{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 133.33333333333334,
  "eval_steps": 500,
  "global_step": 900,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8888888888888888,
      "eval_accuracy": 0.09375,
      "eval_loss": 2.078925609588623,
      "eval_runtime": 2.1652,
      "eval_samples_per_second": 44.337,
      "eval_steps_per_second": 1.386,
      "step": 6
    },
    {
      "epoch": 1.4814814814814814,
      "grad_norm": 4.666333198547363,
      "learning_rate": 5.555555555555556e-06,
      "loss": 2.1005,
      "step": 10
    },
    {
      "epoch": 1.925925925925926,
      "eval_accuracy": 0.125,
      "eval_loss": 2.027583122253418,
      "eval_runtime": 1.9426,
      "eval_samples_per_second": 49.419,
      "eval_steps_per_second": 1.544,
      "step": 13
    },
    {
      "epoch": 2.962962962962963,
      "grad_norm": 6.9924726486206055,
      "learning_rate": 1.1111111111111112e-05,
      "loss": 2.0321,
      "step": 20
    },
    {
      "epoch": 2.962962962962963,
      "eval_accuracy": 0.28125,
      "eval_loss": 1.945550560951233,
      "eval_runtime": 1.8821,
      "eval_samples_per_second": 51.006,
      "eval_steps_per_second": 1.594,
      "step": 20
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.4583333333333333,
      "eval_loss": 1.8392573595046997,
      "eval_runtime": 1.8452,
      "eval_samples_per_second": 52.028,
      "eval_steps_per_second": 1.626,
      "step": 27
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 6.883745193481445,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 1.9151,
      "step": 30
    },
    {
      "epoch": 4.888888888888889,
      "eval_accuracy": 0.5833333333333334,
      "eval_loss": 1.7343072891235352,
      "eval_runtime": 1.8707,
      "eval_samples_per_second": 51.318,
      "eval_steps_per_second": 1.604,
      "step": 33
    },
    {
      "epoch": 5.925925925925926,
      "grad_norm": 12.400025367736816,
      "learning_rate": 2.2222222222222223e-05,
      "loss": 1.7396,
      "step": 40
    },
    {
      "epoch": 5.925925925925926,
      "eval_accuracy": 0.6041666666666666,
      "eval_loss": 1.5971508026123047,
      "eval_runtime": 1.891,
      "eval_samples_per_second": 50.767,
      "eval_steps_per_second": 1.586,
      "step": 40
    },
    {
      "epoch": 6.962962962962963,
      "eval_accuracy": 0.6770833333333334,
      "eval_loss": 1.4546260833740234,
      "eval_runtime": 1.8412,
      "eval_samples_per_second": 52.14,
      "eval_steps_per_second": 1.629,
      "step": 47
    },
    {
      "epoch": 7.407407407407407,
      "grad_norm": 11.476837158203125,
      "learning_rate": 2.777777777777778e-05,
      "loss": 1.5392,
      "step": 50
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7395833333333334,
      "eval_loss": 1.2942534685134888,
      "eval_runtime": 1.9365,
      "eval_samples_per_second": 49.574,
      "eval_steps_per_second": 1.549,
      "step": 54
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 14.580023765563965,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 1.3096,
      "step": 60
    },
    {
      "epoch": 8.88888888888889,
      "eval_accuracy": 0.7395833333333334,
      "eval_loss": 1.1409353017807007,
      "eval_runtime": 1.8719,
      "eval_samples_per_second": 51.286,
      "eval_steps_per_second": 1.603,
      "step": 60
    },
    {
      "epoch": 9.925925925925926,
      "eval_accuracy": 0.8229166666666666,
      "eval_loss": 0.9840934872627258,
      "eval_runtime": 1.8663,
      "eval_samples_per_second": 51.44,
      "eval_steps_per_second": 1.607,
      "step": 67
    },
    {
      "epoch": 10.37037037037037,
      "grad_norm": 15.968297958374023,
      "learning_rate": 3.888888888888889e-05,
      "loss": 1.1062,
      "step": 70
    },
    {
      "epoch": 10.962962962962964,
      "eval_accuracy": 0.8229166666666666,
      "eval_loss": 0.851224958896637,
      "eval_runtime": 1.9128,
      "eval_samples_per_second": 50.189,
      "eval_steps_per_second": 1.568,
      "step": 74
    },
    {
      "epoch": 11.851851851851851,
      "grad_norm": 23.297271728515625,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 0.896,
      "step": 80
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.8541666666666666,
      "eval_loss": 0.7128415703773499,
      "eval_runtime": 1.8414,
      "eval_samples_per_second": 52.136,
      "eval_steps_per_second": 1.629,
      "step": 81
    },
    {
      "epoch": 12.88888888888889,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 0.6365618109703064,
      "eval_runtime": 1.8634,
      "eval_samples_per_second": 51.518,
      "eval_steps_per_second": 1.61,
      "step": 87
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 24.776830673217773,
      "learning_rate": 5e-05,
      "loss": 0.712,
      "step": 90
    },
    {
      "epoch": 13.925925925925926,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.5419111847877502,
      "eval_runtime": 1.874,
      "eval_samples_per_second": 51.228,
      "eval_steps_per_second": 1.601,
      "step": 94
    },
    {
      "epoch": 14.814814814814815,
      "grad_norm": 26.37338638305664,
      "learning_rate": 4.938271604938271e-05,
      "loss": 0.6231,
      "step": 100
    },
    {
      "epoch": 14.962962962962964,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.5082051753997803,
      "eval_runtime": 1.8338,
      "eval_samples_per_second": 52.35,
      "eval_steps_per_second": 1.636,
      "step": 101
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.875,
      "eval_loss": 0.4674176871776581,
      "eval_runtime": 1.878,
      "eval_samples_per_second": 51.118,
      "eval_steps_per_second": 1.597,
      "step": 108
    },
    {
      "epoch": 16.296296296296298,
      "grad_norm": 26.630212783813477,
      "learning_rate": 4.876543209876544e-05,
      "loss": 0.4962,
      "step": 110
    },
    {
      "epoch": 16.88888888888889,
      "eval_accuracy": 0.8541666666666666,
      "eval_loss": 0.4479581415653229,
      "eval_runtime": 1.8306,
