{
  "best_global_step": 44,
  "best_metric": 0.061493076384067535,
  "best_model_checkpoint": "rewrite_results/checkpoint-42",
  "epoch": 4.0,
  "eval_steps": 2,
  "global_step": 56,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.14814814814814814,
      "eval_accuracy": 0.9610916971862766,
      "eval_loss": 0.24771852791309357,
      "eval_runtime": 332.7057,
      "eval_samples_per_second": 0.481,
      "eval_steps_per_second": 0.06,
      "step": 2
    },
    {
      "epoch": 0.2222222222222222,
      "grad_norm": 0.10552783565151713,
      "learning_rate": 0.00027304247779439414,
      "loss": 0.2441,
      "step": 3
    },
    {
      "epoch": 0.2962962962962963,
      "eval_accuracy": 0.9648183236107929,
      "eval_loss": 0.1996598243713379,
      "eval_runtime": 327.4272,
      "eval_samples_per_second": 0.489,
      "eval_steps_per_second": 0.061,
      "step": 4
    },
    {
      "epoch": 0.4444444444444444,
      "grad_norm": 0.08840417088618377,
      "learning_rate": 0.0004,
      "loss": 0.2367,
      "step": 6
    },
    {
      "epoch": 0.4444444444444444,
      "eval_accuracy": 0.9687810581276459,
      "eval_loss": 0.15869076550006866,
      "eval_runtime": 326.6521,
      "eval_samples_per_second": 0.49,
      "eval_steps_per_second": 0.061,
      "step": 6
    },
    {
      "epoch": 0.5925925925925926,
      "eval_accuracy": 0.9710204154483826,
      "eval_loss": 0.14757852256298065,
      "eval_runtime": 330.7473,
      "eval_samples_per_second": 0.484,
      "eval_steps_per_second": 0.06,
      "step": 8
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 0.04166015387300161,
      "learning_rate": 0.00037647058823529414,
      "loss": 0.1895,
      "step": 9
    },
    {
      "epoch": 0.7407407407407407,
      "eval_accuracy": 0.9732184664675948,
      "eval_loss": 0.13183145225048065,
      "eval_runtime": 325.6161,
      "eval_samples_per_second": 0.491,
      "eval_steps_per_second": 0.061,
      "step": 10
    },
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 0.026815989165312845,
      "learning_rate": 0.00035294117647058826,
      "loss": 0.1361,
      "step": 12
    },
    {
      "epoch": 0.8888888888888888,
      "eval_accuracy": 0.9759430638638582,
      "eval_loss": 0.11719267070293427,
      "eval_runtime": 323.9433,
      "eval_samples_per_second": 0.494,
      "eval_steps_per_second": 0.062,
      "step": 12
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9783447362046809,
      "eval_loss": 0.1052589863538742,
      "eval_runtime": 320.7346,
      "eval_samples_per_second": 0.499,
      "eval_steps_per_second": 0.062,
      "step": 14
    },
    {
      "epoch": 1.074074074074074,
      "grad_norm": 0.027111063904139955,
      "learning_rate": 0.0003294117647058824,
      "loss": 0.18,
      "step": 15
    },
    {
      "epoch": 1.1481481481481481,
      "eval_accuracy": 0.9792185262272917,
      "eval_loss": 0.09851272404193878,
      "eval_runtime": 329.3754,
      "eval_samples_per_second": 0.486,
      "eval_steps_per_second": 0.061,
      "step": 16
    },
    {
      "epoch": 1.2962962962962963,
      "grad_norm": 0.02728080298065729,
      "learning_rate": 0.00030588235294117644,
      "loss": 0.1193,
      "step": 18
    },
    {
      "epoch": 1.2962962962962963,
      "eval_accuracy": 0.9797502878582046,
      "eval_loss": 0.09319568425416946,
      "eval_runtime": 327.3319,
      "eval_samples_per_second": 0.489,
      "eval_steps_per_second": 0.061,
      "step": 18
    },
    {
      "epoch": 1.4444444444444444,
      "eval_accuracy": 0.9803937999427419,
      "eval_loss": 0.08752243220806122,
