{
  "best_metric": 1.3439607620239258,
  "best_model_checkpoint": "./outputs_letter3/checkpoint-1736",
  "epoch": 250.0,
  "eval_steps": 500,
  "global_step": 1750,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.015384615384615385,
      "eval_loss": 3.9449453353881836,
      "eval_runtime": 2.0983,
      "eval_samples_per_second": 61.955,
      "eval_steps_per_second": 0.953,
      "step": 7
    },
    {
      "epoch": 1.43,
      "learning_rate": 1.9714285714285718e-05,
      "loss": 3.9333,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.023076923076923078,
      "eval_loss": 3.9366512298583984,
      "eval_runtime": 2.2606,
      "eval_samples_per_second": 57.506,
      "eval_steps_per_second": 0.885,
      "step": 14
    },
    {
      "epoch": 2.86,
      "learning_rate": 1.942857142857143e-05,
      "loss": 3.8939,
      "step": 20
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.03076923076923077,
      "eval_loss": 3.9279532432556152,
      "eval_runtime": 2.3275,
      "eval_samples_per_second": 55.855,
      "eval_steps_per_second": 0.859,
      "step": 21
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.046153846153846156,
      "eval_loss": 3.9167158603668213,
      "eval_runtime": 2.2361,
      "eval_samples_per_second": 58.137,
      "eval_steps_per_second": 0.894,
      "step": 28
    },
    {
      "epoch": 4.29,
      "learning_rate": 1.9142857142857146e-05,
      "loss": 3.8562,
      "step": 30
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.06923076923076923,
      "eval_loss": 3.9032633304595947,
      "eval_runtime": 2.4923,
      "eval_samples_per_second": 52.161,
      "eval_steps_per_second": 0.802,
      "step": 35
    },
    {
      "epoch": 5.71,
      "learning_rate": 1.885714285714286e-05,
      "loss": 3.8008,
      "step": 40
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.07692307692307693,
      "eval_loss": 3.8873634338378906,
      "eval_runtime": 2.5262,
      "eval_samples_per_second": 51.461,
      "eval_steps_per_second": 0.792,
      "step": 42
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.1076923076923077,
      "eval_loss": 3.867002010345459,
      "eval_runtime": 2.5898,
      "eval_samples_per_second": 50.196,
      "eval_steps_per_second": 0.772,
      "step": 49
    },
    {
      "epoch": 7.14,
      "learning_rate": 1.8571428571428575e-05,
      "loss": 3.7555,
      "step": 50
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.1,
      "eval_loss": 3.8494887351989746,
      "eval_runtime": 2.6867,
      "eval_samples_per_second": 48.387,
      "eval_steps_per_second": 0.744,
      "step": 56
    },
    {
      "epoch": 8.57,
      "learning_rate": 1.8285714285714288e-05,
      "loss": 3.6917,
      "step": 60
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.11538461538461539,
      "eval_loss": 3.8304965496063232,
      "eval_runtime": 2.4963,
      "eval_samples_per_second": 52.077,
      "eval_steps_per_second": 0.801,
      "step": 63
    },
    {
      "epoch": 10.0,
      "learning_rate": 1.8e-05,
      "loss": 3.6372,
      "step": 70
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.13846153846153847,
      "eval_loss": 3.8138267993927,
      "eval_runtime": 2.3942,
      "eval_samples_per_second": 54.298,
      "eval_steps_per_second": 0.835,
      "step": 70
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.12307692307692308,
      "eval_loss": 3.796628952026367,
      "eval_runtime": 2.3929,
      "eval_samples_per_second": 54.328,
      "eval_steps_per_second": 0.836,
      "step": 77
    },
    {
      "epoch": 11.43,
      "learning_rate": 1.7714285714285717e-05,
      "loss": 3.5846,
      "step": 80
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.15384615384615385,
      "eval_loss": 3.7766666412353516,
      "eval_runtime": 2.4605,
      "eval_samples_per_second": 52.835,
      "eval_steps_per_second": 0.813,
      "step": 84
    },
    {
      "epoch": 12.86,
      "learning_rate": 1.742857142857143e-05,
      "loss": 3.5047,
      "step": 90
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.23076923076923078,
      "eval_loss": 3.751601219177246,
      "eval_runtime": 2.5029,
      "eval_samples_per_second": 51.94,
      "eval_steps_per_second": 0.799,
      "step": 91
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.23846153846153847,
      "eval_loss": 3.7278525829315186,
      "eval_runtime": 2.6158,
      "eval_samples_per_second": 49.699,
      "eval_steps_per_second": 0.765,
      "step": 98
    },
    {
      "epoch": 14.29,
      "learning_rate": 1.7142857142857142e-05,
      "loss": 3.4547,
      "step": 100
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.23846153846153847,
      "eval_loss": 3.7031078338623047,
      "eval_runtime": 2.5457,
      "eval_samples_per_second": 51.066,
      "eval_steps_per_second": 0.786,
      "step": 105
    },
    {
      "epoch": 15.71,
      "learning_rate": 1.6857142857142858e-05,
      "loss": 3.3796,
      "step": 110
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.2692307692307692,
      "eval_loss": 3.672494649887085,
      "eval_runtime": 2.368,
      "eval_samples_per_second": 54.899,
      "eval_steps_per_second": 0.845,
      "step": 112
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.27692307692307694,
      "eval_loss": 3.646197557449341,
      "eval_runtime": 2.3752,
      "eval_samples_per_second": 54.731,
      "eval_steps_per_second": 0.842,
      "step": 119
    },
    {
      "epoch": 17.14,
      "learning_rate": 1.6571428571428574e-05,
      "loss": 3.3283,
      "step": 120
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.2923076923076923,
      "eval_loss": 3.622591495513916,
      "eval_runtime": 2.3748,
      "eval_samples_per_second": 54.741,
      "eval_steps_per_second": 0.842,
      "step": 126
    },
    {
      "epoch": 18.57,
      "learning_rate": 1.6285714285714287e-05,
      "loss": 3.2728,
      "step": 130
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.2846153846153846,
      "eval_loss": 3.6022486686706543,
      "eval_runtime": 2.4033,
      "eval_samples_per_second": 54.091,
      "eval_steps_per_second": 0.832,
      "step": 133
    },
    {
      "epoch": 20.0,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 3.2229,
      "step": 140
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.27692307692307694,
      "eval_loss": 3.5929768085479736,
      "eval_runtime": 2.4687,
      "eval_samples_per_second": 52.66,
      "eval_steps_per_second": 0.81,
      "step": 140
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.33076923076923076,
      "eval_loss": 3.574838876724243,
      "eval_runtime": 2.5603,
      "eval_samples_per_second": 50.775,
      "eval_steps_per_second": 0.781,
      "step": 147
    },
    {
      "epoch": 21.43,
      "learning_rate": 1.5714285714285715e-05,
      "loss": 3.1514,
      "step": 150
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.3384615384615385,
      "eval_loss": 3.5403976440429688,
      "eval_runtime": 2.3284,
      "eval_samples_per_second": 55.833,
      "eval_steps_per_second": 0.859,
      "step": 154
    },
    {
      "epoch": 22.86,
      "learning_rate": 1.542857142857143e-05,
      "loss": 3.1179,
      "step": 160
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.3384615384615385,
      "eval_loss": 3.514573335647583,
      "eval_runtime": 2.3901,
      "eval_samples_per_second": 54.39,
      "eval_steps_per_second": 0.837,
      "step": 161
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.34615384615384615,
      "eval_loss": 3.491649866104126,
      "eval_runtime": 2.3495,
      "eval_samples_per_second": 55.331,
      "eval_steps_per_second": 0.851,
      "step": 168
    },
    {
      "epoch": 24.29,
      "learning_rate": 1.5142857142857144e-05,
      "loss": 3.0559,
      "step": 170
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.3384615384615385,
      "eval_loss": 3.473316192626953,
      "eval_runtime": 2.387,
      "eval_samples_per_second": 54.461,
      "eval_steps_per_second": 0.838,
      "step": 175
    },
    {
      "epoch": 25.71,
      "learning_rate": 1.4857142857142858e-05,
      "loss": 3.0051,
      "step": 180
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.36153846153846153,
      "eval_loss": 3.453972339630127,
      "eval_runtime": 2.5367,
      "eval_samples_per_second": 51.249,
      "eval_steps_per_second": 0.788,
      "step": 182
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.36923076923076925,
      "eval_loss": 3.4498963356018066,
      "eval_runtime": 2.5601,
      "eval_samples_per_second": 50.779,
      "eval_steps_per_second": 0.781,
      "step": 189
    },
    {
      "epoch": 27.14,
      "learning_rate": 1.4571428571428573e-05,
      "loss": 2.9775,
      "step": 190
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.3769230769230769,
      "eval_loss": 3.435525417327881,
      "eval_runtime": 2.3663,
      "eval_samples_per_second": 54.939,
      "eval_steps_per_second": 0.845,
      "step": 196
    },
    {
      "epoch": 28.57,
      "learning_rate": 1.4285714285714287e-05,
      "loss": 2.9277,
      "step": 200
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.38461538461538464,
      "eval_loss": 3.4166250228881836,
      "eval_runtime": 2.3557,
      "eval_samples_per_second": 55.185,
      "eval_steps_per_second": 0.849,
      "step": 203
    },
    {
      "epoch": 30.0,
      "learning_rate": 1.4e-05,
      "loss": 2.9066,
      "step": 210
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.4,
      "eval_loss": 3.40069842338562,
      "eval_runtime": 2.3914,
      "eval_samples_per_second": 54.362,
      "eval_steps_per_second": 0.836,
      "step": 210
    },
    {
      "epoch": 31.0,
      "eval_accuracy": 0.36923076923076925,
      "eval_loss": 3.3825573921203613,
      "eval_runtime": 2.5239,
      "eval_samples_per_second": 51.509,
      "eval_steps_per_second": 0.792,
      "step": 217
    },
    {
      "epoch": 31.43,
      "learning_rate": 1.3714285714285716e-05,
      "loss": 2.8464,
      "step": 220
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.4076923076923077,
      "eval_loss": 3.369760274887085,
      "eval_runtime": 2.3953,
      "eval_samples_per_second": 54.273,
      "eval_steps_per_second": 0.835,