      "eval_samples_per_second": 52.441,
      "eval_steps_per_second": 1.639,
      "step": 114
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 23.847532272338867,
      "learning_rate": 4.814814814814815e-05,
      "loss": 0.4322,
      "step": 120
    },
    {
      "epoch": 17.925925925925927,
      "eval_accuracy": 0.875,
      "eval_loss": 0.41380929946899414,
      "eval_runtime": 1.8812,
      "eval_samples_per_second": 51.03,
      "eval_steps_per_second": 1.595,
      "step": 121
    },
    {
      "epoch": 18.962962962962962,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.39465710520744324,
      "eval_runtime": 1.8978,
      "eval_samples_per_second": 50.584,
      "eval_steps_per_second": 1.581,
      "step": 128
    },
    {
      "epoch": 19.25925925925926,
      "grad_norm": 23.30731201171875,
      "learning_rate": 4.7530864197530866e-05,
      "loss": 0.3937,
      "step": 130
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.3826509416103363,
      "eval_runtime": 1.8444,
      "eval_samples_per_second": 52.05,
      "eval_steps_per_second": 1.627,
      "step": 135
    },
    {
      "epoch": 20.74074074074074,
      "grad_norm": 22.551395416259766,
      "learning_rate": 4.691358024691358e-05,
      "loss": 0.3377,
      "step": 140
    },
    {
      "epoch": 20.88888888888889,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.36256420612335205,
      "eval_runtime": 1.8807,
      "eval_samples_per_second": 51.044,
      "eval_steps_per_second": 1.595,
      "step": 141
    },
    {
      "epoch": 21.925925925925927,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.3579378128051758,
      "eval_runtime": 1.8493,
      "eval_samples_per_second": 51.911,
      "eval_steps_per_second": 1.622,
      "step": 148
    },
    {
      "epoch": 22.22222222222222,
      "grad_norm": 19.129470825195312,
      "learning_rate": 4.62962962962963e-05,
      "loss": 0.3099,
      "step": 150
    },
    {
      "epoch": 22.962962962962962,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.35575783252716064,
      "eval_runtime": 1.8947,
      "eval_samples_per_second": 50.668,
      "eval_steps_per_second": 1.583,
      "step": 155
    },
    {
      "epoch": 23.703703703703702,
      "grad_norm": 34.541709899902344,
      "learning_rate": 4.567901234567901e-05,
      "loss": 0.2895,
      "step": 160
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.3242819011211395,
      "eval_runtime": 1.9128,
      "eval_samples_per_second": 50.188,
      "eval_steps_per_second": 1.568,
      "step": 162
    },
    {
      "epoch": 24.88888888888889,
      "eval_accuracy": 0.875,
      "eval_loss": 0.34730610251426697,
      "eval_runtime": 1.857,
      "eval_samples_per_second": 51.697,
      "eval_steps_per_second": 1.616,
      "step": 168
    },
    {
      "epoch": 25.185185185185187,
      "grad_norm": 27.23337745666504,
      "learning_rate": 4.506172839506173e-05,
      "loss": 0.2732,
      "step": 170
    },
    {
      "epoch": 25.925925925925927,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.34611785411834717,
      "eval_runtime": 1.9339,
      "eval_samples_per_second": 49.641,
      "eval_steps_per_second": 1.551,
      "step": 175
    },
    {
      "epoch": 26.666666666666668,
      "grad_norm": 26.6937255859375,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 0.2447,
      "step": 180
    },
    {
      "epoch": 26.962962962962962,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.34497448801994324,
      "eval_runtime": 1.9227,
      "eval_samples_per_second": 49.929,
      "eval_steps_per_second": 1.56,
      "step": 182
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.36034080386161804,
      "eval_runtime": 1.8685,
      "eval_samples_per_second": 51.378,
      "eval_steps_per_second": 1.606,
      "step": 189
    },
    {
      "epoch": 28.14814814814815,
      "grad_norm": 22.52947235107422,
      "learning_rate": 4.3827160493827164e-05,
      "loss": 0.2009,
      "step": 190
    },
    {
      "epoch": 28.88888888888889,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.3214483857154846,
      "eval_runtime": 1.8832,
      "eval_samples_per_second": 50.977,
      "eval_steps_per_second": 1.593,
      "step": 195
    },
    {
      "epoch": 29.62962962962963,
      "grad_norm": 23.971797943115234,
      "learning_rate": 4.3209876543209875e-05,
      "loss": 0.2064,
      "step": 200
    },
    {
      "epoch": 29.925925925925927,
      "eval_accuracy": 0.875,
      "eval_loss": 0.30430108308792114,
      "eval_runtime": 1.8494,
      "eval_samples_per_second": 51.909,
      "eval_steps_per_second": 1.622,
      "step": 202
    },
    {
      "epoch": 30.962962962962962,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.2916516661643982,
      "eval_runtime": 1.8828,
      "eval_samples_per_second": 50.987,
      "eval_steps_per_second": 1.593,
      "step": 209
    },
    {
      "epoch": 31.11111111111111,
      "grad_norm": 20.200942993164062,
      "learning_rate": 4.259259259259259e-05,
      "loss": 0.2139,
      "step": 210
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.28600063920021057,
      "eval_runtime": 1.903,
      "eval_samples_per_second": 50.448,
      "eval_steps_per_second": 1.576,
      "step": 216
    },
    {
      "epoch": 32.592592592592595,
      "grad_norm": 49.91157531738281,
      "learning_rate": 4.197530864197531e-05,
      "loss": 0.1732,
      "step": 220
    },
    {
      "epoch": 32.888888888888886,
      "eval_accuracy": 0.8333333333333334,
      "eval_loss": 0.3314121663570404,
      "eval_runtime": 1.8588,
      "eval_samples_per_second": 51.646,
      "eval_steps_per_second": 1.614,
      "step": 222
    },
    {
      "epoch": 33.925925925925924,
      "eval_accuracy": 0.875,
      "eval_loss": 0.339076429605484,
      "eval_runtime": 1.8736,
      "eval_samples_per_second": 51.238,
      "eval_steps_per_second": 1.601,
      "step": 229
    },
    {
      "epoch": 34.074074074074076,
      "grad_norm": 15.412144660949707,
