples_per_second": 75.547,
      "eval_steps_per_second": 2.39,
      "step": 2535
    },
    {
      "epoch": 30.06,
      "learning_rate": 2.1957671957671956e-05,
      "loss": 0.1387,
      "step": 2540
    },
    {
      "epoch": 30.18,
      "learning_rate": 2.1825396825396827e-05,
      "loss": 0.1182,
      "step": 2550
    },
    {
      "epoch": 30.3,
      "learning_rate": 2.1693121693121692e-05,
      "loss": 0.1103,
      "step": 2560
    },
    {
      "epoch": 30.41,
      "learning_rate": 2.1560846560846563e-05,
      "loss": 0.1385,
      "step": 2570
    },
    {
      "epoch": 30.53,
      "learning_rate": 2.1428571428571428e-05,
      "loss": 0.1279,
      "step": 2580
    },
    {
      "epoch": 30.65,
      "learning_rate": 2.1296296296296296e-05,
      "loss": 0.1133,
      "step": 2590
    },
    {
      "epoch": 30.77,
      "learning_rate": 2.1164021164021164e-05,
      "loss": 0.1475,
      "step": 2600
    },
    {
      "epoch": 30.89,
      "learning_rate": 2.1031746031746032e-05,
      "loss": 0.1137,
      "step": 2610
    },
    {
      "epoch": 30.99,
      "eval_accuracy": 0.9034138218151541,
      "eval_loss": 0.2889064848423004,
      "eval_runtime": 15.4125,
      "eval_samples_per_second": 77.924,
      "eval_steps_per_second": 2.466,
      "step": 2619
    },
    {
      "epoch": 31.01,
      "learning_rate": 2.08994708994709e-05,
      "loss": 0.1316,
      "step": 2620
    },
    {
      "epoch": 31.12,
      "learning_rate": 2.076719576719577e-05,
      "loss": 0.1245,
      "step": 2630
    },
    {
      "epoch": 31.24,
      "learning_rate": 2.0634920634920636e-05,
      "loss": 0.1104,
      "step": 2640
    },
    {
      "epoch": 31.36,
      "learning_rate": 2.0502645502645504e-05,
      "loss": 0.1201,
      "step": 2650
    },
    {
      "epoch": 31.48,
      "learning_rate": 2.037037037037037e-05,
      "loss": 0.1269,
      "step": 2660
    },
    {
      "epoch": 31.6,
      "learning_rate": 2.023809523809524e-05,
      "loss": 0.1034,
      "step": 2670
    },
    {
      "epoch": 31.72,
      "learning_rate": 2.0105820105820105e-05,
      "loss": 0.1039,
      "step": 2680
    },
    {
      "epoch": 31.83,
      "learning_rate": 1.9973544973544973e-05,
      "loss": 0.1297,
      "step": 2690
    },
    {
      "epoch": 31.95,
      "learning_rate": 1.984126984126984e-05,
      "loss": 0.1109,
      "step": 2700
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.8917568692756037,
      "eval_loss": 0.3240314722061157,
      "eval_runtime": 15.6837,
      "eval_samples_per_second": 76.576,
      "eval_steps_per_second": 2.423,
      "step": 2704
    },
    {
      "epoch": 32.07,
      "learning_rate": 1.970899470899471e-05,
      "loss": 0.1219,
      "step": 2710
    },
    {
      "epoch": 32.19,
      "learning_rate": 1.9576719576719577e-05,
      "loss": 0.125,
      "step": 2720
    },
    {
      "epoch": 32.31,
      "learning_rate": 1.9444444444444445e-05,
      "loss": 0.1037,
      "step": 2730
    },
    {
      "epoch": 32.43,
      "learning_rate": 1.9312169312169313e-05,
      "loss": 0.0952,
      "step": 2740
    },
    {
      "epoch": 32.54,
      "learning_rate": 1.917989417989418e-05,
      "loss": 0.1125,
      "step": 2750
    },
    {
      "epoch": 32.66,
      "learning_rate": 1.9047619047619046e-05,
      "loss": 0.1036,
      "step": 2760
    },
    {
      "epoch": 32.78,
      "learning_rate": 1.8915343915343918e-05,
      "loss": 0.1192,
      "step": 2770
    },
    {
      "epoch": 32.9,
      "learning_rate": 1.8783068783068782e-05,
      "loss": 0.1254,