      "step": 224
    },
    {
      "epoch": 32.86,
      "learning_rate": 1.3428571428571429e-05,
      "loss": 2.8044,
      "step": 230
    },
    {
      "epoch": 33.0,
      "eval_accuracy": 0.4076923076923077,
      "eval_loss": 3.350893497467041,
      "eval_runtime": 2.5141,
      "eval_samples_per_second": 51.708,
      "eval_steps_per_second": 0.796,
      "step": 231
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.3769230769230769,
      "eval_loss": 3.324298620223999,
      "eval_runtime": 2.5107,
      "eval_samples_per_second": 51.778,
      "eval_steps_per_second": 0.797,
      "step": 238
    },
    {
      "epoch": 34.29,
      "learning_rate": 1.3142857142857145e-05,
      "loss": 2.7699,
      "step": 240
    },
    {
      "epoch": 35.0,
      "eval_accuracy": 0.3923076923076923,
      "eval_loss": 3.320146322250366,
      "eval_runtime": 2.3695,
      "eval_samples_per_second": 54.864,
      "eval_steps_per_second": 0.844,
      "step": 245
    },
    {
      "epoch": 35.71,
      "learning_rate": 1.2857142857142859e-05,
      "loss": 2.7251,
      "step": 250
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.4,
      "eval_loss": 3.301337242126465,
      "eval_runtime": 2.3948,
      "eval_samples_per_second": 54.285,
      "eval_steps_per_second": 0.835,
      "step": 252
    },
    {
      "epoch": 37.0,
      "eval_accuracy": 0.4230769230769231,
      "eval_loss": 3.2935874462127686,
      "eval_runtime": 2.3921,
      "eval_samples_per_second": 54.345,
      "eval_steps_per_second": 0.836,
      "step": 259
    },
    {
      "epoch": 37.14,
      "learning_rate": 1.2571428571428572e-05,
      "loss": 2.6915,
      "step": 260
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.45384615384615384,
      "eval_loss": 3.2827184200286865,
      "eval_runtime": 2.4835,
      "eval_samples_per_second": 52.345,
      "eval_steps_per_second": 0.805,
      "step": 266
    },
    {
      "epoch": 38.57,
      "learning_rate": 1.2285714285714288e-05,
      "loss": 2.6527,
      "step": 270
    },
    {
      "epoch": 39.0,
      "eval_accuracy": 0.46153846153846156,
      "eval_loss": 3.262709379196167,
      "eval_runtime": 2.5052,
      "eval_samples_per_second": 51.892,
      "eval_steps_per_second": 0.798,
      "step": 273
    },
    {
      "epoch": 40.0,
      "learning_rate": 1.2e-05,
      "loss": 2.6541,
      "step": 280
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.46153846153846156,
      "eval_loss": 3.2581355571746826,
      "eval_runtime": 2.5207,
      "eval_samples_per_second": 51.573,
      "eval_steps_per_second": 0.793,
      "step": 280
    },
    {
      "epoch": 41.0,
      "eval_accuracy": 0.4230769230769231,
      "eval_loss": 3.2342007160186768,
      "eval_runtime": 2.363,
      "eval_samples_per_second": 55.016,
      "eval_steps_per_second": 0.846,
      "step": 287
    },
    {
      "epoch": 41.43,
      "learning_rate": 1.1714285714285716e-05,
      "loss": 2.5968,
      "step": 290
    },
    {
      "epoch": 42.0,
      "eval_accuracy": 0.43846153846153846,
      "eval_loss": 3.2211241722106934,
      "eval_runtime": 2.5033,
      "eval_samples_per_second": 51.932,
      "eval_steps_per_second": 0.799,
      "step": 294
    },
    {
      "epoch": 42.86,
      "learning_rate": 1.1428571428571429e-05,
      "loss": 2.573,
      "step": 300
    },
    {
      "epoch": 43.0,
      "eval_accuracy": 0.4076923076923077,
      "eval_loss": 3.212242603302002,
      "eval_runtime": 2.4013,
      "eval_samples_per_second": 54.137,
      "eval_steps_per_second": 0.833,
      "step": 301
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.46153846153846156,
      "eval_loss": 3.2259228229522705,
      "eval_runtime": 2.5062,
      "eval_samples_per_second": 51.871,
      "eval_steps_per_second": 0.798,
      "step": 308
    },
    {
      "epoch": 44.29,
      "learning_rate": 1.1142857142857143e-05,
      "loss": 2.554,
      "step": 310
    },
    {
      "epoch": 45.0,
      "eval_accuracy": 0.4307692307692308,
      "eval_loss": 3.2271060943603516,
      "eval_runtime": 2.357,
      "eval_samples_per_second": 55.154,
      "eval_steps_per_second": 0.849,
      "step": 315
    },
    {
      "epoch": 45.71,
      "learning_rate": 1.0857142857142858e-05,
      "loss": 2.5222,
      "step": 320
    },
    {
      "epoch": 46.0,
      "eval_accuracy": 0.4461538461538462,
      "eval_loss": 3.2208163738250732,
      "eval_runtime": 2.7528,
      "eval_samples_per_second": 47.225,
      "eval_steps_per_second": 0.727,
      "step": 322
    },
    {
      "epoch": 47.0,
      "eval_accuracy": 0.4461538461538462,
      "eval_loss": 3.2139155864715576,
      "eval_runtime": 2.3673,
      "eval_samples_per_second": 54.915,
      "eval_steps_per_second": 0.845,
      "step": 329
    },
    {
      "epoch": 47.14,
      "learning_rate": 1.0571428571428572e-05,
      "loss": 2.5085,
      "step": 330
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.45384615384615384,
      "eval_loss": 3.204026460647583,
      "eval_runtime": 2.3343,
      "eval_samples_per_second": 55.691,
      "eval_steps_per_second": 0.857,
      "step": 336
    },
    {
      "epoch": 48.57,
      "learning_rate": 1.0285714285714285e-05,
      "loss": 2.4593,
      "step": 340
    },
    {
      "epoch": 49.0,
      "eval_accuracy": 0.49230769230769234,
      "eval_loss": 3.205333948135376,
      "eval_runtime": 2.3891,
      "eval_samples_per_second": 54.414,
      "eval_steps_per_second": 0.837,
      "step": 343
    },
    {
      "epoch": 50.0,
      "learning_rate": 1e-05,
      "loss": 2.4585,
      "step": 350
    },
    {
      "epoch": 50.0,
      "eval_accuracy": 0.47692307692307695,
      "eval_loss": 3.1822495460510254,
      "eval_runtime": 2.3305,
      "eval_samples_per_second": 55.781,
      "eval_steps_per_second": 0.858,
      "step": 350
    },
    {
      "epoch": 51.0,
      "eval_accuracy": 0.46923076923076923,
      "eval_loss": 3.1696622371673584,
      "eval_runtime": 2.3113,
      "eval_samples_per_second": 56.246,
      "eval_steps_per_second": 0.865,
      "step": 357
    },
    {
      "epoch": 51.43,
      "learning_rate": 9.714285714285715e-06,
      "loss": 2.4228,
      "step": 360
    },
    {
      "epoch": 52.0,
      "eval_accuracy": 0.46923076923076923,
      "eval_loss": 3.1589303016662598,
      "eval_runtime": 2.5283,
      "eval_samples_per_second": 51.419,
      "eval_steps_per_second": 0.791,
      "step": 364
    },
    {
      "epoch": 52.86,
      "learning_rate": 9.42857142857143e-06,
      "loss": 2.3954,
      "step": 370
    },
    {
      "epoch": 53.0,
      "eval_accuracy": 0.47692307692307695,
      "eval_loss": 3.1375324726104736,
      "eval_runtime": 2.3577,
      "eval_samples_per_second": 55.139,
      "eval_steps_per_second": 0.848,
      "step": 371
    },
    {
      "epoch": 54.0,
      "eval_accuracy": 0.45384615384615384,
      "eval_loss": 3.1092171669006348,
      "eval_runtime": 2.3799,
      "eval_samples_per_second": 54.625,
      "eval_steps_per_second": 0.84,
      "step": 378
    },
    {
      "epoch": 54.29,
      "learning_rate": 9.142857142857144e-06,
      "loss": 2.3641,
      "step": 380
    },
    {
      "epoch": 55.0,
      "eval_accuracy": 0.47692307692307695,
      "eval_loss": 3.099919319152832,
      "eval_runtime": 2.3488,
      "eval_samples_per_second": 55.348,
      "eval_steps_per_second": 0.852,
      "step": 385
    },
    {
      "epoch": 55.71,
      "learning_rate": 8.857142857142858e-06,
      "loss": 2.3651,
      "step": 390
    },
    {
      "epoch": 56.0,
      "eval_accuracy": 0.46153846153846156,
      "eval_loss": 3.0859622955322266,
      "eval_runtime": 2.5127,
      "eval_samples_per_second": 51.737,
      "eval_steps_per_second": 0.796,
      "step": 392
    },
    {
      "epoch": 57.0,
      "eval_accuracy": 0.46153846153846156,
      "eval_loss": 3.0813159942626953,
      "eval_runtime": 2.449,
      "eval_samples_per_second": 53.082,
      "eval_steps_per_second": 0.817,
      "step": 399
    },
    {
      "epoch": 57.14,
      "learning_rate": 8.571428571428571e-06,
      "loss": 2.3182,
      "step": 400
    },
    {
      "epoch": 58.0,
      "eval_accuracy": 0.49230769230769234,
      "eval_loss": 3.0692226886749268,
      "eval_runtime": 2.3556,
      "eval_samples_per_second": 55.188,
      "eval_steps_per_second": 0.849,
      "step": 406
    },
    {
      "epoch": 58.57,
      "learning_rate": 8.285714285714287e-06,
      "loss": 2.3029,
      "step": 410
    },
    {
      "epoch": 59.0,
      "eval_accuracy": 0.4846153846153846,
      "eval_loss": 3.061020612716675,
      "eval_runtime": 2.4546,
      "eval_samples_per_second": 52.961,
      "eval_steps_per_second": 0.815,
      "step": 413
    },
    {
      "epoch": 60.0,
      "learning_rate": 8.000000000000001e-06,
      "loss": 2.2988,
      "step": 420
    },
    {
      "epoch": 60.0,
      "eval_accuracy": 0.46153846153846156,
      "eval_loss": 3.0626943111419678,
      "eval_runtime": 2.3415,
      "eval_samples_per_second": 55.521,
      "eval_steps_per_second": 0.854,
      "step": 420
    },
    {
      "epoch": 61.0,
      "eval_accuracy": 0.46923076923076923,
      "eval_loss": 3.051997184753418,
      "eval_runtime": 2.492,
      "eval_samples_per_second": 52.167,
      "eval_steps_per_second": 0.803,
      "step": 427
    },
    {
      "epoch": 61.43,
      "learning_rate": 7.714285714285716e-06,
      "loss": 2.2865,
      "step": 430
    },
    {
      "epoch": 62.0,
      "eval_accuracy": 0.45384615384615384,
      "eval_loss": 3.039459705352783,
      "eval_runtime": 2.4542,
      "eval_samples_per_second": 52.97,
      "eval_steps_per_second": 0.815,
      "step": 434
    },
    {
      "epoch": 62.86,
      "learning_rate": 7.428571428571429e-06,
      "loss": 2.2623,
      "step": 440
    },
    {
      "epoch": 63.0,
      "eval_accuracy": 0.46153846153846156,
      "eval_loss": 3.0356502532958984,
      "eval_runtime": 2.3341,
      "eval_samples_per_second": 55.697,
      "eval_steps_per_second": 0.857,
      "step": 441
    },
    {
      "epoch": 64.0,
      "eval_accuracy": 0.46153846153846156,
      "eval_loss": 3.0332510471343994,
      "eval_runtime": 2.3535,