      "learning_rate": 4.135802469135803e-05,
      "loss": 0.2009,
      "step": 230
    },
    {
      "epoch": 34.96296296296296,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.3118279278278351,
      "eval_runtime": 1.8956,
      "eval_samples_per_second": 50.645,
      "eval_steps_per_second": 1.583,
      "step": 236
    },
    {
      "epoch": 35.55555555555556,
      "grad_norm": 18.91287612915039,
      "learning_rate": 4.074074074074074e-05,
      "loss": 0.1683,
      "step": 240
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.875,
      "eval_loss": 0.3162083327770233,
      "eval_runtime": 1.8823,
      "eval_samples_per_second": 51.002,
      "eval_steps_per_second": 1.594,
      "step": 243
    },
    {
      "epoch": 36.888888888888886,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.3011355996131897,
      "eval_runtime": 1.8611,
      "eval_samples_per_second": 51.582,
      "eval_steps_per_second": 1.612,
      "step": 249
    },
    {
      "epoch": 37.03703703703704,
      "grad_norm": 35.87955856323242,
      "learning_rate": 4.012345679012346e-05,
      "loss": 0.16,
      "step": 250
    },
    {
      "epoch": 37.925925925925924,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.2981296479701996,
      "eval_runtime": 1.9139,
      "eval_samples_per_second": 50.159,
      "eval_steps_per_second": 1.567,
      "step": 256
    },
    {
      "epoch": 38.51851851851852,
      "grad_norm": 12.093141555786133,
      "learning_rate": 3.950617283950617e-05,
      "loss": 0.1448,
      "step": 260
    },
    {
      "epoch": 38.96296296296296,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.34168219566345215,
      "eval_runtime": 1.8741,
      "eval_samples_per_second": 51.226,
      "eval_steps_per_second": 1.601,
      "step": 263
    },
    {
      "epoch": 40.0,
      "grad_norm": 36.96841049194336,
      "learning_rate": 3.888888888888889e-05,
      "loss": 0.1272,
      "step": 270
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.35579848289489746,
      "eval_runtime": 1.8827,
      "eval_samples_per_second": 50.992,
      "eval_steps_per_second": 1.593,
      "step": 270
    },
    {
      "epoch": 40.888888888888886,
      "eval_accuracy": 0.8541666666666666,
      "eval_loss": 0.3948463499546051,
      "eval_runtime": 1.8989,
      "eval_samples_per_second": 50.557,
      "eval_steps_per_second": 1.58,
      "step": 276
    },
    {
      "epoch": 41.48148148148148,
      "grad_norm": 17.500049591064453,
      "learning_rate": 3.82716049382716e-05,
      "loss": 0.1578,
      "step": 280
    },
    {
      "epoch": 41.925925925925924,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.36678051948547363,
      "eval_runtime": 1.9696,
      "eval_samples_per_second": 48.741,
      "eval_steps_per_second": 1.523,
      "step": 283
    },
    {
      "epoch": 42.96296296296296,
      "grad_norm": 19.923572540283203,
      "learning_rate": 3.7654320987654326e-05,
      "loss": 0.1604,
      "step": 290
    },
    {
      "epoch": 42.96296296296296,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.33422672748565674,
      "eval_runtime": 1.9103,
      "eval_samples_per_second": 50.253,
      "eval_steps_per_second": 1.57,
      "step": 290
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.9166666666666666,
      "eval_loss": 0.314091295003891,
      "eval_runtime": 1.8512,
      "eval_samples_per_second": 51.858,
      "eval_steps_per_second": 1.621,
      "step": 297
    },
    {
      "epoch": 44.44444444444444,
      "grad_norm": 21.997079849243164,
      "learning_rate": 3.7037037037037037e-05,
      "loss": 0.1251,
      "step": 300
    },
    {
      "epoch": 44.888888888888886,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.3266027271747589,
      "eval_runtime": 1.8596,
      "eval_samples_per_second": 51.625,
      "eval_steps_per_second": 1.613,
      "step": 303
    },
    {
      "epoch": 45.925925925925924,
      "grad_norm": 27.58420753479004,
      "learning_rate": 3.6419753086419754e-05,
      "loss": 0.1449,
      "step": 310
    },
    {
      "epoch": 45.925925925925924,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.34379109740257263,
      "eval_runtime": 1.9283,
      "eval_samples_per_second": 49.786,
      "eval_steps_per_second": 1.556,
      "step": 310
    },
    {
      "epoch": 46.96296296296296,
      "eval_accuracy": 0.875,
      "eval_loss": 0.3382701873779297,
      "eval_runtime": 1.8689,
      "eval_samples_per_second": 51.368,
      "eval_steps_per_second": 1.605,
      "step": 317
    },
    {
      "epoch": 47.407407407407405,
      "grad_norm": 14.891225814819336,
      "learning_rate": 3.580246913580247e-05,
      "loss": 0.1134,
      "step": 320
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.33413389325141907,
      "eval_runtime": 1.887,
      "eval_samples_per_second": 50.875,
      "eval_steps_per_second": 1.59,
      "step": 324
    },
    {
      "epoch": 48.888888888888886,
      "grad_norm": 22.530115127563477,
      "learning_rate": 3.518518518518519e-05,
      "loss": 0.1558,
      "step": 330
    },
    {
      "epoch": 48.888888888888886,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.2854965031147003,
      "eval_runtime": 1.9011,
      "eval_samples_per_second": 50.497,
      "eval_steps_per_second": 1.578,
      "step": 330
    },
    {
      "epoch": 49.925925925925924,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.2842860221862793,
      "eval_runtime": 1.9386,
      "eval_samples_per_second": 49.52,
      "eval_steps_per_second": 1.547,
      "step": 337
    },
    {
      "epoch": 50.370370370370374,
      "grad_norm": 45.90536880493164,
      "learning_rate": 3.45679012345679e-05,
      "loss": 0.1433,
      "step": 340
    },
    {
      "epoch": 50.96296296296296,
      "eval_accuracy": 0.84375,
      "eval_loss": 0.2878771126270294,
      "eval_runtime": 1.8358,
      "eval_samples_per_second": 52.292,
      "eval_steps_per_second": 1.634,
      "step": 344
    },
    {