      "step": 2780
    },
    {
      "epoch": 32.99,
      "eval_accuracy": 0.9150707743547044,
      "eval_loss": 0.2847664952278137,
      "eval_runtime": 15.3587,
      "eval_samples_per_second": 78.196,
      "eval_steps_per_second": 2.474,
      "step": 2788
    },
    {
      "epoch": 33.02,
      "learning_rate": 1.8650793650793654e-05,
      "loss": 0.0989,
      "step": 2790
    },
    {
      "epoch": 33.14,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.1006,
      "step": 2800
    },
    {
      "epoch": 33.25,
      "learning_rate": 1.8386243386243386e-05,
      "loss": 0.1213,
      "step": 2810
    },
    {
      "epoch": 33.37,
      "learning_rate": 1.8253968253968254e-05,
      "loss": 0.1153,
      "step": 2820
    },
    {
      "epoch": 33.49,
      "learning_rate": 1.8121693121693122e-05,
      "loss": 0.1199,
      "step": 2830
    },
    {
      "epoch": 33.61,
      "learning_rate": 1.798941798941799e-05,
      "loss": 0.1034,
      "step": 2840
    },
    {
      "epoch": 33.73,
      "learning_rate": 1.785714285714286e-05,
      "loss": 0.1013,
      "step": 2850
    },
    {
      "epoch": 33.85,
      "learning_rate": 1.7724867724867723e-05,
      "loss": 0.1197,
      "step": 2860
    },
    {
      "epoch": 33.96,
      "learning_rate": 1.7592592592592595e-05,
      "loss": 0.1425,
      "step": 2870
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.9092422980849292,
      "eval_loss": 0.27417123317718506,
      "eval_runtime": 15.4475,
      "eval_samples_per_second": 77.747,
      "eval_steps_per_second": 2.46,
      "step": 2873
    },
    {
      "epoch": 34.08,
      "learning_rate": 1.746031746031746e-05,
      "loss": 0.0986,
      "step": 2880
    },
    {
      "epoch": 34.2,
      "learning_rate": 1.732804232804233e-05,
      "loss": 0.0942,
      "step": 2890
    },
    {
      "epoch": 34.32,
      "learning_rate": 1.7195767195767195e-05,
      "loss": 0.1186,
      "step": 2900
    },
    {
      "epoch": 34.44,
      "learning_rate": 1.7063492063492063e-05,
      "loss": 0.0965,
      "step": 2910
    },
    {
      "epoch": 34.56,
      "learning_rate": 1.693121693121693e-05,
      "loss": 0.1119,
      "step": 2920
    },
    {
      "epoch": 34.67,
      "learning_rate": 1.67989417989418e-05,
      "loss": 0.1238,
      "step": 2930
    },
    {
      "epoch": 34.79,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.125,
      "step": 2940
    },
    {
      "epoch": 34.91,
      "learning_rate": 1.6534391534391536e-05,
      "loss": 0.1164,
      "step": 2950
    },
    {
      "epoch": 34.99,
      "eval_accuracy": 0.9150707743547044,
      "eval_loss": 0.3028131425380707,
      "eval_runtime": 15.5308,
      "eval_samples_per_second": 77.33,
      "eval_steps_per_second": 2.447,
      "step": 2957
    },
    {
      "epoch": 35.03,
      "learning_rate": 1.6402116402116404e-05,
      "loss": 0.0972,
      "step": 2960
    },
    {
      "epoch": 35.15,
      "learning_rate": 1.626984126984127e-05,
      "loss": 0.1096,
      "step": 2970
    },
    {
      "epoch": 35.27,
      "learning_rate": 1.6137566137566136e-05,
      "loss": 0.1163,
      "step": 2980
    },
    {
      "epoch": 35.38,
      "learning_rate": 1.6005291005291008e-05,
      "loss": 0.1034,
      "step": 2990
    },
    {
      "epoch": 35.5,
      "learning_rate": 1.5873015873015872e-05,
      "loss": 0.0962,
      "step": 3000
    },
    {
      "epoch": 35.62,
      "learning_rate": 1.574074074074074e-05,
      "loss": 0.0955,
      "step": 3010
    },
    {
      "epoch": 35.74,