      "eval_samples_per_second": 55.236,
      "eval_steps_per_second": 0.85,
      "step": 448
    },
    {
      "epoch": 64.29,
      "learning_rate": 7.1428571428571436e-06,
      "loss": 2.2252,
      "step": 450
    },
    {
      "epoch": 65.0,
      "eval_accuracy": 0.47692307692307695,
      "eval_loss": 3.022916316986084,
      "eval_runtime": 2.5362,
      "eval_samples_per_second": 51.258,
      "eval_steps_per_second": 0.789,
      "step": 455
    },
    {
      "epoch": 65.71,
      "learning_rate": 6.857142857142858e-06,
      "loss": 2.2339,
      "step": 460
    },
    {
      "epoch": 66.0,
      "eval_accuracy": 0.47692307692307695,
      "eval_loss": 3.020322799682617,
      "eval_runtime": 2.3519,
      "eval_samples_per_second": 55.275,
      "eval_steps_per_second": 0.85,
      "step": 462
    },
    {
      "epoch": 67.0,
      "eval_accuracy": 0.49230769230769234,
      "eval_loss": 3.0076119899749756,
      "eval_runtime": 2.359,
      "eval_samples_per_second": 55.107,
      "eval_steps_per_second": 0.848,
      "step": 469
    },
    {
      "epoch": 67.14,
      "learning_rate": 6.571428571428572e-06,
      "loss": 2.2017,
      "step": 470
    },
    {
      "epoch": 68.0,
      "eval_accuracy": 0.4846153846153846,
      "eval_loss": 2.9876248836517334,
      "eval_runtime": 2.372,
      "eval_samples_per_second": 54.805,
      "eval_steps_per_second": 0.843,
      "step": 476
    },
    {
      "epoch": 68.57,
      "learning_rate": 6.285714285714286e-06,
      "loss": 2.1972,
      "step": 480
    },
    {
      "epoch": 69.0,
      "eval_accuracy": 0.49230769230769234,
      "eval_loss": 2.971572160720825,
      "eval_runtime": 2.3479,
      "eval_samples_per_second": 55.368,
      "eval_steps_per_second": 0.852,
      "step": 483
    },
    {
      "epoch": 70.0,
      "learning_rate": 6e-06,
      "loss": 2.1964,
      "step": 490
    },
    {
      "epoch": 70.0,
      "eval_accuracy": 0.5,
      "eval_loss": 2.96323823928833,
      "eval_runtime": 2.4351,
      "eval_samples_per_second": 53.387,
      "eval_steps_per_second": 0.821,
      "step": 490
    },
    {
      "epoch": 71.0,
      "eval_accuracy": 0.49230769230769234,
      "eval_loss": 2.9597153663635254,
      "eval_runtime": 2.4929,
      "eval_samples_per_second": 52.149,
      "eval_steps_per_second": 0.802,
      "step": 497
    },
    {
      "epoch": 71.43,
      "learning_rate": 5.7142857142857145e-06,
      "loss": 2.1775,
      "step": 500
    },
    {
      "epoch": 72.0,
      "eval_accuracy": 0.5,
      "eval_loss": 2.9580914974212646,
      "eval_runtime": 2.5027,
      "eval_samples_per_second": 51.943,
      "eval_steps_per_second": 0.799,
      "step": 504
    },
    {
      "epoch": 72.86,
      "learning_rate": 5.428571428571429e-06,
      "loss": 2.1619,
      "step": 510
    },
    {
      "epoch": 73.0,
      "eval_accuracy": 0.5076923076923077,
      "eval_loss": 2.9516122341156006,
      "eval_runtime": 2.4828,
      "eval_samples_per_second": 52.361,
      "eval_steps_per_second": 0.806,
      "step": 511
    },
    {
      "epoch": 74.0,
      "eval_accuracy": 0.5153846153846153,
      "eval_loss": 2.935607433319092,
      "eval_runtime": 2.3758,
      "eval_samples_per_second": 54.719,
      "eval_steps_per_second": 0.842,
      "step": 518
    },
    {
      "epoch": 74.29,
      "learning_rate": 5.142857142857142e-06,
      "loss": 2.1633,
      "step": 520
    },
    {
      "epoch": 75.0,
      "eval_accuracy": 0.5076923076923077,
      "eval_loss": 2.928561210632324,
      "eval_runtime": 2.3327,
      "eval_samples_per_second": 55.729,
      "eval_steps_per_second": 0.857,
      "step": 525
    },
    {
      "epoch": 75.71,
      "learning_rate": 4.857142857142858e-06,
      "loss": 2.1207,
      "step": 530
    },
    {
      "epoch": 76.0,
      "eval_accuracy": 0.5153846153846153,
      "eval_loss": 2.9266109466552734,
      "eval_runtime": 2.3813,
      "eval_samples_per_second": 54.591,
      "eval_steps_per_second": 0.84,
      "step": 532
    },
    {
      "epoch": 77.0,
      "eval_accuracy": 0.5230769230769231,
      "eval_loss": 2.920504093170166,
      "eval_runtime": 2.4777,
      "eval_samples_per_second": 52.468,
      "eval_steps_per_second": 0.807,
      "step": 539
    },
    {
      "epoch": 77.14,
      "learning_rate": 4.571428571428572e-06,
      "loss": 2.1353,
      "step": 540
    },
    {
      "epoch": 78.0,
      "eval_accuracy": 0.5153846153846153,
      "eval_loss": 2.9131386280059814,
      "eval_runtime": 2.3865,
      "eval_samples_per_second": 54.473,
      "eval_steps_per_second": 0.838,
      "step": 546
    },
    {
      "epoch": 78.57,
      "learning_rate": 4.2857142857142855e-06,
      "loss": 2.1075,
      "step": 550
    },
    {
      "epoch": 79.0,
      "eval_accuracy": 0.5230769230769231,
      "eval_loss": 2.907454252243042,
      "eval_runtime": 2.3713,
      "eval_samples_per_second": 54.823,
      "eval_steps_per_second": 0.843,
      "step": 553
    },
    {
      "epoch": 80.0,
      "learning_rate": 4.000000000000001e-06,
      "loss": 2.1025,
      "step": 560
    },
    {
      "epoch": 80.0,
      "eval_accuracy": 0.5230769230769231,
      "eval_loss": 2.907280921936035,
      "eval_runtime": 2.3736,
      "eval_samples_per_second": 54.768,
      "eval_steps_per_second": 0.843,
      "step": 560
    },
    {
      "epoch": 81.0,
      "eval_accuracy": 0.5153846153846153,
      "eval_loss": 2.9174439907073975,
      "eval_runtime": 2.3602,
      "eval_samples_per_second": 55.079,
      "eval_steps_per_second": 0.847,
      "step": 567
    },
    {
      "epoch": 81.43,
      "learning_rate": 3.7142857142857146e-06,
      "loss": 2.1031,
      "step": 570
    },
    {
      "epoch": 82.0,
      "eval_accuracy": 0.5307692307692308,
      "eval_loss": 2.913137197494507,
      "eval_runtime": 2.4712,
      "eval_samples_per_second": 52.605,
      "eval_steps_per_second": 0.809,
      "step": 574
    },
    {
      "epoch": 82.86,
      "learning_rate": 3.428571428571429e-06,
      "loss": 2.0932,
      "step": 580
    },
    {
      "epoch": 83.0,
      "eval_accuracy": 0.5307692307692308,
      "eval_loss": 2.909245491027832,
      "eval_runtime": 2.3787,
      "eval_samples_per_second": 54.651,
      "eval_steps_per_second": 0.841,
      "step": 581
    },
    {
      "epoch": 84.0,
      "eval_accuracy": 0.5307692307692308,
      "eval_loss": 2.8977668285369873,
      "eval_runtime": 4.0301,
      "eval_samples_per_second": 32.257,
      "eval_steps_per_second": 0.496,
      "step": 588
    },
    {
      "epoch": 84.29,
      "learning_rate": 3.142857142857143e-06,
      "loss": 2.0861,
      "step": 590
    },
    {
      "epoch": 85.0,
      "eval_accuracy": 0.5307692307692308,
      "eval_loss": 2.8871328830718994,
      "eval_runtime": 3.7091,
      "eval_samples_per_second": 35.049,
      "eval_steps_per_second": 0.539,
      "step": 595
    },
    {
      "epoch": 85.71,
      "learning_rate": 2.8571428571428573e-06,
      "loss": 2.0478,
      "step": 600
    },
    {
      "epoch": 86.0,
      "eval_accuracy": 0.5384615384615384,
      "eval_loss": 2.882897138595581,
      "eval_runtime": 4.3893,
      "eval_samples_per_second": 29.617,
      "eval_steps_per_second": 0.456,
      "step": 602
    },
    {
      "epoch": 87.0,
      "eval_accuracy": 0.5461538461538461,
      "eval_loss": 2.880449056625366,
      "eval_runtime": 2.9082,
      "eval_samples_per_second": 44.701,
      "eval_steps_per_second": 0.688,
      "step": 609
    },
    {
      "epoch": 87.14,
      "learning_rate": 2.571428571428571e-06,
      "loss": 2.0815,
      "step": 610
    },
    {
      "epoch": 88.0,
      "eval_accuracy": 0.5461538461538461,
      "eval_loss": 2.872459888458252,
      "eval_runtime": 2.8921,
      "eval_samples_per_second": 44.95,
      "eval_steps_per_second": 0.692,
      "step": 616
    },
    {
      "epoch": 88.57,
      "learning_rate": 2.285714285714286e-06,
      "loss": 2.0756,
      "step": 620
    },
    {
      "epoch": 89.0,
      "eval_accuracy": 0.5461538461538461,
      "eval_loss": 2.8694419860839844,
      "eval_runtime": 3.9643,
      "eval_samples_per_second": 32.793,
      "eval_steps_per_second": 0.505,
      "step": 623
    },
    {
      "epoch": 90.0,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 2.065,
      "step": 630
    },
    {
      "epoch": 90.0,
      "eval_accuracy": 0.5461538461538461,
      "eval_loss": 2.866473913192749,
      "eval_runtime": 3.1035,
      "eval_samples_per_second": 41.888,
      "eval_steps_per_second": 0.644,
      "step": 630
    },
    {
      "epoch": 91.0,
      "eval_accuracy": 0.5461538461538461,
      "eval_loss": 2.861470937728882,
      "eval_runtime": 3.1207,
      "eval_samples_per_second": 41.657,
      "eval_steps_per_second": 0.641,
      "step": 637
    },
    {
      "epoch": 91.43,
      "learning_rate": 1.7142857142857145e-06,
      "loss": 2.0572,
      "step": 640
    },
    {
      "epoch": 92.0,
      "eval_accuracy": 0.5461538461538461,
      "eval_loss": 2.859877347946167,
      "eval_runtime": 2.6425,
      "eval_samples_per_second": 49.197,
      "eval_steps_per_second": 0.757,
      "step": 644
    },
    {
      "epoch": 92.86,
      "learning_rate": 1.4285714285714286e-06,
      "loss": 2.0358,
      "step": 650
    },
    {
      "epoch": 93.0,
      "eval_accuracy": 0.5461538461538461,
      "eval_loss": 2.8619918823242188,
      "eval_runtime": 2.4174,
      "eval_samples_per_second": 53.776,
      "eval_steps_per_second": 0.827,
      "step": 651
    },
    {
      "epoch": 94.0,
      "eval_accuracy": 0.5461538461538461,
      "eval_loss": 2.862945079803467,
      "eval_runtime": 2.4216,
      "eval_samples_per_second": 53.684,
      "eval_steps_per_second": 0.826,
      "step": 658
    },
    {
      "epoch": 94.29,
      "learning_rate": 1.142857142857143e-06,
      "loss": 2.0663,
      "step": 660
    },
    {
      "epoch": 95.0,
      "eval_accuracy": 0.5538461538461539,
      "eval_loss": 2.862516403198242,
      "eval_runtime": 2.3749,
      "eval_samples_per_second": 54.739,
      "eval_steps_per_second": 0.842,
      "step": 665
    },
    {
      "epoch": 95.71,
      "learning_rate": 8.571428571428572e-07,
      "loss": 2.0353,