      "epoch": 51.851851851851855,
      "grad_norm": 30.33245849609375,
      "learning_rate": 3.395061728395062e-05,
      "loss": 0.1207,
      "step": 350
    },
    {
      "epoch": 52.0,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.28866705298423767,
      "eval_runtime": 1.8609,
      "eval_samples_per_second": 51.587,
      "eval_steps_per_second": 1.612,
      "step": 351
    },
    {
      "epoch": 52.888888888888886,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.3173443377017975,
      "eval_runtime": 1.858,
      "eval_samples_per_second": 51.668,
      "eval_steps_per_second": 1.615,
      "step": 357
    },
    {
      "epoch": 53.333333333333336,
      "grad_norm": 14.928536415100098,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.1006,
      "step": 360
    },
    {
      "epoch": 53.925925925925924,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.2926301658153534,
      "eval_runtime": 2.1238,
      "eval_samples_per_second": 45.203,
      "eval_steps_per_second": 1.413,
      "step": 364
    },
    {
      "epoch": 54.81481481481482,
      "grad_norm": 16.707883834838867,
      "learning_rate": 3.271604938271605e-05,
      "loss": 0.1053,
      "step": 370
    },
    {
      "epoch": 54.96296296296296,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.2791050672531128,
      "eval_runtime": 1.8468,
      "eval_samples_per_second": 51.982,
      "eval_steps_per_second": 1.624,
      "step": 371
    },
    {
      "epoch": 56.0,
      "eval_accuracy": 0.875,
      "eval_loss": 0.3276265859603882,
      "eval_runtime": 1.8761,
      "eval_samples_per_second": 51.17,
      "eval_steps_per_second": 1.599,
      "step": 378
    },
    {
      "epoch": 56.2962962962963,
      "grad_norm": 19.896167755126953,
      "learning_rate": 3.209876543209876e-05,
      "loss": 0.106,
      "step": 380
    },
    {
      "epoch": 56.888888888888886,
      "eval_accuracy": 0.875,
      "eval_loss": 0.32238277792930603,
      "eval_runtime": 1.9352,
      "eval_samples_per_second": 49.608,
      "eval_steps_per_second": 1.55,
      "step": 384
    },
    {
      "epoch": 57.77777777777778,
      "grad_norm": 23.663541793823242,
      "learning_rate": 3.148148148148148e-05,
      "loss": 0.1058,
      "step": 390
    },
    {
      "epoch": 57.925925925925924,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.33849918842315674,
      "eval_runtime": 1.9952,
      "eval_samples_per_second": 48.116,
      "eval_steps_per_second": 1.504,
      "step": 391
    },
    {
      "epoch": 58.96296296296296,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.34936726093292236,
      "eval_runtime": 1.8826,
      "eval_samples_per_second": 50.992,
      "eval_steps_per_second": 1.594,
      "step": 398
    },
    {
      "epoch": 59.25925925925926,
      "grad_norm": 6.200643539428711,
      "learning_rate": 3.08641975308642e-05,
      "loss": 0.0962,
      "step": 400
    },
    {
      "epoch": 60.0,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.2797829508781433,
      "eval_runtime": 1.8593,
      "eval_samples_per_second": 51.633,
      "eval_steps_per_second": 1.614,
      "step": 405
    },
    {
      "epoch": 60.74074074074074,
      "grad_norm": 16.507617950439453,
      "learning_rate": 3.0246913580246916e-05,
      "loss": 0.0883,
      "step": 410
    },
    {
      "epoch": 60.888888888888886,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.29343104362487793,
      "eval_runtime": 1.8755,
      "eval_samples_per_second": 51.187,
      "eval_steps_per_second": 1.6,
      "step": 411
    },
    {
      "epoch": 61.925925925925924,
      "eval_accuracy": 0.875,
      "eval_loss": 0.2956160008907318,
      "eval_runtime": 2.1241,
      "eval_samples_per_second": 45.195,
      "eval_steps_per_second": 1.412,
      "step": 418
    },
    {
      "epoch": 62.22222222222222,
      "grad_norm": 19.266647338867188,
      "learning_rate": 2.962962962962963e-05,
      "loss": 0.084,
      "step": 420
    },
    {
      "epoch": 62.96296296296296,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.291843980550766,
      "eval_runtime": 1.864,
      "eval_samples_per_second": 51.503,
      "eval_steps_per_second": 1.609,
      "step": 425
    },
    {
      "epoch": 63.7037037037037,
      "grad_norm": 27.770219802856445,
      "learning_rate": 2.9012345679012347e-05,
      "loss": 0.0808,
      "step": 430
    },
    {
      "epoch": 64.0,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.3416362702846527,
      "eval_runtime": 1.8651,
      "eval_samples_per_second": 51.471,
      "eval_steps_per_second": 1.608,
      "step": 432
    },
    {
      "epoch": 64.88888888888889,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.35024556517601013,
      "eval_runtime": 1.8463,
      "eval_samples_per_second": 51.995,
      "eval_steps_per_second": 1.625,
      "step": 438
    },
    {
      "epoch": 65.18518518518519,
      "grad_norm": 14.995842933654785,
      "learning_rate": 2.839506172839506e-05,
      "loss": 0.0804,
      "step": 440
    },
    {
      "epoch": 65.92592592592592,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.2984811067581177,
      "eval_runtime": 2.0874,
      "eval_samples_per_second": 45.989,
      "eval_steps_per_second": 1.437,
      "step": 445
    },
    {
      "epoch": 66.66666666666667,
      "grad_norm": 34.63008117675781,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.0854,
      "step": 450
    },
    {
      "epoch": 66.96296296296296,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.2791596055030823,
      "eval_runtime": 1.88,
      "eval_samples_per_second": 51.064,
      "eval_steps_per_second": 1.596,
      "step": 452
    },
    {
      "epoch": 68.0,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.3643587827682495,
      "eval_runtime": 1.8436,
      "eval_samples_per_second": 52.072,
      "eval_steps_per_second": 1.627,
      "step": 459
    },
    {
      "epoch": 68.14814814814815,
      "grad_norm": 32.28718566894531,