      "learning_rate": 1.560846560846561e-05,
      "loss": 0.1082,
      "step": 3020
    },
    {
      "epoch": 35.86,
      "learning_rate": 1.5476190476190476e-05,
      "loss": 0.0875,
      "step": 3030
    },
    {
      "epoch": 35.98,
      "learning_rate": 1.5343915343915344e-05,
      "loss": 0.0781,
      "step": 3040
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.9059117402164862,
      "eval_loss": 0.33245834708213806,
      "eval_runtime": 15.6522,
      "eval_samples_per_second": 76.73,
      "eval_steps_per_second": 2.428,
      "step": 3042
    },
    {
      "epoch": 36.09,
      "learning_rate": 1.5211640211640213e-05,
      "loss": 0.1148,
      "step": 3050
    },
    {
      "epoch": 36.21,
      "learning_rate": 1.5079365079365079e-05,
      "loss": 0.0983,
      "step": 3060
    },
    {
      "epoch": 36.33,
      "learning_rate": 1.4947089947089949e-05,
      "loss": 0.1018,
      "step": 3070
    },
    {
      "epoch": 36.45,
      "learning_rate": 1.4814814814814815e-05,
      "loss": 0.1093,
      "step": 3080
    },
    {
      "epoch": 36.57,
      "learning_rate": 1.4682539682539683e-05,
      "loss": 0.1056,
      "step": 3090
    },
    {
      "epoch": 36.69,
      "learning_rate": 1.455026455026455e-05,
      "loss": 0.0843,
      "step": 3100
    },
    {
      "epoch": 36.8,
      "learning_rate": 1.4417989417989419e-05,
      "loss": 0.0945,
      "step": 3110
    },
    {
      "epoch": 36.92,
      "learning_rate": 1.4285714285714285e-05,
      "loss": 0.1004,
      "step": 3120
    },
    {
      "epoch": 36.99,
      "eval_accuracy": 0.9017485428809325,
      "eval_loss": 0.314680814743042,
      "eval_runtime": 15.403,
      "eval_samples_per_second": 77.972,
      "eval_steps_per_second": 2.467,
      "step": 3126
    },
    {
      "epoch": 37.04,
      "learning_rate": 1.4153439153439155e-05,
      "loss": 0.0903,
      "step": 3130
    },
    {
      "epoch": 37.16,
      "learning_rate": 1.4021164021164022e-05,
      "loss": 0.0932,
      "step": 3140
    },
    {
      "epoch": 37.28,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.0949,
      "step": 3150
    },
    {
      "epoch": 37.4,
      "learning_rate": 1.3756613756613756e-05,
      "loss": 0.0923,
      "step": 3160
    },
    {
      "epoch": 37.51,
      "learning_rate": 1.3624338624338626e-05,
      "loss": 0.0837,
      "step": 3170
    },
    {
      "epoch": 37.63,
      "learning_rate": 1.3492063492063492e-05,
      "loss": 0.1104,
      "step": 3180
    },
    {
      "epoch": 37.75,
      "learning_rate": 1.3359788359788362e-05,
      "loss": 0.1168,
      "step": 3190
    },
    {
      "epoch": 37.87,
      "learning_rate": 1.3227513227513228e-05,
      "loss": 0.098,
      "step": 3200
    },
    {
      "epoch": 37.99,
      "learning_rate": 1.3095238095238096e-05,
      "loss": 0.0955,
      "step": 3210
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.9117402164862615,
      "eval_loss": 0.31323495507240295,
      "eval_runtime": 15.436,
      "eval_samples_per_second": 77.805,
      "eval_steps_per_second": 2.462,
      "step": 3211
    },
    {
      "epoch": 38.11,
      "learning_rate": 1.2962962962962962e-05,
      "loss": 0.1021,
      "step": 3220
    },
    {
      "epoch": 38.22,
      "learning_rate": 1.2830687830687832e-05,
      "loss": 0.082,
      "step": 3230
    },
    {
      "epoch": 38.34,
      "learning_rate": 1.2698412698412699e-05,
      "loss": 0.1104,
      "step": 3240
    },
    {
      "epoch": 38.46,
      "learning_rate": 1.2566137566137568e-05,
      "loss": 0.0954,
      "step": 3250