      "step": 670
    },
    {
      "epoch": 96.0,
      "eval_accuracy": 0.5538461538461539,
      "eval_loss": 2.8627583980560303,
      "eval_runtime": 2.3527,
      "eval_samples_per_second": 55.257,
      "eval_steps_per_second": 0.85,
      "step": 672
    },
    {
      "epoch": 97.0,
      "eval_accuracy": 0.5538461538461539,
      "eval_loss": 2.8628671169281006,
      "eval_runtime": 2.5748,
      "eval_samples_per_second": 50.49,
      "eval_steps_per_second": 0.777,
      "step": 679
    },
    {
      "epoch": 97.14,
      "learning_rate": 5.714285714285715e-07,
      "loss": 2.0506,
      "step": 680
    },
    {
      "epoch": 98.0,
      "eval_accuracy": 0.5538461538461539,
      "eval_loss": 2.862170457839966,
      "eval_runtime": 2.3987,
      "eval_samples_per_second": 54.196,
      "eval_steps_per_second": 0.834,
      "step": 686
    },
    {
      "epoch": 98.57,
      "learning_rate": 2.8571428571428575e-07,
      "loss": 2.0494,
      "step": 690
    },
    {
      "epoch": 99.0,
      "eval_accuracy": 0.5538461538461539,
      "eval_loss": 2.8621606826782227,
      "eval_runtime": 2.676,
      "eval_samples_per_second": 48.58,
      "eval_steps_per_second": 0.747,
      "step": 693
    },
    {
      "epoch": 100.0,
      "learning_rate": 0.0,
      "loss": 2.0566,
      "step": 700
    },
    {
      "epoch": 100.0,
      "eval_accuracy": 0.5538461538461539,
      "eval_loss": 2.862189292907715,
      "eval_runtime": 2.8132,
      "eval_samples_per_second": 46.211,
      "eval_steps_per_second": 0.711,
      "step": 700
    },
    {
      "epoch": 100.0,
      "step": 700,
      "total_flos": 4.031389254057984e+18,
      "train_loss": 2.6316724504743303,
      "train_runtime": 4344.3701,
      "train_samples_per_second": 11.97,
      "train_steps_per_second": 0.161
    },
    {
      "epoch": 101.0,
      "eval_accuracy": 0.5153846153846153,
      "eval_loss": 2.84305477142334,
      "eval_runtime": 2.1617,
      "eval_samples_per_second": 60.137,
      "eval_steps_per_second": 0.925,
      "step": 707
    },
    {
      "epoch": 101.43,
      "learning_rate": 1.9761904761904763e-05,
      "loss": 2.067,
      "step": 710
    },
    {
      "epoch": 102.0,
      "eval_accuracy": 0.5,
      "eval_loss": 2.8063366413116455,
      "eval_runtime": 2.1563,
      "eval_samples_per_second": 60.288,
      "eval_steps_per_second": 0.928,
      "step": 714
    },
    {
      "epoch": 102.86,
      "learning_rate": 1.9523809523809524e-05,
      "loss": 2.0234,
      "step": 720
    },
    {
      "epoch": 103.0,
      "eval_accuracy": 0.5230769230769231,
      "eval_loss": 2.7845773696899414,
      "eval_runtime": 2.1676,
      "eval_samples_per_second": 59.974,
      "eval_steps_per_second": 0.923,
      "step": 721
    },
    {
      "epoch": 104.0,
      "eval_accuracy": 0.5307692307692308,
      "eval_loss": 2.762021541595459,
      "eval_runtime": 2.2767,
      "eval_samples_per_second": 57.1,
      "eval_steps_per_second": 0.878,
      "step": 728
    },
    {
      "epoch": 104.29,
      "learning_rate": 1.928571428571429e-05,
      "loss": 1.9979,
      "step": 730
    },
    {
      "epoch": 105.0,
      "eval_accuracy": 0.5538461538461539,
      "eval_loss": 2.7421603202819824,
      "eval_runtime": 2.2859,
      "eval_samples_per_second": 56.871,
      "eval_steps_per_second": 0.875,
      "step": 735
    },
    {
      "epoch": 105.71,
      "learning_rate": 1.904761904761905e-05,
      "loss": 1.9515,
      "step": 740
    },
    {
      "epoch": 106.0,
      "eval_accuracy": 0.5615384615384615,
      "eval_loss": 2.722872734069824,
      "eval_runtime": 2.3487,
      "eval_samples_per_second": 55.35,
      "eval_steps_per_second": 0.852,
      "step": 742
    },
    {
      "epoch": 107.0,
      "eval_accuracy": 0.5384615384615384,
      "eval_loss": 2.7145817279815674,
      "eval_runtime": 2.3976,
      "eval_samples_per_second": 54.221,
      "eval_steps_per_second": 0.834,
      "step": 749
    },
    {
      "epoch": 107.14,
      "learning_rate": 1.880952380952381e-05,
      "loss": 1.9095,
      "step": 750
    },
    {
      "epoch": 108.0,
      "eval_accuracy": 0.5538461538461539,
      "eval_loss": 2.6905264854431152,
      "eval_runtime": 2.3887,
      "eval_samples_per_second": 54.422,
      "eval_steps_per_second": 0.837,
      "step": 756
    },
    {
      "epoch": 108.57,
      "learning_rate": 1.8571428571428575e-05,
      "loss": 1.8759,
      "step": 760
    },
    {
      "epoch": 109.0,
      "eval_accuracy": 0.5538461538461539,
      "eval_loss": 2.6840174198150635,
      "eval_runtime": 2.369,
      "eval_samples_per_second": 54.876,
      "eval_steps_per_second": 0.844,
      "step": 763
    },
    {
      "epoch": 110.0,
      "learning_rate": 1.8333333333333333e-05,
      "loss": 1.8561,
      "step": 770
    },
    {
      "epoch": 110.0,
      "eval_accuracy": 0.5461538461538461,
      "eval_loss": 2.6554973125457764,
      "eval_runtime": 2.3964,
      "eval_samples_per_second": 54.248,
      "eval_steps_per_second": 0.835,
      "step": 770
    },
    {
      "epoch": 111.0,
      "eval_accuracy": 0.5692307692307692,
      "eval_loss": 2.6465163230895996,
      "eval_runtime": 2.5418,
      "eval_samples_per_second": 51.146,
      "eval_steps_per_second": 0.787,
      "step": 777
    },
    {
      "epoch": 111.43,
      "learning_rate": 1.8095238095238097e-05,
      "loss": 1.8147,
      "step": 780
    },
    {
      "epoch": 112.0,
      "eval_accuracy": 0.5384615384615384,
      "eval_loss": 2.6321260929107666,
      "eval_runtime": 2.6607,
      "eval_samples_per_second": 48.859,
      "eval_steps_per_second": 0.752,
      "step": 784
    },
    {
      "epoch": 112.86,
      "learning_rate": 1.785714285714286e-05,
      "loss": 1.799,
      "step": 790
    },
    {
      "epoch": 113.0,
      "eval_accuracy": 0.5384615384615384,
      "eval_loss": 2.6352975368499756,
      "eval_runtime": 2.3449,
      "eval_samples_per_second": 55.439,
      "eval_steps_per_second": 0.853,
      "step": 791
    },
    {
      "epoch": 114.0,
      "eval_accuracy": 0.5461538461538461,
      "eval_loss": 2.6188342571258545,
      "eval_runtime": 2.5585,
      "eval_samples_per_second": 50.811,
      "eval_steps_per_second": 0.782,
      "step": 798
    },
    {
      "epoch": 114.29,
      "learning_rate": 1.761904761904762e-05,
      "loss": 1.7536,
      "step": 800
    },
    {
      "epoch": 115.0,
      "eval_accuracy": 0.5461538461538461,
      "eval_loss": 2.6063857078552246,
      "eval_runtime": 2.5441,
      "eval_samples_per_second": 51.098,
      "eval_steps_per_second": 0.786,
      "step": 805
    },
    {
      "epoch": 115.71,
      "learning_rate": 1.7380952380952384e-05,
      "loss": 1.7162,
      "step": 810
    },
    {
      "epoch": 116.0,
      "eval_accuracy": 0.5538461538461539,
      "eval_loss": 2.591280937194824,
      "eval_runtime": 2.5836,
      "eval_samples_per_second": 50.316,
      "eval_steps_per_second": 0.774,
      "step": 812
    },
    {
      "epoch": 117.0,
      "eval_accuracy": 0.5923076923076923,
      "eval_loss": 2.5749051570892334,
      "eval_runtime": 2.5441,
      "eval_samples_per_second": 51.099,
      "eval_steps_per_second": 0.786,
      "step": 819
    },
    {
      "epoch": 117.14,
      "learning_rate": 1.7142857142857142e-05,
      "loss": 1.7225,
      "step": 820
    },
    {
      "epoch": 118.0,
      "eval_accuracy": 0.5846153846153846,
      "eval_loss": 2.563140869140625,
      "eval_runtime": 2.6117,
      "eval_samples_per_second": 49.777,
      "eval_steps_per_second": 0.766,
      "step": 826
    },
    {
      "epoch": 118.57,
      "learning_rate": 1.6904761904761906e-05,
      "loss": 1.6841,
      "step": 830
    },
    {
      "epoch": 119.0,
      "eval_accuracy": 0.5923076923076923,
      "eval_loss": 2.547883987426758,
      "eval_runtime": 2.5689,
      "eval_samples_per_second": 50.605,
      "eval_steps_per_second": 0.779,
      "step": 833
    },
    {
      "epoch": 120.0,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 1.6388,
      "step": 840
    },
    {
      "epoch": 120.0,
      "eval_accuracy": 0.5846153846153846,
      "eval_loss": 2.5402629375457764,
      "eval_runtime": 2.4843,
      "eval_samples_per_second": 52.328,
      "eval_steps_per_second": 0.805,
      "step": 840
    },
    {
      "epoch": 120.0,
      "step": 840,
      "total_flos": 4.837667104869581e+18,
      "train_loss": 0.3072653747740246,
      "train_runtime": 892.0889,
      "train_samples_per_second": 69.948,
      "train_steps_per_second": 0.942
    },
    {
      "epoch": 121.0,
      "eval_accuracy": 0.6692307692307692,
      "eval_loss": 2.35441255569458,
      "eval_runtime": 2.1277,
      "eval_samples_per_second": 61.098,
      "eval_steps_per_second": 0.94,
      "step": 847
    },
    {
      "epoch": 121.43,
      "learning_rate": 1.985714285714286e-05,
      "loss": 1.4504,
      "step": 850
    },
    {
      "epoch": 122.0,
      "eval_accuracy": 0.6461538461538462,
      "eval_loss": 2.3751370906829834,
      "eval_runtime": 2.1033,
      "eval_samples_per_second": 61.809,
      "eval_steps_per_second": 0.951,
      "step": 854
    },
    {
      "epoch": 122.86,
      "learning_rate": 1.9714285714285718e-05,
      "loss": 1.3963,
      "step": 860
    },
    {
      "epoch": 123.0,
      "eval_accuracy": 0.6538461538461539,
      "eval_loss": 2.3588593006134033,
      "eval_runtime": 2.0842,
      "eval_samples_per_second": 62.373,
      "eval_steps_per_second": 0.96,
      "step": 861
    },
    {
      "epoch": 124.0,
      "eval_accuracy": 0.6538461538461539,
      "eval_loss": 2.353787899017334,
      "eval_runtime": 2.1043,
      "eval_samples_per_second": 61.78,
      "eval_steps_per_second": 0.95,
      "step": 868
    },
    {
      "epoch": 124.29,
      "learning_rate": 1.9571428571428572e-05,
      "loss": 1.4087,
      "step": 870
    },
    {
      "epoch": 125.0,
      "eval_accuracy": 0.6461538461538462,
      "eval_loss": 2.326305627822876,
      "eval_runtime": 2.1627,
      "eval_samples_per_second": 60.11,
      "eval_steps_per_second": 0.925,
      "step": 875
    },
    {
      "epoch": 125.71,
      "learning_rate": 1.942857142857143e-05,
      "loss": 1.3564,