      "eval_runtime": 325.6299,
      "eval_samples_per_second": 0.491,
      "eval_steps_per_second": 0.061,
      "step": 20
    },
    {
      "epoch": 1.5185185185185186,
      "grad_norm": 0.024562359210379074,
      "learning_rate": 0.0002823529411764706,
      "loss": 0.0823,
      "step": 21
    },
    {
      "epoch": 1.5925925925925926,
      "eval_accuracy": 0.980617014762694,
      "eval_loss": 0.08400166034698486,
      "eval_runtime": 330.0542,
      "eval_samples_per_second": 0.485,
      "eval_steps_per_second": 0.061,
      "step": 22
    },
    {
      "epoch": 1.7407407407407407,
      "grad_norm": 0.04720330319206125,
      "learning_rate": 0.00025882352941176474,
      "loss": 0.1175,
      "step": 24
    },
    {
      "epoch": 1.7407407407407407,
      "eval_accuracy": 0.9813755814844891,
      "eval_loss": 0.07777883112430573,
      "eval_runtime": 330.5389,
      "eval_samples_per_second": 0.484,
      "eval_steps_per_second": 0.061,
      "step": 24
    },
    {
      "epoch": 1.8888888888888888,
      "eval_accuracy": 0.9826866606161857,
      "eval_loss": 0.07366437464952469,
      "eval_runtime": 330.0141,
      "eval_samples_per_second": 0.485,
      "eval_steps_per_second": 0.061,
      "step": 26
    },
    {
      "epoch": 1.9629629629629628,
      "grad_norm": 0.05091774794908078,
      "learning_rate": 0.00023529411764705883,
      "loss": 0.0898,
      "step": 27
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9836173653422126,
      "eval_loss": 0.07038821280002594,
      "eval_runtime": 330.8638,
      "eval_samples_per_second": 0.484,
      "eval_steps_per_second": 0.06,
      "step": 28
    },
    {
      "epoch": 2.148148148148148,
      "grad_norm": 0.02332212687636519,
      "learning_rate": 0.00021176470588235295,
      "loss": 0.0948,
      "step": 30
    },
    {
      "epoch": 2.148148148148148,
      "eval_accuracy": 0.9838456743046095,
      "eval_loss": 0.06887274235486984,
      "eval_runtime": 334.1801,
      "eval_samples_per_second": 0.479,
      "eval_steps_per_second": 0.06,
      "step": 30
    },
    {
      "epoch": 2.2962962962962963,
      "eval_accuracy": 0.9841358041079982,
      "eval_loss": 0.06697963178157806,
      "eval_runtime": 327.8238,
      "eval_samples_per_second": 0.488,
      "eval_steps_per_second": 0.061,
      "step": 32
    },
    {
      "epoch": 2.3703703703703702,
      "grad_norm": 0.023466685722830175,
      "learning_rate": 0.00018823529411764707,
      "loss": 0.0739,
      "step": 33
    },
    {
      "epoch": 2.4444444444444446,
      "eval_accuracy": 0.9841183559501172,
      "eval_loss": 0.06529785692691803,
      "eval_runtime": 324.9504,
      "eval_samples_per_second": 0.492,
      "eval_steps_per_second": 0.062,
      "step": 34
    },
    {
      "epoch": 2.5925925925925926,
      "grad_norm": 0.02499598734491485,
      "learning_rate": 0.0001647058823529412,
      "loss": 0.0526,
      "step": 36
    },
    {
      "epoch": 2.5925925925925926,
      "eval_accuracy": 0.984462930071049,
      "eval_loss": 0.06430461257696152,
      "eval_runtime": 328.33,
      "eval_samples_per_second": 0.487,
      "eval_steps_per_second": 0.061,
      "step": 36
    },
    {
      "epoch": 2.7407407407407405,
      "eval_accuracy": 0.9845490069370237,
      "eval_loss": 0.06337571144104004,
      "eval_runtime": 323.0777,
      "eval_samples_per_second": 0.495,
      "eval_steps_per_second": 0.062,
      "step": 38
    },
    {
      "epoch": 2.814814814814815,
      "grad_norm": 0.02235923851811797,
      "learning_rate": 0.0001411764705882353,
      "loss": 0.0564,
