{
  "best_metric": 0.967032967032967,
  "best_model_checkpoint": "document-crop/checkpoint-72",
  "epoch": 19.310344827586206,
  "eval_steps": 500,
  "global_step": 140,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9655172413793104,
      "eval_accuracy": 0.9120879120879121,
      "eval_loss": 0.3032292127609253,
      "eval_runtime": 2.5391,
      "eval_samples_per_second": 35.839,
      "eval_steps_per_second": 2.363,
      "step": 7
    },
    {
      "epoch": 1.9310344827586206,
      "eval_accuracy": 0.8241758241758241,
      "eval_loss": 0.3624477684497833,
      "eval_runtime": 2.5753,
      "eval_samples_per_second": 35.335,
      "eval_steps_per_second": 2.33,
      "step": 14
    },
    {
      "epoch": 2.0689655172413794,
      "grad_norm": 17.86758804321289,
      "learning_rate": 4.960317460317461e-05,
      "loss": 0.48,
      "step": 15
    },
    {
      "epoch": 2.896551724137931,
      "eval_accuracy": 0.8351648351648352,
      "eval_loss": 0.3741466701030731,
      "eval_runtime": 2.5668,
      "eval_samples_per_second": 35.453,
      "eval_steps_per_second": 2.338,
      "step": 21
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8901098901098901,
      "eval_loss": 0.23425163328647614,
      "eval_runtime": 2.5564,
      "eval_samples_per_second": 35.596,
      "eval_steps_per_second": 2.347,
      "step": 29
    },
    {
      "epoch": 4.137931034482759,
      "grad_norm": 19.562450408935547,
      "learning_rate": 4.3650793650793655e-05,
      "loss": 0.3749,
      "step": 30
    },
    {
      "epoch": 4.9655172413793105,
      "eval_accuracy": 0.9120879120879121,
      "eval_loss": 0.20266202092170715,
      "eval_runtime": 2.6328,
      "eval_samples_per_second": 34.564,
      "eval_steps_per_second": 2.279,
      "step": 36
    },
    {
      "epoch": 5.931034482758621,
      "eval_accuracy": 0.945054945054945,
      "eval_loss": 0.19039569795131683,
      "eval_runtime": 2.6292,
      "eval_samples_per_second": 34.611,
      "eval_steps_per_second": 2.282,
      "step": 43
    },
    {
      "epoch": 6.206896551724138,
      "grad_norm": 13.474059104919434,
      "learning_rate": 3.76984126984127e-05,
      "loss": 0.2323,
      "step": 45
    },
    {
      "epoch": 6.896551724137931,
      "eval_accuracy": 0.9120879120879121,
      "eval_loss": 0.19586102664470673,
      "eval_runtime": 2.6461,
      "eval_samples_per_second": 34.39,
      "eval_steps_per_second": 2.267,
      "step": 50
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8901098901098901,
      "eval_loss": 0.2905385494232178,
      "eval_runtime": 2.6111,
      "eval_samples_per_second": 34.852,
      "eval_steps_per_second": 2.298,
      "step": 58
    },
    {
      "epoch": 8.275862068965518,
      "grad_norm": 24.003192901611328,
      "learning_rate": 3.1746031746031745e-05,
      "loss": 0.1447,
      "step": 60
    },
    {
      "epoch": 8.96551724137931,
      "eval_accuracy": 0.9010989010989011,
      "eval_loss": 0.3998720347881317,
      "eval_runtime": 2.6866,
      "eval_samples_per_second": 33.872,
      "eval_steps_per_second": 2.233,
      "step": 65
    },
    {
      "epoch": 9.931034482758621,
      "eval_accuracy": 0.967032967032967,
      "eval_loss": 0.10719592869281769,
      "eval_runtime": 2.6592,
      "eval_samples_per_second": 34.22,
      "eval_steps_per_second": 2.256,
      "step": 72
    },
    {
      "epoch": 10.344827586206897,
      "grad_norm": 12.631780624389648,
      "learning_rate": 2.5793650793650796e-05,
      "loss": 0.125,
      "step": 75
    },
    {
      "epoch": 10.89655172413793,