      "step": 880
    },
    {
      "epoch": 126.0,
      "eval_accuracy": 0.6230769230769231,
      "eval_loss": 2.3176610469818115,
      "eval_runtime": 2.2283,
      "eval_samples_per_second": 58.34,
      "eval_steps_per_second": 0.898,
      "step": 882
    },
    {
      "epoch": 127.0,
      "eval_accuracy": 0.6615384615384615,
      "eval_loss": 2.263662099838257,
      "eval_runtime": 2.1946,
      "eval_samples_per_second": 59.236,
      "eval_steps_per_second": 0.911,
      "step": 889
    },
    {
      "epoch": 127.14,
      "learning_rate": 1.928571428571429e-05,
      "loss": 1.373,
      "step": 890
    },
    {
      "epoch": 128.0,
      "eval_accuracy": 0.6538461538461539,
      "eval_loss": 2.2609002590179443,
      "eval_runtime": 2.3121,
      "eval_samples_per_second": 56.225,
      "eval_steps_per_second": 0.865,
      "step": 896
    },
    {
      "epoch": 128.57,
      "learning_rate": 1.9142857142857146e-05,
      "loss": 1.346,
      "step": 900
    },
    {
      "epoch": 129.0,
      "eval_accuracy": 0.6538461538461539,
      "eval_loss": 2.259404182434082,
      "eval_runtime": 2.1822,
      "eval_samples_per_second": 59.572,
      "eval_steps_per_second": 0.916,
      "step": 903
    },
    {
      "epoch": 130.0,
      "learning_rate": 1.9e-05,
      "loss": 1.3307,
      "step": 910
    },
    {
      "epoch": 130.0,
      "eval_accuracy": 0.6538461538461539,
      "eval_loss": 2.2564494609832764,
      "eval_runtime": 2.3375,
      "eval_samples_per_second": 55.615,
      "eval_steps_per_second": 0.856,
      "step": 910
    },
    {
      "epoch": 131.0,
      "eval_accuracy": 0.6307692307692307,
      "eval_loss": 2.2525382041931152,
      "eval_runtime": 2.2827,
      "eval_samples_per_second": 56.95,
      "eval_steps_per_second": 0.876,
      "step": 917
    },
    {
      "epoch": 131.43,
      "learning_rate": 1.885714285714286e-05,
      "loss": 1.3124,
      "step": 920
    },
    {
      "epoch": 132.0,
      "eval_accuracy": 0.6461538461538462,
      "eval_loss": 2.218500852584839,
      "eval_runtime": 2.2212,
      "eval_samples_per_second": 58.526,
      "eval_steps_per_second": 0.9,
      "step": 924
    },
    {
      "epoch": 132.86,
      "learning_rate": 1.8714285714285717e-05,
      "loss": 1.2848,
      "step": 930
    },
    {
      "epoch": 133.0,
      "eval_accuracy": 0.6461538461538462,
      "eval_loss": 2.2194907665252686,
      "eval_runtime": 2.3384,
      "eval_samples_per_second": 55.594,
      "eval_steps_per_second": 0.855,
      "step": 931
    },
    {
      "epoch": 134.0,
      "eval_accuracy": 0.6307692307692307,
      "eval_loss": 2.204814910888672,
      "eval_runtime": 2.2103,
      "eval_samples_per_second": 58.814,
      "eval_steps_per_second": 0.905,
      "step": 938
    },
    {
      "epoch": 134.29,
      "learning_rate": 1.8571428571428575e-05,
      "loss": 1.256,
      "step": 940
    },
    {
      "epoch": 135.0,
      "eval_accuracy": 0.6461538461538462,
      "eval_loss": 2.193894624710083,
      "eval_runtime": 2.2027,
      "eval_samples_per_second": 59.018,
      "eval_steps_per_second": 0.908,
      "step": 945
    },
    {
      "epoch": 135.71,
      "learning_rate": 1.842857142857143e-05,
      "loss": 1.24,
      "step": 950
    },
    {
      "epoch": 136.0,
      "eval_accuracy": 0.6461538461538462,
      "eval_loss": 2.1721882820129395,
      "eval_runtime": 2.3539,
      "eval_samples_per_second": 55.226,
      "eval_steps_per_second": 0.85,
      "step": 952
    },
    {
      "epoch": 137.0,
      "eval_accuracy": 0.6846153846153846,
      "eval_loss": 2.1497113704681396,
      "eval_runtime": 2.1949,
      "eval_samples_per_second": 59.23,
      "eval_steps_per_second": 0.911,
      "step": 959
    },
    {
      "epoch": 137.14,
      "learning_rate": 1.8285714285714288e-05,
      "loss": 1.2318,
      "step": 960
    },
    {
      "epoch": 138.0,
      "eval_accuracy": 0.6846153846153846,
      "eval_loss": 2.1332032680511475,
      "eval_runtime": 2.2346,
      "eval_samples_per_second": 58.176,
      "eval_steps_per_second": 0.895,
      "step": 966
    },
    {
      "epoch": 138.57,
      "learning_rate": 1.8142857142857146e-05,
      "loss": 1.1905,
      "step": 970
    },
    {
      "epoch": 139.0,
      "eval_accuracy": 0.6923076923076923,
      "eval_loss": 2.133913278579712,
      "eval_runtime": 2.1776,
      "eval_samples_per_second": 59.698,
      "eval_steps_per_second": 0.918,
      "step": 973
    },
    {
      "epoch": 140.0,
      "learning_rate": 1.8e-05,
      "loss": 1.186,
      "step": 980
    },
    {
      "epoch": 140.0,
      "eval_accuracy": 0.6692307692307692,
      "eval_loss": 2.1346120834350586,
      "eval_runtime": 2.3208,
      "eval_samples_per_second": 56.016,
      "eval_steps_per_second": 0.862,
      "step": 980
    },
    {
      "epoch": 141.0,
      "eval_accuracy": 0.7,
      "eval_loss": 2.092773914337158,
      "eval_runtime": 2.3738,
      "eval_samples_per_second": 54.764,
      "eval_steps_per_second": 0.843,
      "step": 987
    },
    {
      "epoch": 141.43,
      "learning_rate": 1.785714285714286e-05,
      "loss": 1.1759,
      "step": 990
    },
    {
      "epoch": 142.0,
      "eval_accuracy": 0.6846153846153846,
      "eval_loss": 2.1067099571228027,
      "eval_runtime": 2.356,
      "eval_samples_per_second": 55.178,
      "eval_steps_per_second": 0.849,
      "step": 994
    },
    {
      "epoch": 142.86,
      "learning_rate": 1.7714285714285717e-05,
      "loss": 1.132,
      "step": 1000
    },
    {
      "epoch": 143.0,
      "eval_accuracy": 0.6692307692307692,
      "eval_loss": 2.091548442840576,
      "eval_runtime": 2.3185,
      "eval_samples_per_second": 56.07,
      "eval_steps_per_second": 0.863,
      "step": 1001
    },
    {
      "epoch": 144.0,
      "eval_accuracy": 0.676923076923077,
      "eval_loss": 2.0903213024139404,
      "eval_runtime": 2.1804,
      "eval_samples_per_second": 59.622,
      "eval_steps_per_second": 0.917,
      "step": 1008
    },
    {
      "epoch": 144.29,
      "learning_rate": 1.757142857142857e-05,
      "loss": 1.1206,
      "step": 1010
    },
    {
      "epoch": 145.0,
      "eval_accuracy": 0.676923076923077,
      "eval_loss": 2.1074228286743164,
      "eval_runtime": 2.3685,
      "eval_samples_per_second": 54.887,
      "eval_steps_per_second": 0.844,
      "step": 1015
    },
    {
      "epoch": 145.71,
      "learning_rate": 1.742857142857143e-05,
      "loss": 1.1096,
      "step": 1020
    },
    {
      "epoch": 146.0,
      "eval_accuracy": 0.7230769230769231,
      "eval_loss": 2.0536632537841797,
      "eval_runtime": 2.5866,
      "eval_samples_per_second": 50.259,
      "eval_steps_per_second": 0.773,
      "step": 1022
    },
    {
      "epoch": 147.0,
      "eval_accuracy": 0.7,
      "eval_loss": 2.061009407043457,
      "eval_runtime": 2.3206,
      "eval_samples_per_second": 56.021,
      "eval_steps_per_second": 0.862,
      "step": 1029
    },
    {
      "epoch": 147.14,
      "learning_rate": 1.7285714285714287e-05,
      "loss": 1.0814,
      "step": 1030
    },
    {
      "epoch": 148.0,
      "eval_accuracy": 0.7076923076923077,
      "eval_loss": 2.0476534366607666,
      "eval_runtime": 2.1912,
      "eval_samples_per_second": 59.33,
      "eval_steps_per_second": 0.913,
      "step": 1036
    },
    {
      "epoch": 148.57,
      "learning_rate": 1.7142857142857142e-05,
      "loss": 1.0756,
      "step": 1040
    },
    {
      "epoch": 149.0,
      "eval_accuracy": 0.7076923076923077,
      "eval_loss": 2.0184433460235596,
      "eval_runtime": 2.2139,
      "eval_samples_per_second": 58.72,
      "eval_steps_per_second": 0.903,
      "step": 1043
    },
    {
      "epoch": 150.0,
      "learning_rate": 1.7e-05,
      "loss": 1.0638,
      "step": 1050
    },
    {
      "epoch": 150.0,
      "eval_accuracy": 0.6923076923076923,
      "eval_loss": 2.0189385414123535,
      "eval_runtime": 3.1458,
      "eval_samples_per_second": 41.325,
      "eval_steps_per_second": 0.636,
      "step": 1050
    },
    {
      "epoch": 151.0,
      "eval_accuracy": 0.7,
      "eval_loss": 2.01568865776062,
      "eval_runtime": 2.3668,
      "eval_samples_per_second": 54.926,
      "eval_steps_per_second": 0.845,
      "step": 1057
    },
    {
      "epoch": 151.43,
      "learning_rate": 1.6857142857142858e-05,
      "loss": 1.0343,
      "step": 1060
    },
    {
      "epoch": 152.0,
      "eval_accuracy": 0.6923076923076923,
      "eval_loss": 1.9978599548339844,
      "eval_runtime": 2.3546,
      "eval_samples_per_second": 55.211,
      "eval_steps_per_second": 0.849,
      "step": 1064
    },
    {
      "epoch": 152.86,
      "learning_rate": 1.6714285714285716e-05,
      "loss": 1.0187,
      "step": 1070
    },
    {
      "epoch": 153.0,
      "eval_accuracy": 0.7230769230769231,
      "eval_loss": 1.9824119806289673,
      "eval_runtime": 2.2036,
      "eval_samples_per_second": 58.995,
      "eval_steps_per_second": 0.908,
      "step": 1071
    },
    {
      "epoch": 154.0,
      "eval_accuracy": 0.7076923076923077,
      "eval_loss": 1.968167781829834,
      "eval_runtime": 2.422,
      "eval_samples_per_second": 53.675,
      "eval_steps_per_second": 0.826,
      "step": 1078
    },
    {
      "epoch": 154.29,
      "learning_rate": 1.6571428571428574e-05,
      "loss": 0.9993,
      "step": 1080
    },
    {
      "epoch": 155.0,
      "eval_accuracy": 0.7230769230769231,
      "eval_loss": 1.9394822120666504,
      "eval_runtime": 2.3562,
      "eval_samples_per_second": 55.175,
      "eval_steps_per_second": 0.849,
      "step": 1085
    },
    {
      "epoch": 155.71,
      "learning_rate": 1.642857142857143e-05,
      "loss": 1.0029,
      "step": 1090
    },
    {
      "epoch": 156.0,
      "eval_accuracy": 0.7153846153846154,
      "eval_loss": 1.920530080795288,
      "eval_runtime": 2.3989,
      "eval_samples_per_second": 54.191,
      "eval_steps_per_second": 0.834,
      "step": 1092
    },
    {
      "epoch": 157.0,
      "eval_accuracy": 0.7076923076923077,
      "eval_loss": 1.906773567199707,
      "eval_runtime": 2.3766,
      "eval_samples_per_second": 54.699,
      "eval_steps_per_second": 0.842,
      "step": 1099