      "learning_rate": 2.7160493827160493e-05,
      "loss": 0.0887,
      "step": 460
    },
    {
      "epoch": 68.88888888888889,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.2684231400489807,
      "eval_runtime": 1.8572,
      "eval_samples_per_second": 51.691,
      "eval_steps_per_second": 1.615,
      "step": 465
    },
    {
      "epoch": 69.62962962962963,
      "grad_norm": 22.44804573059082,
      "learning_rate": 2.654320987654321e-05,
      "loss": 0.0671,
      "step": 470
    },
    {
      "epoch": 69.92592592592592,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.28022265434265137,
      "eval_runtime": 2.1506,
      "eval_samples_per_second": 44.638,
      "eval_steps_per_second": 1.395,
      "step": 472
    },
    {
      "epoch": 70.96296296296296,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.2900524437427521,
      "eval_runtime": 1.859,
      "eval_samples_per_second": 51.641,
      "eval_steps_per_second": 1.614,
      "step": 479
    },
    {
      "epoch": 71.11111111111111,
      "grad_norm": 4.379269123077393,
      "learning_rate": 2.5925925925925925e-05,
      "loss": 0.0704,
      "step": 480
    },
    {
      "epoch": 72.0,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.3098466694355011,
      "eval_runtime": 1.838,
      "eval_samples_per_second": 52.23,
      "eval_steps_per_second": 1.632,
      "step": 486
    },
    {
      "epoch": 72.5925925925926,
      "grad_norm": 17.398576736450195,
      "learning_rate": 2.5308641975308646e-05,
      "loss": 0.0802,
      "step": 490
    },
    {
      "epoch": 72.88888888888889,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.2960352897644043,
      "eval_runtime": 1.867,
      "eval_samples_per_second": 51.421,
      "eval_steps_per_second": 1.607,
      "step": 492
    },
    {
      "epoch": 73.92592592592592,
      "eval_accuracy": 0.875,
      "eval_loss": 0.2757139503955841,
      "eval_runtime": 2.0356,
      "eval_samples_per_second": 47.161,
      "eval_steps_per_second": 1.474,
      "step": 499
    },
    {
      "epoch": 74.07407407407408,
      "grad_norm": 18.339569091796875,
      "learning_rate": 2.4691358024691357e-05,
      "loss": 0.09,
      "step": 500
    },
    {
      "epoch": 74.96296296296296,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.31044331192970276,
      "eval_runtime": 1.8732,
      "eval_samples_per_second": 51.249,
      "eval_steps_per_second": 1.602,
      "step": 506
    },
    {
      "epoch": 75.55555555555556,
      "grad_norm": 8.075996398925781,
      "learning_rate": 2.4074074074074074e-05,
      "loss": 0.0772,
      "step": 510
    },
    {
      "epoch": 76.0,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.3120499849319458,
      "eval_runtime": 1.9108,
      "eval_samples_per_second": 50.239,
      "eval_steps_per_second": 1.57,
      "step": 513
    },
    {
      "epoch": 76.88888888888889,
      "eval_accuracy": 0.9166666666666666,
      "eval_loss": 0.2803463637828827,
      "eval_runtime": 1.8689,
      "eval_samples_per_second": 51.368,
      "eval_steps_per_second": 1.605,
      "step": 519
    },
    {
      "epoch": 77.03703703703704,
      "grad_norm": 3.0919692516326904,
      "learning_rate": 2.345679012345679e-05,
      "loss": 0.0725,
      "step": 520
    },
    {
      "epoch": 77.92592592592592,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.28252843022346497,
      "eval_runtime": 1.9144,
      "eval_samples_per_second": 50.146,
      "eval_steps_per_second": 1.567,
      "step": 526
    },
    {
      "epoch": 78.51851851851852,
      "grad_norm": 14.103160858154297,
      "learning_rate": 2.2839506172839506e-05,
      "loss": 0.0684,
      "step": 530
    },
    {
      "epoch": 78.96296296296296,
      "eval_accuracy": 0.875,
      "eval_loss": 0.325451523065567,
      "eval_runtime": 1.8932,
      "eval_samples_per_second": 50.708,
      "eval_steps_per_second": 1.585,
      "step": 533
    },
    {
      "epoch": 80.0,
      "grad_norm": 25.228979110717773,
      "learning_rate": 2.2222222222222223e-05,
      "loss": 0.0732,
      "step": 540
    },
    {
      "epoch": 80.0,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.3091324269771576,
      "eval_runtime": 1.8455,
      "eval_samples_per_second": 52.019,
      "eval_steps_per_second": 1.626,
      "step": 540
    },
    {
      "epoch": 80.88888888888889,
      "eval_accuracy": 0.9166666666666666,
      "eval_loss": 0.2875919044017792,
      "eval_runtime": 1.8878,
      "eval_samples_per_second": 50.852,
      "eval_steps_per_second": 1.589,
      "step": 546
    },
    {
      "epoch": 81.48148148148148,
      "grad_norm": 9.110090255737305,
      "learning_rate": 2.1604938271604937e-05,
      "loss": 0.0743,
      "step": 550
    },
    {
      "epoch": 81.92592592592592,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.30346551537513733,
      "eval_runtime": 2.0498,
      "eval_samples_per_second": 46.834,
      "eval_steps_per_second": 1.464,
      "step": 553
    },
    {
      "epoch": 82.96296296296296,
      "grad_norm": 23.57141876220703,
      "learning_rate": 2.0987654320987655e-05,
      "loss": 0.0807,
      "step": 560
    },
    {
      "epoch": 82.96296296296296,
      "eval_accuracy": 0.9270833333333334,
      "eval_loss": 0.2750629186630249,
      "eval_runtime": 1.8436,
      "eval_samples_per_second": 52.073,
      "eval_steps_per_second": 1.627,
      "step": 560
    },
    {
      "epoch": 84.0,
      "eval_accuracy": 0.9166666666666666,
      "eval_loss": 0.2656916677951813,
      "eval_runtime": 1.884,
      "eval_samples_per_second": 50.955,
      "eval_steps_per_second": 1.592,
      "step": 567
    },
    {
      "epoch": 84.44444444444444,
      "grad_norm": 7.4042277336120605,
      "learning_rate": 2.037037037037037e-05,
      "loss": 0.0799,
      "step": 570
    },
    {
      "epoch": 84.88888888888889,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.2810324728488922,
      "eval_runtime": 1.8651,
      "eval_samples_per_second": 51.472,
