{
  "best_metric": 0.013712849373137744,
  "best_model_checkpoint": "distilbert-base-uncased-finetuned-cola/run-3/checkpoint-856",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 1070,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 0.6547805666923523,
      "eval_matthews_correlation": -0.017704091110815466,
      "eval_runtime": 0.9942,
      "eval_samples_per_second": 1049.128,
      "eval_steps_per_second": 66.388,
      "step": 214
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.6239611506462097,
      "eval_matthews_correlation": -0.04188001855089693,
      "eval_runtime": 1.0297,
      "eval_samples_per_second": 1012.907,
      "eval_steps_per_second": 64.096,
      "step": 428
    },
    {
      "epoch": 2.34,
      "grad_norm": 4.167867660522461,
      "learning_rate": 6.747447840087923e-07,
      "loss": 0.6494,
      "step": 500
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.6182331442832947,
      "eval_matthews_correlation": -0.01117597298795153,
      "eval_runtime": 1.005,
      "eval_samples_per_second": 1037.811,
      "eval_steps_per_second": 65.672,
      "step": 642
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.6167928576469421,
      "eval_matthews_correlation": 0.013712849373137744,
      "eval_runtime": 0.9611,
      "eval_samples_per_second": 1085.204,
      "eval_steps_per_second": 68.671,
      "step": 856
    },
    {
      "epoch": 4.67,
      "grad_norm": 4.26352596282959,
      "learning_rate": 8.286339452739555e-08,
      "loss": 0.5956,
      "step": 1000
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.6157423257827759,
      "eval_matthews_correlation": -0.007887379670285008,
      "eval_runtime": 0.9724,
      "eval_samples_per_second": 1072.634,
      "eval_steps_per_second": 67.875,
      "step": 1070
    }
  ],
  "logging_steps": 500,
  "max_steps": 1070,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 16570584925236.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 1.2666261734901892e-06,
    "num_train_epochs": 5,
    "per_device_train_batch_size": 4,
    "seed": 6
  }
}