    },
    {
      "epoch": 157.14,
      "learning_rate": 1.6285714285714287e-05,
      "loss": 0.9703,
      "step": 1100
    },
    {
      "epoch": 158.0,
      "eval_accuracy": 0.7384615384615385,
      "eval_loss": 1.899588704109192,
      "eval_runtime": 2.3516,
      "eval_samples_per_second": 55.282,
      "eval_steps_per_second": 0.85,
      "step": 1106
    },
    {
      "epoch": 158.57,
      "learning_rate": 1.6142857142857145e-05,
      "loss": 0.9532,
      "step": 1110
    },
    {
      "epoch": 159.0,
      "eval_accuracy": 0.7076923076923077,
      "eval_loss": 1.8944636583328247,
      "eval_runtime": 2.2108,
      "eval_samples_per_second": 58.803,
      "eval_steps_per_second": 0.905,
      "step": 1113
    },
    {
      "epoch": 160.0,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.9526,
      "step": 1120
    },
    {
      "epoch": 160.0,
      "eval_accuracy": 0.7307692307692307,
      "eval_loss": 1.8939263820648193,
      "eval_runtime": 2.2202,
      "eval_samples_per_second": 58.553,
      "eval_steps_per_second": 0.901,
      "step": 1120
    },
    {
      "epoch": 161.0,
      "eval_accuracy": 0.7230769230769231,
      "eval_loss": 1.8937941789627075,
      "eval_runtime": 2.3259,
      "eval_samples_per_second": 55.892,
      "eval_steps_per_second": 0.86,
      "step": 1127
    },
    {
      "epoch": 161.43,
      "learning_rate": 1.5857142857142857e-05,
      "loss": 0.9365,
      "step": 1130
    },
    {
      "epoch": 162.0,
      "eval_accuracy": 0.7384615384615385,
      "eval_loss": 1.8544682264328003,
      "eval_runtime": 2.1829,
      "eval_samples_per_second": 59.554,
      "eval_steps_per_second": 0.916,
      "step": 1134
    },
    {
      "epoch": 162.86,
      "learning_rate": 1.5714285714285715e-05,
      "loss": 0.9301,
      "step": 1140
    },
    {
      "epoch": 163.0,
      "eval_accuracy": 0.7384615384615385,
      "eval_loss": 1.860898733139038,
      "eval_runtime": 2.2861,
      "eval_samples_per_second": 56.864,
      "eval_steps_per_second": 0.875,
      "step": 1141
    },
    {
      "epoch": 164.0,
      "eval_accuracy": 0.7230769230769231,
      "eval_loss": 1.8540517091751099,
      "eval_runtime": 2.2092,
      "eval_samples_per_second": 58.844,
      "eval_steps_per_second": 0.905,
      "step": 1148
    },
    {
      "epoch": 164.29,
      "learning_rate": 1.5571428571428573e-05,
      "loss": 0.8856,
      "step": 1150
    },
    {
      "epoch": 165.0,
      "eval_accuracy": 0.7076923076923077,
      "eval_loss": 1.884601354598999,
      "eval_runtime": 2.2447,
      "eval_samples_per_second": 57.915,
      "eval_steps_per_second": 0.891,
      "step": 1155
    },
    {
      "epoch": 165.71,
      "learning_rate": 1.542857142857143e-05,
      "loss": 0.8801,
      "step": 1160
    },
    {
      "epoch": 166.0,
      "eval_accuracy": 0.7076923076923077,
      "eval_loss": 1.867126226425171,
      "eval_runtime": 2.2028,
      "eval_samples_per_second": 59.017,
      "eval_steps_per_second": 0.908,
      "step": 1162
    },
    {
      "epoch": 167.0,
      "eval_accuracy": 0.7230769230769231,
      "eval_loss": 1.8465133905410767,
      "eval_runtime": 2.3398,
      "eval_samples_per_second": 55.559,
      "eval_steps_per_second": 0.855,
      "step": 1169
    },
    {
      "epoch": 167.14,
      "learning_rate": 1.5285714285714286e-05,
      "loss": 0.8898,
      "step": 1170
    },
    {
      "epoch": 168.0,
      "eval_accuracy": 0.7230769230769231,
      "eval_loss": 1.8428664207458496,
      "eval_runtime": 2.1983,
      "eval_samples_per_second": 59.138,
      "eval_steps_per_second": 0.91,
      "step": 1176
    },
    {
      "epoch": 168.57,
      "learning_rate": 1.5142857142857144e-05,
      "loss": 0.8729,
      "step": 1180
    },
    {
      "epoch": 169.0,
      "eval_accuracy": 0.7384615384615385,
      "eval_loss": 1.832564115524292,
      "eval_runtime": 2.357,
      "eval_samples_per_second": 55.155,
      "eval_steps_per_second": 0.849,
      "step": 1183
    },
    {
      "epoch": 170.0,
      "learning_rate": 1.5000000000000002e-05,
      "loss": 0.8632,
      "step": 1190
    },
    {
      "epoch": 170.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.810370683670044,
      "eval_runtime": 2.4307,
      "eval_samples_per_second": 53.483,
      "eval_steps_per_second": 0.823,
      "step": 1190
    },
    {
      "epoch": 171.0,
      "eval_accuracy": 0.7384615384615385,
      "eval_loss": 1.8017692565917969,
      "eval_runtime": 2.2052,
      "eval_samples_per_second": 58.951,
      "eval_steps_per_second": 0.907,
      "step": 1197
    },
    {
      "epoch": 171.43,
      "learning_rate": 1.4857142857142858e-05,
      "loss": 0.8423,
      "step": 1200
    },
    {
      "epoch": 172.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.7875727415084839,
      "eval_runtime": 2.3658,
      "eval_samples_per_second": 54.949,
      "eval_steps_per_second": 0.845,
      "step": 1204
    },
    {
      "epoch": 172.86,
      "learning_rate": 1.4714285714285716e-05,
      "loss": 0.828,
      "step": 1210
    },
    {
      "epoch": 173.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.787375569343567,
      "eval_runtime": 2.2488,
      "eval_samples_per_second": 57.808,
      "eval_steps_per_second": 0.889,
      "step": 1211
    },
    {
      "epoch": 174.0,
      "eval_accuracy": 0.7692307692307693,
      "eval_loss": 1.7785577774047852,
      "eval_runtime": 2.3295,
      "eval_samples_per_second": 55.807,
      "eval_steps_per_second": 0.859,
      "step": 1218
    },
    {
      "epoch": 174.29,
      "learning_rate": 1.4571428571428573e-05,
      "loss": 0.8063,
      "step": 1220
    },
    {
      "epoch": 175.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.7820396423339844,
      "eval_runtime": 2.2553,
      "eval_samples_per_second": 57.642,
      "eval_steps_per_second": 0.887,
      "step": 1225
    },
    {
      "epoch": 175.71,
      "learning_rate": 1.4428571428571429e-05,
      "loss": 0.8084,
      "step": 1230
    },
    {
      "epoch": 176.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.800405502319336,
      "eval_runtime": 2.303,
      "eval_samples_per_second": 56.447,
      "eval_steps_per_second": 0.868,
      "step": 1232
    },
    {
      "epoch": 177.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.787461757659912,
      "eval_runtime": 2.2538,
      "eval_samples_per_second": 57.679,
      "eval_steps_per_second": 0.887,
      "step": 1239
    },
    {
      "epoch": 177.14,
      "learning_rate": 1.4285714285714287e-05,
      "loss": 0.7948,
      "step": 1240
    },
    {
      "epoch": 178.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.7639268636703491,
      "eval_runtime": 2.3516,
      "eval_samples_per_second": 55.281,
      "eval_steps_per_second": 0.85,
      "step": 1246
    },
    {
      "epoch": 178.57,
      "learning_rate": 1.4142857142857145e-05,
      "loss": 0.7687,
      "step": 1250
    },
    {
      "epoch": 179.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.7797616720199585,
      "eval_runtime": 2.2016,
      "eval_samples_per_second": 59.047,
      "eval_steps_per_second": 0.908,
      "step": 1253
    },
    {
      "epoch": 180.0,
      "learning_rate": 1.4e-05,
      "loss": 0.7598,
      "step": 1260
    },
    {
      "epoch": 180.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.7775954008102417,
      "eval_runtime": 2.2652,
      "eval_samples_per_second": 57.389,
      "eval_steps_per_second": 0.883,
      "step": 1260
    },
    {
      "epoch": 181.0,
      "eval_accuracy": 0.7384615384615385,
      "eval_loss": 1.7582831382751465,
      "eval_runtime": 2.3301,
      "eval_samples_per_second": 55.792,
      "eval_steps_per_second": 0.858,
      "step": 1267
    },
    {
      "epoch": 181.43,
      "learning_rate": 1.3857142857142858e-05,
      "loss": 0.7622,
      "step": 1270
    },
    {
      "epoch": 182.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.7546014785766602,
      "eval_runtime": 2.3427,
      "eval_samples_per_second": 55.493,
      "eval_steps_per_second": 0.854,
      "step": 1274
    },
    {
      "epoch": 182.86,
      "learning_rate": 1.3714285714285716e-05,
      "loss": 0.754,
      "step": 1280
    },
    {
      "epoch": 183.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.7243136167526245,
      "eval_runtime": 2.2386,
      "eval_samples_per_second": 58.071,
      "eval_steps_per_second": 0.893,
      "step": 1281
    },
    {
      "epoch": 184.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.7218064069747925,
      "eval_runtime": 2.3487,
      "eval_samples_per_second": 55.35,
      "eval_steps_per_second": 0.852,
      "step": 1288
    },
    {
      "epoch": 184.29,
      "learning_rate": 1.3571428571428574e-05,
      "loss": 0.7397,
      "step": 1290
    },
    {
      "epoch": 185.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.6983821392059326,
      "eval_runtime": 2.2077,
      "eval_samples_per_second": 58.886,
      "eval_steps_per_second": 0.906,
      "step": 1295
    },
    {
      "epoch": 185.71,
      "learning_rate": 1.3428571428571429e-05,
      "loss": 0.7171,
      "step": 1300
    },
    {
      "epoch": 186.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.6678060293197632,
      "eval_runtime": 2.3645,
      "eval_samples_per_second": 54.98,
      "eval_steps_per_second": 0.846,
      "step": 1302
    },
    {
      "epoch": 187.0,
      "eval_accuracy": 0.7692307692307693,
      "eval_loss": 1.6591731309890747,
      "eval_runtime": 2.3411,
      "eval_samples_per_second": 55.529,
      "eval_steps_per_second": 0.854,
      "step": 1309
    },
    {
      "epoch": 187.14,
      "learning_rate": 1.3285714285714287e-05,
      "loss": 0.7351,
      "step": 1310
    },
    {
      "epoch": 188.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.67545747756958,
      "eval_runtime": 2.2255,
      "eval_samples_per_second": 58.414,
      "eval_steps_per_second": 0.899,
      "step": 1316
    },
    {
      "epoch": 188.57,
      "learning_rate": 1.3142857142857145e-05,
      "loss": 0.717,
      "step": 1320
    },
    {
      "epoch": 189.0,
      "eval_accuracy": 0.7846153846153846,
      "eval_loss": 1.6684316396713257,