      "step": 39
    },
    {
      "epoch": 2.888888888888889,
      "eval_accuracy": 0.9847284644961529,
      "eval_loss": 0.062485575675964355,
      "eval_runtime": 328.1316,
      "eval_samples_per_second": 0.488,
      "eval_steps_per_second": 0.061,
      "step": 40
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.03433345990626862,
      "learning_rate": 0.00011764705882352942,
      "loss": 0.0678,
      "step": 42
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9849732649347395,
      "eval_loss": 0.061602283269166946,
      "eval_runtime": 301.0464,
      "eval_samples_per_second": 0.531,
      "eval_steps_per_second": 0.066,
      "step": 42
    },
    {
      "epoch": 3.148148148148148,
      "eval_accuracy": 0.9852119445846472,
      "eval_loss": 0.061493076384067535,
      "eval_runtime": 331.7193,
      "eval_samples_per_second": 0.482,
      "eval_steps_per_second": 0.06,
      "step": 44
    },
    {
      "epoch": 3.2222222222222223,
      "grad_norm": 0.022600132657406128,
      "learning_rate": 9.411764705882353e-05,
      "loss": 0.0499,
      "step": 45
    },
    {
      "epoch": 3.2962962962962963,
      "eval_accuracy": 0.9853399724949231,
      "eval_loss": 0.06156101077795029,
      "eval_runtime": 327.3955,
      "eval_samples_per_second": 0.489,
      "eval_steps_per_second": 0.061,
      "step": 46
    },
    {
      "epoch": 3.4444444444444446,
      "grad_norm": 0.018635515413178816,
      "learning_rate": 7.058823529411765e-05,
      "loss": 0.0437,
      "step": 48
    },
    {
      "epoch": 3.4444444444444446,
      "eval_accuracy": 0.9852882434474587,
      "eval_loss": 0.06198165938258171,
      "eval_runtime": 335.6061,
      "eval_samples_per_second": 0.477,
      "eval_steps_per_second": 0.06,
      "step": 48
    },
    {
      "epoch": 3.5925925925925926,
      "eval_accuracy": 0.9850508846328356,
      "eval_loss": 0.06207697466015816,
      "eval_runtime": 324.7443,
      "eval_samples_per_second": 0.493,
      "eval_steps_per_second": 0.062,
      "step": 50
    },
    {
      "epoch": 3.6666666666666665,
      "grad_norm": 0.021001664364668726,
      "learning_rate": 4.705882352941177e-05,
      "loss": 0.0421,
      "step": 51
    },
    {
      "epoch": 3.7407407407407405,
      "eval_accuracy": 0.9851825126851306,
      "eval_loss": 0.06235222890973091,
      "eval_runtime": 330.039,
      "eval_samples_per_second": 0.485,
      "eval_steps_per_second": 0.061,
      "step": 52
    },
    {
      "epoch": 3.888888888888889,
      "grad_norm": 0.02399059372934013,
      "learning_rate": 2.3529411764705884e-05,
      "loss": 0.0557,
      "step": 54
    },
    {
      "epoch": 3.888888888888889,
      "eval_accuracy": 0.985193215099636,
      "eval_loss": 0.06244580075144768,
      "eval_runtime": 328.4431,
      "eval_samples_per_second": 0.487,
      "eval_steps_per_second": 0.061,
      "step": 54
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9850785367286026,
      "eval_loss": 0.062325846403837204,
      "eval_runtime": 334.0117,
      "eval_samples_per_second": 0.479,
      "eval_steps_per_second": 0.06,
      "step": 56
    },
    {
      "epoch": 4.0,
      "step": 56,
      "total_flos": 83324244590592.0,
      "train_loss": 0.10440204206055828,
      "train_runtime": 17139.0998,
      "train_samples_per_second": 0.101,
      "train_steps_per_second": 0.003
    }
  ],
  "logging_steps": 3,
  "max_steps": 56,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 6,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 83324244590592.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}