      "eval_steps_per_second": 1.608,
      "step": 573
    },
    {
      "epoch": 85.92592592592592,
      "grad_norm": 9.805841445922852,
      "learning_rate": 1.9753086419753087e-05,
      "loss": 0.0632,
      "step": 580
    },
    {
      "epoch": 85.92592592592592,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.303717702627182,
      "eval_runtime": 1.9401,
      "eval_samples_per_second": 49.481,
      "eval_steps_per_second": 1.546,
      "step": 580
    },
    {
      "epoch": 86.96296296296296,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.33565008640289307,
      "eval_runtime": 1.8817,
      "eval_samples_per_second": 51.018,
      "eval_steps_per_second": 1.594,
      "step": 587
    },
    {
      "epoch": 87.4074074074074,
      "grad_norm": 27.298439025878906,
      "learning_rate": 1.91358024691358e-05,
      "loss": 0.0579,
      "step": 590
    },
    {
      "epoch": 88.0,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.3170994818210602,
      "eval_runtime": 1.8439,
      "eval_samples_per_second": 52.063,
      "eval_steps_per_second": 1.627,
      "step": 594
    },
    {
      "epoch": 88.88888888888889,
      "grad_norm": 12.730273246765137,
      "learning_rate": 1.8518518518518518e-05,
      "loss": 0.0593,
      "step": 600
    },
    {
      "epoch": 88.88888888888889,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.32229921221733093,
      "eval_runtime": 1.8672,
      "eval_samples_per_second": 51.414,
      "eval_steps_per_second": 1.607,
      "step": 600
    },
    {
      "epoch": 89.92592592592592,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.29770705103874207,
      "eval_runtime": 1.9546,
      "eval_samples_per_second": 49.115,
      "eval_steps_per_second": 1.535,
      "step": 607
    },
    {
      "epoch": 90.37037037037037,
      "grad_norm": 7.479738712310791,
      "learning_rate": 1.7901234567901236e-05,
      "loss": 0.0418,
      "step": 610
    },
    {
      "epoch": 90.96296296296296,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.3380300998687744,
      "eval_runtime": 1.8381,
      "eval_samples_per_second": 52.228,
      "eval_steps_per_second": 1.632,
      "step": 614
    },
    {
      "epoch": 91.85185185185185,
      "grad_norm": 19.313344955444336,
      "learning_rate": 1.728395061728395e-05,
      "loss": 0.0647,
      "step": 620
    },
    {
      "epoch": 92.0,
      "eval_accuracy": 0.875,
      "eval_loss": 0.28629523515701294,
      "eval_runtime": 1.9227,
      "eval_samples_per_second": 49.928,
      "eval_steps_per_second": 1.56,
      "step": 621
    },
    {
      "epoch": 92.88888888888889,
      "eval_accuracy": 0.9166666666666666,
      "eval_loss": 0.2898975610733032,
      "eval_runtime": 1.8782,
      "eval_samples_per_second": 51.114,
      "eval_steps_per_second": 1.597,
      "step": 627
    },
    {
      "epoch": 93.33333333333333,
      "grad_norm": 11.266192436218262,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.0649,
      "step": 630
    },
    {
      "epoch": 93.92592592592592,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.2853105962276459,
      "eval_runtime": 1.9709,
      "eval_samples_per_second": 48.708,
      "eval_steps_per_second": 1.522,
      "step": 634
    },
    {
      "epoch": 94.81481481481481,
      "grad_norm": 9.165306091308594,
      "learning_rate": 1.604938271604938e-05,
      "loss": 0.0538,
      "step": 640
    },
    {
      "epoch": 94.96296296296296,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.24523399770259857,
      "eval_runtime": 1.9081,
      "eval_samples_per_second": 50.312,
      "eval_steps_per_second": 1.572,
      "step": 641
    },
    {
      "epoch": 96.0,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.2568932771682739,
      "eval_runtime": 1.8246,
      "eval_samples_per_second": 52.614,
      "eval_steps_per_second": 1.644,
      "step": 648
    },
    {
      "epoch": 96.29629629629629,
      "grad_norm": 6.884490489959717,
      "learning_rate": 1.54320987654321e-05,
      "loss": 0.0483,
      "step": 650
    },
    {
      "epoch": 96.88888888888889,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.26870599389076233,
      "eval_runtime": 1.8508,
      "eval_samples_per_second": 51.87,
      "eval_steps_per_second": 1.621,
      "step": 654
    },
    {
      "epoch": 97.77777777777777,
      "grad_norm": 10.973348617553711,
      "learning_rate": 1.4814814814814815e-05,
      "loss": 0.0597,
      "step": 660
    },
    {
      "epoch": 97.92592592592592,
      "eval_accuracy": 0.875,
      "eval_loss": 0.3083449900150299,
      "eval_runtime": 1.991,
      "eval_samples_per_second": 48.217,
      "eval_steps_per_second": 1.507,
      "step": 661
    },
    {
      "epoch": 98.96296296296296,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.2929399311542511,
      "eval_runtime": 1.8639,
      "eval_samples_per_second": 51.505,
      "eval_steps_per_second": 1.61,
      "step": 668
    },
    {
      "epoch": 99.25925925925925,
      "grad_norm": 7.356732368469238,
      "learning_rate": 1.419753086419753e-05,
      "loss": 0.0544,
      "step": 670
    },
    {
      "epoch": 100.0,
      "eval_accuracy": 0.875,
      "eval_loss": 0.32528772950172424,
      "eval_runtime": 1.9688,
      "eval_samples_per_second": 48.76,
      "eval_steps_per_second": 1.524,
      "step": 675
    },
    {
      "epoch": 100.74074074074075,
      "grad_norm": 5.596048831939697,
      "learning_rate": 1.3580246913580247e-05,
      "loss": 0.0585,
      "step": 680
    },
    {
      "epoch": 100.88888888888889,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.33938369154930115,
      "eval_runtime": 1.8565,
      "eval_samples_per_second": 51.711,
      "eval_steps_per_second": 1.616,
      "step": 681
    },
    {
      "epoch": 101.92592592592592,
      "eval_accuracy": 0.8541666666666666,
      "eval_loss": 0.37479040026664734,
      "eval_runtime": 2.0134,
      "eval_samples_per_second": 47.68,
      "eval_steps_per_second": 1.49,