      "eval_runtime": 2.2461,
      "eval_samples_per_second": 57.879,
      "eval_steps_per_second": 0.89,
      "step": 1323
    },
    {
      "epoch": 190.0,
      "learning_rate": 1.3000000000000001e-05,
      "loss": 0.703,
      "step": 1330
    },
    {
      "epoch": 190.0,
      "eval_accuracy": 0.7692307692307693,
      "eval_loss": 1.6624796390533447,
      "eval_runtime": 2.3197,
      "eval_samples_per_second": 56.043,
      "eval_steps_per_second": 0.862,
      "step": 1330
    },
    {
      "epoch": 191.0,
      "eval_accuracy": 0.7769230769230769,
      "eval_loss": 1.6624727249145508,
      "eval_runtime": 2.2578,
      "eval_samples_per_second": 57.577,
      "eval_steps_per_second": 0.886,
      "step": 1337
    },
    {
      "epoch": 191.43,
      "learning_rate": 1.2857142857142859e-05,
      "loss": 0.7036,
      "step": 1340
    },
    {
      "epoch": 192.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.6422407627105713,
      "eval_runtime": 2.4009,
      "eval_samples_per_second": 54.146,
      "eval_steps_per_second": 0.833,
      "step": 1344
    },
    {
      "epoch": 192.86,
      "learning_rate": 1.2714285714285715e-05,
      "loss": 0.698,
      "step": 1350
    },
    {
      "epoch": 193.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.6632704734802246,
      "eval_runtime": 2.2565,
      "eval_samples_per_second": 57.612,
      "eval_steps_per_second": 0.886,
      "step": 1351
    },
    {
      "epoch": 194.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.663210153579712,
      "eval_runtime": 2.3463,
      "eval_samples_per_second": 55.406,
      "eval_steps_per_second": 0.852,
      "step": 1358
    },
    {
      "epoch": 194.29,
      "learning_rate": 1.2571428571428572e-05,
      "loss": 0.7002,
      "step": 1360
    },
    {
      "epoch": 195.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.6452277898788452,
      "eval_runtime": 2.2223,
      "eval_samples_per_second": 58.497,
      "eval_steps_per_second": 0.9,
      "step": 1365
    },
    {
      "epoch": 195.71,
      "learning_rate": 1.242857142857143e-05,
      "loss": 0.6662,
      "step": 1370
    },
    {
      "epoch": 196.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.6403251886367798,
      "eval_runtime": 2.3447,
      "eval_samples_per_second": 55.444,
      "eval_steps_per_second": 0.853,
      "step": 1372
    },
    {
      "epoch": 197.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.627591848373413,
      "eval_runtime": 2.3495,
      "eval_samples_per_second": 55.332,
      "eval_steps_per_second": 0.851,
      "step": 1379
    },
    {
      "epoch": 197.14,
      "learning_rate": 1.2285714285714288e-05,
      "loss": 0.6771,
      "step": 1380
    },
    {
      "epoch": 198.0,
      "eval_accuracy": 0.7769230769230769,
      "eval_loss": 1.6097954511642456,
      "eval_runtime": 2.2213,
      "eval_samples_per_second": 58.524,
      "eval_steps_per_second": 0.9,
      "step": 1386
    },
    {
      "epoch": 198.57,
      "learning_rate": 1.2142857142857142e-05,
      "loss": 0.6588,
      "step": 1390
    },
    {
      "epoch": 199.0,
      "eval_accuracy": 0.7692307692307693,
      "eval_loss": 1.622355580329895,
      "eval_runtime": 2.2376,
      "eval_samples_per_second": 58.098,
      "eval_steps_per_second": 0.894,
      "step": 1393
    },
    {
      "epoch": 200.0,
      "learning_rate": 1.2e-05,
      "loss": 0.6549,
      "step": 1400
    },
    {
      "epoch": 200.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.6263408660888672,
      "eval_runtime": 2.3312,
      "eval_samples_per_second": 55.765,
      "eval_steps_per_second": 0.858,
      "step": 1400
    },
    {
      "epoch": 200.0,
      "step": 1400,
      "total_flos": 8.062778508115968e+18,
      "train_loss": 0.39147548845836094,
      "train_runtime": 3403.5317,
      "train_samples_per_second": 30.556,
      "train_steps_per_second": 0.411
    },
    {
      "epoch": 201.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.6449843645095825,
      "eval_runtime": 2.095,
      "eval_samples_per_second": 62.053,
      "eval_steps_per_second": 0.955,
      "step": 1407
    },
    {
      "epoch": 201.43,
      "learning_rate": 1.988571428571429e-05,
      "loss": 0.706,
      "step": 1410
    },
    {
      "epoch": 202.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.646350622177124,
      "eval_runtime": 2.1984,
      "eval_samples_per_second": 59.134,
      "eval_steps_per_second": 0.91,
      "step": 1414
    },
    {
      "epoch": 202.86,
      "learning_rate": 1.9771428571428574e-05,
      "loss": 0.6746,
      "step": 1420
    },
    {
      "epoch": 203.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.6431909799575806,
      "eval_runtime": 2.1397,
      "eval_samples_per_second": 60.755,
      "eval_steps_per_second": 0.935,
      "step": 1421
    },
    {
      "epoch": 204.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.6220970153808594,
      "eval_runtime": 2.2125,
      "eval_samples_per_second": 58.756,
      "eval_steps_per_second": 0.904,
      "step": 1428
    },
    {
      "epoch": 204.29,
      "learning_rate": 1.9657142857142858e-05,
      "loss": 0.6282,
      "step": 1430
    },
    {
      "epoch": 205.0,
      "eval_accuracy": 0.7769230769230769,
      "eval_loss": 1.5988324880599976,
      "eval_runtime": 2.3558,
      "eval_samples_per_second": 55.182,
      "eval_steps_per_second": 0.849,
      "step": 1435
    },
    {
      "epoch": 205.71,
      "learning_rate": 1.9542857142857143e-05,
      "loss": 0.618,
      "step": 1440
    },
    {
      "epoch": 206.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.6031513214111328,
      "eval_runtime": 2.2121,
      "eval_samples_per_second": 58.767,
      "eval_steps_per_second": 0.904,
      "step": 1442
    },
    {
      "epoch": 207.0,
      "eval_accuracy": 0.7923076923076923,
      "eval_loss": 1.561190128326416,
      "eval_runtime": 2.2132,
      "eval_samples_per_second": 58.739,
      "eval_steps_per_second": 0.904,
      "step": 1449
    },
    {
      "epoch": 207.14,
      "learning_rate": 1.942857142857143e-05,
      "loss": 0.6053,
      "step": 1450
    },
    {
      "epoch": 208.0,
      "eval_accuracy": 0.7769230769230769,
      "eval_loss": 1.5888561010360718,
      "eval_runtime": 2.1847,
      "eval_samples_per_second": 59.504,
      "eval_steps_per_second": 0.915,
      "step": 1456
    },
    {
      "epoch": 208.57,
      "learning_rate": 1.9314285714285718e-05,
      "loss": 0.5977,
      "step": 1460
    },
    {
      "epoch": 209.0,
      "eval_accuracy": 0.7692307692307693,
      "eval_loss": 1.566961646080017,
      "eval_runtime": 2.2023,
      "eval_samples_per_second": 59.03,
      "eval_steps_per_second": 0.908,
      "step": 1463
    },
    {
      "epoch": 210.0,
      "learning_rate": 1.9200000000000003e-05,
      "loss": 0.6131,
      "step": 1470
    },
    {
      "epoch": 210.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.5975637435913086,
      "eval_runtime": 2.3397,
      "eval_samples_per_second": 55.564,
      "eval_steps_per_second": 0.855,
      "step": 1470
    },
    {
      "epoch": 211.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.5593103170394897,
      "eval_runtime": 2.2361,
      "eval_samples_per_second": 58.136,
      "eval_steps_per_second": 0.894,
      "step": 1477
    },
    {
      "epoch": 211.43,
      "learning_rate": 1.9085714285714287e-05,
      "loss": 0.5753,
      "step": 1480
    },
    {
      "epoch": 212.0,
      "eval_accuracy": 0.7384615384615385,
      "eval_loss": 1.5974284410476685,
      "eval_runtime": 2.3233,
      "eval_samples_per_second": 55.955,
      "eval_steps_per_second": 0.861,
      "step": 1484
    },
    {
      "epoch": 212.86,
      "learning_rate": 1.8971428571428575e-05,
      "loss": 0.576,
      "step": 1490
    },
    {
      "epoch": 213.0,
      "eval_accuracy": 0.7769230769230769,
      "eval_loss": 1.573266863822937,
      "eval_runtime": 2.2312,
      "eval_samples_per_second": 58.264,
      "eval_steps_per_second": 0.896,
      "step": 1491
    },
    {
      "epoch": 214.0,
      "eval_accuracy": 0.7769230769230769,
      "eval_loss": 1.497517466545105,
      "eval_runtime": 2.3929,
      "eval_samples_per_second": 54.328,
      "eval_steps_per_second": 0.836,
      "step": 1498
    },
    {
      "epoch": 214.29,
      "learning_rate": 1.885714285714286e-05,
      "loss": 0.5744,
      "step": 1500
    },
    {
      "epoch": 215.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.562217116355896,
      "eval_runtime": 2.2169,
      "eval_samples_per_second": 58.642,
      "eval_steps_per_second": 0.902,
      "step": 1505
    },
    {
      "epoch": 215.71,
      "learning_rate": 1.8742857142857143e-05,
      "loss": 0.5658,
      "step": 1510
    },
    {
      "epoch": 216.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.5334222316741943,
      "eval_runtime": 2.3054,
      "eval_samples_per_second": 56.388,
      "eval_steps_per_second": 0.868,
      "step": 1512
    },
    {
      "epoch": 217.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.5438138246536255,
      "eval_runtime": 2.3277,
      "eval_samples_per_second": 55.849,
      "eval_steps_per_second": 0.859,
      "step": 1519
    },
    {
      "epoch": 217.14,
      "learning_rate": 1.8628571428571428e-05,
      "loss": 0.5507,
      "step": 1520
    },
    {
      "epoch": 218.0,
      "eval_accuracy": 0.7307692307692307,
      "eval_loss": 1.5617175102233887,
      "eval_runtime": 2.3945,
      "eval_samples_per_second": 54.291,
      "eval_steps_per_second": 0.835,
      "step": 1526
    },
    {
      "epoch": 218.57,
      "learning_rate": 1.8514285714285716e-05,
      "loss": 0.5474,
      "step": 1530
    },
    {
      "epoch": 219.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.542279601097107,
      "eval_runtime": 2.1837,
      "eval_samples_per_second": 59.533,
      "eval_steps_per_second": 0.916,
      "step": 1533
    },
    {
      "epoch": 220.0,
      "learning_rate": 1.8400000000000003e-05,
      "loss": 0.5095,
      "step": 1540
    },
    {
      "epoch": 220.0,
      "eval_accuracy": 0.7769230769230769,
      "eval_loss": 1.509877324104309,