      "eval_accuracy": 0.9010989010989011,
      "eval_loss": 0.2653524875640869,
      "eval_runtime": 2.6798,
      "eval_samples_per_second": 33.958,
      "eval_steps_per_second": 2.239,
      "step": 79
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.945054945054945,
      "eval_loss": 0.17991583049297333,
      "eval_runtime": 2.7033,
      "eval_samples_per_second": 33.663,
      "eval_steps_per_second": 2.22,
      "step": 87
    },
    {
      "epoch": 12.413793103448276,
      "grad_norm": 22.67920684814453,
      "learning_rate": 1.984126984126984e-05,
      "loss": 0.1979,
      "step": 90
    },
    {
      "epoch": 12.96551724137931,
      "eval_accuracy": 0.9120879120879121,
      "eval_loss": 0.25460970401763916,
      "eval_runtime": 2.605,
      "eval_samples_per_second": 34.933,
      "eval_steps_per_second": 2.303,
      "step": 94
    },
    {
      "epoch": 13.931034482758621,
      "eval_accuracy": 0.9010989010989011,
      "eval_loss": 0.25756293535232544,
      "eval_runtime": 2.6309,
      "eval_samples_per_second": 34.589,
      "eval_steps_per_second": 2.281,
      "step": 101
    },
    {
      "epoch": 14.482758620689655,
      "grad_norm": 3.1801395416259766,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.0633,
      "step": 105
    },
    {
      "epoch": 14.89655172413793,
      "eval_accuracy": 0.9340659340659341,
      "eval_loss": 0.1996360421180725,
      "eval_runtime": 2.7759,
      "eval_samples_per_second": 32.782,
      "eval_steps_per_second": 2.161,
      "step": 108
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9560439560439561,
      "eval_loss": 0.1824313998222351,
      "eval_runtime": 2.6954,
      "eval_samples_per_second": 33.762,
      "eval_steps_per_second": 2.226,
      "step": 116
    },
    {
      "epoch": 16.551724137931036,
      "grad_norm": 0.8367649912834167,
      "learning_rate": 7.936507936507936e-06,
      "loss": 0.0311,
      "step": 120
    },
    {
      "epoch": 16.96551724137931,
      "eval_accuracy": 0.9560439560439561,
      "eval_loss": 0.18344764411449432,
      "eval_runtime": 2.7125,
      "eval_samples_per_second": 33.549,
      "eval_steps_per_second": 2.212,
      "step": 123
    },
    {
      "epoch": 17.93103448275862,
      "eval_accuracy": 0.9230769230769231,
      "eval_loss": 0.2769739627838135,
      "eval_runtime": 2.693,
      "eval_samples_per_second": 33.791,
      "eval_steps_per_second": 2.228,
      "step": 130
    },
    {
      "epoch": 18.620689655172413,
      "grad_norm": 0.14054416120052338,
      "learning_rate": 1.984126984126984e-06,
      "loss": 0.0154,
      "step": 135
    },
    {
      "epoch": 18.896551724137932,
      "eval_accuracy": 0.945054945054945,
      "eval_loss": 0.2113342434167862,
      "eval_runtime": 2.672,
      "eval_samples_per_second": 34.056,
      "eval_steps_per_second": 2.245,
      "step": 137
    },
    {
      "epoch": 19.310344827586206,
      "eval_accuracy": 0.9340659340659341,
      "eval_loss": 0.20573703944683075,
      "eval_runtime": 2.6636,
      "eval_samples_per_second": 34.164,
      "eval_steps_per_second": 2.253,
      "step": 140
    },
    {
      "epoch": 19.310344827586206,
      "step": 140,
      "total_flos": 6.939281828084982e+17,
      "train_loss": 0.17856683683182512,
      "train_runtime": 404.7192,
      "train_samples_per_second": 22.485,
      "train_steps_per_second": 0.346
    }
  ],
  "logging_steps": 15,
  "max_steps": 140,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 6.939281828084982e+17,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}