      "step": 688
    },
    {
      "epoch": 102.22222222222223,
      "grad_norm": 13.269804000854492,
      "learning_rate": 1.2962962962962962e-05,
      "loss": 0.0563,
      "step": 690
    },
    {
      "epoch": 102.96296296296296,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.38897252082824707,
      "eval_runtime": 1.8519,
      "eval_samples_per_second": 51.838,
      "eval_steps_per_second": 1.62,
      "step": 695
    },
    {
      "epoch": 103.70370370370371,
      "grad_norm": 16.924348831176758,
      "learning_rate": 1.2345679012345678e-05,
      "loss": 0.059,
      "step": 700
    },
    {
      "epoch": 104.0,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.3460318148136139,
      "eval_runtime": 1.8788,
      "eval_samples_per_second": 51.096,
      "eval_steps_per_second": 1.597,
      "step": 702
    },
    {
      "epoch": 104.88888888888889,
      "eval_accuracy": 0.875,
      "eval_loss": 0.33083054423332214,
      "eval_runtime": 1.9045,
      "eval_samples_per_second": 50.406,
      "eval_steps_per_second": 1.575,
      "step": 708
    },
    {
      "epoch": 105.18518518518519,
      "grad_norm": 34.068790435791016,
      "learning_rate": 1.1728395061728396e-05,
      "loss": 0.0601,
      "step": 710
    },
    {
      "epoch": 105.92592592592592,
      "eval_accuracy": 0.875,
      "eval_loss": 0.3228204846382141,
      "eval_runtime": 1.8931,
      "eval_samples_per_second": 50.711,
      "eval_steps_per_second": 1.585,
      "step": 715
    },
    {
      "epoch": 106.66666666666667,
      "grad_norm": 28.148273468017578,
      "learning_rate": 1.1111111111111112e-05,
      "loss": 0.0512,
      "step": 720
    },
    {
      "epoch": 106.96296296296296,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.3189867436885834,
      "eval_runtime": 1.9115,
      "eval_samples_per_second": 50.224,
      "eval_steps_per_second": 1.569,
      "step": 722
    },
    {
      "epoch": 108.0,
      "eval_accuracy": 0.875,
      "eval_loss": 0.30278849601745605,
      "eval_runtime": 1.8687,
      "eval_samples_per_second": 51.373,
      "eval_steps_per_second": 1.605,
      "step": 729
    },
    {
      "epoch": 108.14814814814815,
      "grad_norm": 8.840377807617188,
      "learning_rate": 1.0493827160493827e-05,
      "loss": 0.0346,
      "step": 730
    },
    {
      "epoch": 108.88888888888889,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.3065723478794098,
      "eval_runtime": 1.849,
      "eval_samples_per_second": 51.919,
      "eval_steps_per_second": 1.622,
      "step": 735
    },
    {
      "epoch": 109.62962962962963,
      "grad_norm": 5.847916126251221,
      "learning_rate": 9.876543209876543e-06,
      "loss": 0.0434,
      "step": 740
    },
    {
      "epoch": 109.92592592592592,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.29524290561676025,
      "eval_runtime": 1.8637,
      "eval_samples_per_second": 51.511,
      "eval_steps_per_second": 1.61,
      "step": 742
    },
    {
      "epoch": 110.96296296296296,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.3053508698940277,
      "eval_runtime": 1.875,
      "eval_samples_per_second": 51.201,
      "eval_steps_per_second": 1.6,
      "step": 749
    },
    {
      "epoch": 111.11111111111111,
      "grad_norm": 13.934989929199219,
      "learning_rate": 9.259259259259259e-06,
      "loss": 0.0466,
      "step": 750
    },
    {
      "epoch": 112.0,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.30871471762657166,
      "eval_runtime": 1.8837,
      "eval_samples_per_second": 50.965,
      "eval_steps_per_second": 1.593,
      "step": 756
    },
    {
      "epoch": 112.5925925925926,
      "grad_norm": 4.37272834777832,
      "learning_rate": 8.641975308641975e-06,
      "loss": 0.0402,
      "step": 760
    },
    {
      "epoch": 112.88888888888889,
      "eval_accuracy": 0.875,
      "eval_loss": 0.3211623728275299,
      "eval_runtime": 1.8685,
      "eval_samples_per_second": 51.377,
      "eval_steps_per_second": 1.606,
      "step": 762
    },
    {
      "epoch": 113.92592592592592,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.3235282897949219,
      "eval_runtime": 1.8591,
      "eval_samples_per_second": 51.637,
      "eval_steps_per_second": 1.614,
      "step": 769
    },
    {
      "epoch": 114.07407407407408,
      "grad_norm": 17.743671417236328,
      "learning_rate": 8.02469135802469e-06,
      "loss": 0.0491,
      "step": 770
    },
    {
      "epoch": 114.96296296296296,
      "eval_accuracy": 0.90625,
      "eval_loss": 0.3134639263153076,
      "eval_runtime": 1.8752,
      "eval_samples_per_second": 51.195,
      "eval_steps_per_second": 1.6,
      "step": 776
    },
    {
      "epoch": 115.55555555555556,
      "grad_norm": 14.649134635925293,
      "learning_rate": 7.4074074074074075e-06,
      "loss": 0.0495,
      "step": 780
    },
    {
      "epoch": 116.0,
      "eval_accuracy": 0.8958333333333334,
      "eval_loss": 0.2991209924221039,
      "eval_runtime": 1.9038,
      "eval_samples_per_second": 50.426,
      "eval_steps_per_second": 1.576,
      "step": 783
    },
    {
      "epoch": 116.88888888888889,
      "eval_accuracy": 0.8854166666666666,
      "eval_loss": 0.3050777018070221,
      "eval_runtime": 1.8271,
      "eval_samples_per_second": 52.541,
      "eval_steps_per_second": 1.642,
      "step": 789
    },
    {
      "epoch": 117.03703703703704,
      "grad_norm": 14.181517601013184,
      "learning_rate": 6.790123456790123e-06,
      "loss": 0.0536,
      "step": 790
    },
    {
      "epoch": 117.92592592592592,
      "eval_accuracy": 0.875,
      "eval_loss": 0.33390840888023376,
      "eval_runtime": 1.8859,
      "eval_samples_per_second": 50.904,
      "eval_steps_per_second": 1.591,
      "step": 796
    },
    {
      "epoch": 118.51851851851852,
      "grad_norm": 13.237271308898926,
      "learning_rate": 6.172839506172839e-06,
      "loss": 0.0419,
      "step": 800
    },
    {