      "eval_runtime": 2.1925,
      "eval_samples_per_second": 59.293,
      "eval_steps_per_second": 0.912,
      "step": 1540
    },
    {
      "epoch": 221.0,
      "eval_accuracy": 0.7692307692307693,
      "eval_loss": 1.5009371042251587,
      "eval_runtime": 2.1917,
      "eval_samples_per_second": 59.315,
      "eval_steps_per_second": 0.913,
      "step": 1547
    },
    {
      "epoch": 221.43,
      "learning_rate": 1.8285714285714288e-05,
      "loss": 0.5263,
      "step": 1550
    },
    {
      "epoch": 222.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.4717506170272827,
      "eval_runtime": 2.1826,
      "eval_samples_per_second": 59.561,
      "eval_steps_per_second": 0.916,
      "step": 1554
    },
    {
      "epoch": 222.86,
      "learning_rate": 1.8171428571428572e-05,
      "loss": 0.4934,
      "step": 1560
    },
    {
      "epoch": 223.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.5150481462478638,
      "eval_runtime": 2.1689,
      "eval_samples_per_second": 59.939,
      "eval_steps_per_second": 0.922,
      "step": 1561
    },
    {
      "epoch": 224.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.5133140087127686,
      "eval_runtime": 2.2493,
      "eval_samples_per_second": 57.795,
      "eval_steps_per_second": 0.889,
      "step": 1568
    },
    {
      "epoch": 224.29,
      "learning_rate": 1.8057142857142857e-05,
      "loss": 0.4935,
      "step": 1570
    },
    {
      "epoch": 225.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.5053058862686157,
      "eval_runtime": 2.2888,
      "eval_samples_per_second": 56.799,
      "eval_steps_per_second": 0.874,
      "step": 1575
    },
    {
      "epoch": 225.71,
      "learning_rate": 1.7942857142857144e-05,
      "loss": 0.4808,
      "step": 1580
    },
    {
      "epoch": 226.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.5152372121810913,
      "eval_runtime": 2.3287,
      "eval_samples_per_second": 55.824,
      "eval_steps_per_second": 0.859,
      "step": 1582
    },
    {
      "epoch": 227.0,
      "eval_accuracy": 0.7384615384615385,
      "eval_loss": 1.521767497062683,
      "eval_runtime": 2.2464,
      "eval_samples_per_second": 57.871,
      "eval_steps_per_second": 0.89,
      "step": 1589
    },
    {
      "epoch": 227.14,
      "learning_rate": 1.7828571428571432e-05,
      "loss": 0.5039,
      "step": 1590
    },
    {
      "epoch": 228.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.4631297588348389,
      "eval_runtime": 2.1868,
      "eval_samples_per_second": 59.449,
      "eval_steps_per_second": 0.915,
      "step": 1596
    },
    {
      "epoch": 228.57,
      "learning_rate": 1.7714285714285717e-05,
      "loss": 0.4516,
      "step": 1600
    },
    {
      "epoch": 229.0,
      "eval_accuracy": 0.7692307692307693,
      "eval_loss": 1.4795747995376587,
      "eval_runtime": 2.3162,
      "eval_samples_per_second": 56.126,
      "eval_steps_per_second": 0.863,
      "step": 1603
    },
    {
      "epoch": 230.0,
      "learning_rate": 1.76e-05,
      "loss": 0.4655,
      "step": 1610
    },
    {
      "epoch": 230.0,
      "eval_accuracy": 0.7769230769230769,
      "eval_loss": 1.4648711681365967,
      "eval_runtime": 2.3186,
      "eval_samples_per_second": 56.068,
      "eval_steps_per_second": 0.863,
      "step": 1610
    },
    {
      "epoch": 231.0,
      "eval_accuracy": 0.7846153846153846,
      "eval_loss": 1.4528415203094482,
      "eval_runtime": 2.2367,
      "eval_samples_per_second": 58.12,
      "eval_steps_per_second": 0.894,
      "step": 1617
    },
    {
      "epoch": 231.43,
      "learning_rate": 1.748571428571429e-05,
      "loss": 0.4487,
      "step": 1620
    },
    {
      "epoch": 232.0,
      "eval_accuracy": 0.7769230769230769,
      "eval_loss": 1.4157787561416626,
      "eval_runtime": 2.2157,
      "eval_samples_per_second": 58.673,
      "eval_steps_per_second": 0.903,
      "step": 1624
    },
    {
      "epoch": 232.86,
      "learning_rate": 1.7371428571428573e-05,
      "loss": 0.453,
      "step": 1630
    },
    {
      "epoch": 233.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.4378057718276978,
      "eval_runtime": 2.2298,
      "eval_samples_per_second": 58.301,
      "eval_steps_per_second": 0.897,
      "step": 1631
    },
    {
      "epoch": 234.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.4233053922653198,
      "eval_runtime": 2.2254,
      "eval_samples_per_second": 58.416,
      "eval_steps_per_second": 0.899,
      "step": 1638
    },
    {
      "epoch": 234.29,
      "learning_rate": 1.7257142857142857e-05,
      "loss": 0.4248,
      "step": 1640
    },
    {
      "epoch": 235.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.4437103271484375,
      "eval_runtime": 2.2018,
      "eval_samples_per_second": 59.042,
      "eval_steps_per_second": 0.908,
      "step": 1645
    },
    {
      "epoch": 235.71,
      "learning_rate": 1.7142857142857142e-05,
      "loss": 0.4103,
      "step": 1650
    },
    {
      "epoch": 236.0,
      "eval_accuracy": 0.7384615384615385,
      "eval_loss": 1.4489341974258423,
      "eval_runtime": 2.3267,
      "eval_samples_per_second": 55.872,
      "eval_steps_per_second": 0.86,
      "step": 1652
    },
    {
      "epoch": 237.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.437820315361023,
      "eval_runtime": 2.2082,
      "eval_samples_per_second": 58.871,
      "eval_steps_per_second": 0.906,
      "step": 1659
    },
    {
      "epoch": 237.14,
      "learning_rate": 1.702857142857143e-05,
      "loss": 0.4146,
      "step": 1660
    },
    {
      "epoch": 238.0,
      "eval_accuracy": 0.7384615384615385,
      "eval_loss": 1.4532761573791504,
      "eval_runtime": 2.2013,
      "eval_samples_per_second": 59.057,
      "eval_steps_per_second": 0.909,
      "step": 1666
    },
    {
      "epoch": 238.57,
      "learning_rate": 1.6914285714285717e-05,
      "loss": 0.4313,
      "step": 1670
    },
    {
      "epoch": 239.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.444754958152771,
      "eval_runtime": 2.2195,
      "eval_samples_per_second": 58.573,
      "eval_steps_per_second": 0.901,
      "step": 1673
    },
    {
      "epoch": 240.0,
      "learning_rate": 1.6800000000000002e-05,
      "loss": 0.408,
      "step": 1680
    },
    {
      "epoch": 240.0,
      "eval_accuracy": 0.7692307692307693,
      "eval_loss": 1.428984522819519,
      "eval_runtime": 2.3419,
      "eval_samples_per_second": 55.51,
      "eval_steps_per_second": 0.854,
      "step": 1680
    },
    {
      "epoch": 241.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.4546881914138794,
      "eval_runtime": 2.2356,
      "eval_samples_per_second": 58.149,
      "eval_steps_per_second": 0.895,
      "step": 1687
    },
    {
      "epoch": 241.43,
      "learning_rate": 1.6685714285714286e-05,
      "loss": 0.4145,
      "step": 1690
    },
    {
      "epoch": 242.0,
      "eval_accuracy": 0.7461538461538462,
      "eval_loss": 1.3930408954620361,
      "eval_runtime": 2.2411,
      "eval_samples_per_second": 58.006,
      "eval_steps_per_second": 0.892,
      "step": 1694
    },
    {
      "epoch": 242.86,
      "learning_rate": 1.6571428571428574e-05,
      "loss": 0.4193,
      "step": 1700
    },
    {
      "epoch": 243.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.4186941385269165,
      "eval_runtime": 2.3409,
      "eval_samples_per_second": 55.534,
      "eval_steps_per_second": 0.854,
      "step": 1701
    },
    {
      "epoch": 244.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.3854211568832397,
      "eval_runtime": 2.288,
      "eval_samples_per_second": 56.817,
      "eval_steps_per_second": 0.874,
      "step": 1708
    },
    {
      "epoch": 244.29,
      "learning_rate": 1.645714285714286e-05,
      "loss": 0.3951,
      "step": 1710
    },
    {
      "epoch": 245.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.3977891206741333,
      "eval_runtime": 2.2608,
      "eval_samples_per_second": 57.502,
      "eval_steps_per_second": 0.885,
      "step": 1715
    },
    {
      "epoch": 245.71,
      "learning_rate": 1.6342857142857146e-05,
      "loss": 0.3871,
      "step": 1720
    },
    {
      "epoch": 246.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.3596566915512085,
      "eval_runtime": 2.3275,
      "eval_samples_per_second": 55.853,
      "eval_steps_per_second": 0.859,
      "step": 1722
    },
    {
      "epoch": 247.0,
      "eval_accuracy": 0.7692307692307693,
      "eval_loss": 1.346314549446106,
      "eval_runtime": 2.3219,
      "eval_samples_per_second": 55.989,
      "eval_steps_per_second": 0.861,
      "step": 1729
    },
    {
      "epoch": 247.14,
      "learning_rate": 1.622857142857143e-05,
      "loss": 0.3679,
      "step": 1730
    },
    {
      "epoch": 248.0,
      "eval_accuracy": 0.7615384615384615,
      "eval_loss": 1.3439607620239258,
      "eval_runtime": 2.249,
      "eval_samples_per_second": 57.803,
      "eval_steps_per_second": 0.889,
      "step": 1736
    },
    {
      "epoch": 248.57,
      "learning_rate": 1.6114285714285715e-05,
      "loss": 0.3788,
      "step": 1740
    },
    {
      "epoch": 249.0,
      "eval_accuracy": 0.7692307692307693,
      "eval_loss": 1.3646103143692017,
      "eval_runtime": 2.1853,
      "eval_samples_per_second": 59.489,
      "eval_steps_per_second": 0.915,
      "step": 1743
    },
    {
      "epoch": 250.0,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.379,
      "step": 1750
    },
    {
      "epoch": 250.0,
      "eval_accuracy": 0.7538461538461538,
      "eval_loss": 1.3682621717453003,
      "eval_runtime": 2.2333,
      "eval_samples_per_second": 58.209,
      "eval_steps_per_second": 0.896,
      "step": 1750
    },
    {
      "epoch": 250.0,
      "step": 1750,
      "total_flos": 1.007847313514496e+19,
      "train_loss": 0.09994067628043038,
      "train_runtime": 2169.5998,
      "train_samples_per_second": 59.919,
      "train_steps_per_second": 0.807
    }
  ],
  "logging_steps": 10,
  "max_steps": 1750,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 250,
  "save_steps": 500,
  "total_flos": 1.007847313514496e+19,
  "train_batch_size": 80,
  "trial_name": null,
  "trial_params": null
}