    },
    {
      "epoch": 38.58,
      "learning_rate": 1.2433862433862433e-05,
      "loss": 0.084,
      "step": 3260
    },
    {
      "epoch": 38.7,
      "learning_rate": 1.2301587301587301e-05,
      "loss": 0.0762,
      "step": 3270
    },
    {
      "epoch": 38.82,
      "learning_rate": 1.2169312169312169e-05,
      "loss": 0.0789,
      "step": 3280
    },
    {
      "epoch": 38.93,
      "learning_rate": 1.2037037037037037e-05,
      "loss": 0.0942,
      "step": 3290
    },
    {
      "epoch": 38.99,
      "eval_accuracy": 0.9142381348875936,
      "eval_loss": 0.3243829607963562,
      "eval_runtime": 16.3904,
      "eval_samples_per_second": 73.275,
      "eval_steps_per_second": 2.318,
      "step": 3295
    },
    {
      "epoch": 39.05,
      "learning_rate": 1.1904761904761905e-05,
      "loss": 0.0888,
      "step": 3300
    },
    {
      "epoch": 39.17,
      "learning_rate": 1.1772486772486773e-05,
      "loss": 0.0895,
      "step": 3310
    },
    {
      "epoch": 39.29,
      "learning_rate": 1.164021164021164e-05,
      "loss": 0.0973,
      "step": 3320
    },
    {
      "epoch": 39.41,
      "learning_rate": 1.1507936507936508e-05,
      "loss": 0.0648,
      "step": 3330
    },
    {
      "epoch": 39.53,
      "learning_rate": 1.1375661375661376e-05,
      "loss": 0.11,
      "step": 3340
    },
    {
      "epoch": 39.64,
      "learning_rate": 1.1243386243386244e-05,
      "loss": 0.0698,
      "step": 3350
    },
    {
      "epoch": 39.76,
      "learning_rate": 1.1111111111111112e-05,
      "loss": 0.0988,
      "step": 3360
    },
    {
      "epoch": 39.88,
      "learning_rate": 1.0978835978835978e-05,
      "loss": 0.0803,
      "step": 3370
    },
    {
      "epoch": 40.0,
      "learning_rate": 1.0846560846560846e-05,
      "loss": 0.1077,
      "step": 3380
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.9134054954204829,
      "eval_loss": 0.32237449288368225,
      "eval_runtime": 15.4633,
      "eval_samples_per_second": 77.668,
      "eval_steps_per_second": 2.457,
      "step": 3380
    },
    {
      "epoch": 40.12,
      "learning_rate": 1.0714285714285714e-05,
      "loss": 0.063,
      "step": 3390
    },
    {
      "epoch": 40.24,
      "learning_rate": 1.0582010582010582e-05,
      "loss": 0.077,
      "step": 3400
    },
    {
      "epoch": 40.36,
      "learning_rate": 1.044973544973545e-05,
      "loss": 0.0976,
      "step": 3410
    },
    {
      "epoch": 40.47,
      "learning_rate": 1.0317460317460318e-05,
      "loss": 0.1123,
      "step": 3420
    },
    {
      "epoch": 40.59,
      "learning_rate": 1.0185185185185185e-05,
      "loss": 0.0968,
      "step": 3430
    },
    {
      "epoch": 40.71,
      "learning_rate": 1.0052910052910053e-05,
      "loss": 0.0808,
      "step": 3440
    },
    {
      "epoch": 40.83,
      "learning_rate": 9.92063492063492e-06,
      "loss": 0.0861,
      "step": 3450
    },
    {
      "epoch": 40.95,
      "learning_rate": 9.788359788359789e-06,
      "loss": 0.079,
      "step": 3460
    },
    {
      "epoch": 40.99,
      "eval_accuracy": 0.9134054954204829,
      "eval_loss": 0.33289557695388794,
      "eval_runtime": 15.7987,
      "eval_samples_per_second": 76.019,
      "eval_steps_per_second": 2.405,
      "step": 3464
    },
    {
      "epoch": 41.07,
      "learning_rate": 9.656084656084657e-06,
      "loss": 0.0968,
      "step": 3470
    },
    {
      "epoch": 41.18,
      "learning_rate": 9.523809523809523e-06,
      "loss": 0.0785,
      "step": 3480
    },
    {
      "epoch": 41.3,