      "epoch": 118.96296296296296,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.3370848596096039,
      "eval_runtime": 1.8605,
      "eval_samples_per_second": 51.6,
      "eval_steps_per_second": 1.612,
      "step": 803
    },
    {
      "epoch": 120.0,
      "grad_norm": 5.592768669128418,
      "learning_rate": 5.555555555555556e-06,
      "loss": 0.0333,
      "step": 810
    },
    {
      "epoch": 120.0,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.3375813066959381,
      "eval_runtime": 1.8801,
      "eval_samples_per_second": 51.06,
      "eval_steps_per_second": 1.596,
      "step": 810
    },
    {
      "epoch": 120.88888888888889,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.33790695667266846,
      "eval_runtime": 1.8455,
      "eval_samples_per_second": 52.018,
      "eval_steps_per_second": 1.626,
      "step": 816
    },
    {
      "epoch": 121.48148148148148,
      "grad_norm": 4.372032165527344,
      "learning_rate": 4.938271604938272e-06,
      "loss": 0.0376,
      "step": 820
    },
    {
      "epoch": 121.92592592592592,
      "eval_accuracy": 0.8541666666666666,
      "eval_loss": 0.337320476770401,
      "eval_runtime": 1.8367,
      "eval_samples_per_second": 52.267,
      "eval_steps_per_second": 1.633,
      "step": 823
    },
    {
      "epoch": 122.96296296296296,
      "grad_norm": 10.392346382141113,
      "learning_rate": 4.3209876543209875e-06,
      "loss": 0.0397,
      "step": 830
    },
    {
      "epoch": 122.96296296296296,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.34366607666015625,
      "eval_runtime": 1.8522,
      "eval_samples_per_second": 51.831,
      "eval_steps_per_second": 1.62,
      "step": 830
    },
    {
      "epoch": 124.0,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.35845816135406494,
      "eval_runtime": 1.8547,
      "eval_samples_per_second": 51.76,
      "eval_steps_per_second": 1.618,
      "step": 837
    },
    {
      "epoch": 124.44444444444444,
      "grad_norm": 5.668208122253418,
      "learning_rate": 3.7037037037037037e-06,
      "loss": 0.0299,
      "step": 840
    },
    {
      "epoch": 124.88888888888889,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.3514343202114105,
      "eval_runtime": 1.8661,
      "eval_samples_per_second": 51.445,
      "eval_steps_per_second": 1.608,
      "step": 843
    },
    {
      "epoch": 125.92592592592592,
      "grad_norm": 21.28321647644043,
      "learning_rate": 3.0864197530864196e-06,
      "loss": 0.0468,
      "step": 850
    },
    {
      "epoch": 125.92592592592592,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.3397478759288788,
      "eval_runtime": 1.8626,
      "eval_samples_per_second": 51.54,
      "eval_steps_per_second": 1.611,
      "step": 850
    },
    {
      "epoch": 126.96296296296296,
      "eval_accuracy": 0.8541666666666666,
      "eval_loss": 0.33162921667099,
      "eval_runtime": 1.9031,
      "eval_samples_per_second": 50.444,
      "eval_steps_per_second": 1.576,
      "step": 857
    },
    {
      "epoch": 127.4074074074074,
      "grad_norm": 2.5001776218414307,
      "learning_rate": 2.469135802469136e-06,
      "loss": 0.0351,
      "step": 860
    },
    {
      "epoch": 128.0,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.33344319462776184,
      "eval_runtime": 1.8441,
      "eval_samples_per_second": 52.058,
      "eval_steps_per_second": 1.627,
      "step": 864
    },
    {
      "epoch": 128.88888888888889,
      "grad_norm": 9.004130363464355,
      "learning_rate": 1.8518518518518519e-06,
      "loss": 0.0439,
      "step": 870
    },
    {
      "epoch": 128.88888888888889,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.3323952555656433,
      "eval_runtime": 1.8475,
      "eval_samples_per_second": 51.963,
      "eval_steps_per_second": 1.624,
      "step": 870
    },
    {
      "epoch": 129.92592592592592,
      "eval_accuracy": 0.8645833333333334,
      "eval_loss": 0.3289691209793091,
      "eval_runtime": 1.8689,
      "eval_samples_per_second": 51.368,
      "eval_steps_per_second": 1.605,
      "step": 877
    },
    {
      "epoch": 130.37037037037038,
      "grad_norm": 12.271383285522461,
      "learning_rate": 1.234567901234568e-06,
      "loss": 0.0478,
      "step": 880
    },
    {
      "epoch": 130.96296296296296,
      "eval_accuracy": 0.875,
      "eval_loss": 0.325620174407959,
      "eval_runtime": 1.8869,
      "eval_samples_per_second": 50.876,
      "eval_steps_per_second": 1.59,
      "step": 884
    },
    {
      "epoch": 131.85185185185185,
      "grad_norm": 8.459063529968262,
      "learning_rate": 6.17283950617284e-07,
      "loss": 0.0434,
      "step": 890
    },
    {
      "epoch": 132.0,
      "eval_accuracy": 0.875,
      "eval_loss": 0.32529863715171814,
      "eval_runtime": 1.8653,
      "eval_samples_per_second": 51.466,
      "eval_steps_per_second": 1.608,
      "step": 891
    },
    {
      "epoch": 132.88888888888889,
      "eval_accuracy": 0.875,
      "eval_loss": 0.325122594833374,
      "eval_runtime": 1.835,
      "eval_samples_per_second": 52.315,
      "eval_steps_per_second": 1.635,
      "step": 897
    },
    {
      "epoch": 133.33333333333334,
      "grad_norm": 9.571310997009277,
      "learning_rate": 0.0,
      "loss": 0.0374,
      "step": 900
    },
    {
      "epoch": 133.33333333333334,
      "eval_accuracy": 0.875,
      "eval_loss": 0.3251601457595825,
      "eval_runtime": 1.8591,
      "eval_samples_per_second": 51.637,
      "eval_steps_per_second": 1.614,
      "step": 900
    },
    {
      "epoch": 133.33333333333334,
      "step": 900,
      "total_flos": 2.900008367869133e+18,
      "train_loss": 0.254776107304626,
      "train_runtime": 3168.008,
      "train_samples_per_second": 40.909,
      "train_steps_per_second": 0.284
    }
  ],
  "logging_steps": 10,
  "max_steps": 900,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 150,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.900008367869133e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}