      "learning_rate": 9.391534391534391e-06,
      "loss": 0.0711,
      "step": 3490
    },
    {
      "epoch": 41.42,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.0785,
      "step": 3500
    },
    {
      "epoch": 41.54,
      "learning_rate": 9.126984126984127e-06,
      "loss": 0.0888,
      "step": 3510
    },
    {
      "epoch": 41.66,
      "learning_rate": 8.994708994708995e-06,
      "loss": 0.0727,
      "step": 3520
    },
    {
      "epoch": 41.78,
      "learning_rate": 8.862433862433862e-06,
      "loss": 0.0806,
      "step": 3530
    },
    {
      "epoch": 41.89,
      "learning_rate": 8.73015873015873e-06,
      "loss": 0.0826,
      "step": 3540
    },
    {
      "epoch": 42.0,
      "eval_accuracy": 0.9109075770191507,
      "eval_loss": 0.33648884296417236,
      "eval_runtime": 15.7707,
      "eval_samples_per_second": 76.154,
      "eval_steps_per_second": 2.41,
      "step": 3549
    },
    {
      "epoch": 42.01,
      "learning_rate": 8.597883597883598e-06,
      "loss": 0.0678,
      "step": 3550
    },
    {
      "epoch": 42.13,
      "learning_rate": 8.465608465608466e-06,
      "loss": 0.0653,
      "step": 3560
    },
    {
      "epoch": 42.25,
      "learning_rate": 8.333333333333334e-06,
      "loss": 0.0788,
      "step": 3570
    },
    {
      "epoch": 42.37,
      "learning_rate": 8.201058201058202e-06,
      "loss": 0.0798,
      "step": 3580
    },
    {
      "epoch": 42.49,
      "learning_rate": 8.068783068783068e-06,
      "loss": 0.0832,
      "step": 3590
    },
    {
      "epoch": 42.6,
      "learning_rate": 7.936507936507936e-06,
      "loss": 0.0819,
      "step": 3600
    },
    {
      "epoch": 42.72,
      "learning_rate": 7.804232804232804e-06,
      "loss": 0.0813,
      "step": 3610
    },
    {
      "epoch": 42.84,
      "learning_rate": 7.671957671957672e-06,
      "loss": 0.0738,
      "step": 3620
    },
    {
      "epoch": 42.96,
      "learning_rate": 7.5396825396825394e-06,
      "loss": 0.07,
      "step": 3630
    },
    {
      "epoch": 42.99,
      "eval_accuracy": 0.9100749375520399,
      "eval_loss": 0.3693588972091675,
      "eval_runtime": 15.7316,
      "eval_samples_per_second": 76.343,
      "eval_steps_per_second": 2.416,
      "step": 3633
    },
    {
      "epoch": 43.08,
      "learning_rate": 7.4074074074074075e-06,
      "loss": 0.0717,
      "step": 3640
    },
    {
      "epoch": 43.2,
      "learning_rate": 7.275132275132275e-06,
      "loss": 0.0835,
      "step": 3650
    },
    {
      "epoch": 43.31,
      "learning_rate": 7.142857142857143e-06,
      "loss": 0.0765,
      "step": 3660
    },
    {
      "epoch": 43.43,
      "learning_rate": 7.010582010582011e-06,
      "loss": 0.0721,
      "step": 3670
    },
    {
      "epoch": 43.55,
      "learning_rate": 6.878306878306878e-06,
      "loss": 0.0733,
      "step": 3680
    },
    {
      "epoch": 43.67,
      "learning_rate": 6.746031746031746e-06,
      "loss": 0.0709,
      "step": 3690
    },
    {
      "epoch": 43.79,
      "learning_rate": 6.613756613756614e-06,
      "loss": 0.09,
      "step": 3700
    },
    {
      "epoch": 43.91,
      "learning_rate": 6.481481481481481e-06,
      "loss": 0.0698,
      "step": 3710
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.9184013322231473,
      "eval_loss": 0.3492058217525482,
      "eval_runtime": 15.97,
      "eval_samples_per_second": 75.204,
      "eval_steps_per_second": 2.379,
      "step": 3718
    }
  ],
  "max_steps": 4200,
  "num_train_epochs": 50,
  "total_flos": 4.035107025911808e+19,
  "trial_name": null,
  "trial_params": null
}