{
  "best_metric": 0.5168316831683168,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-3/checkpoint-297",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 891,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.8371890187263489,
      "learning_rate": 0.000508582557644255,
      "loss": 0.309,
      "step": 297
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5168316831683168,
      "eval_f1": 0.1643835616438356,
      "eval_loss": 0.30045250058174133,
      "eval_mcc": 0.059239742807176775,
      "eval_precision": 0.6,
      "eval_recall": 0.09523809523809523,
      "eval_runtime": 0.9336,
      "eval_samples_per_second": 540.929,
      "eval_steps_per_second": 17.138,
      "step": 297
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.21939373016357422,
      "learning_rate": 0.0002542912788221275,
      "loss": 0.3016,
      "step": 594
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5128712871287129,
      "eval_f1": 0.14583333333333331,
      "eval_loss": 0.29779767990112305,
      "eval_mcc": 0.046724423554504804,
      "eval_precision": 0.5833333333333334,
      "eval_recall": 0.08333333333333333,
      "eval_runtime": 0.9339,
      "eval_samples_per_second": 540.765,
      "eval_steps_per_second": 17.133,
      "step": 594
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.4567031264305115,
      "learning_rate": 0.0,
      "loss": 0.2962,
      "step": 891
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5128712871287129,
      "eval_f1": 0.13986013986013984,
      "eval_loss": 0.293793261051178,
      "eval_mcc": 0.047945551909639166,
      "eval_precision": 0.5882352941176471,
      "eval_recall": 0.07936507936507936,
      "eval_runtime": 0.9343,
      "eval_samples_per_second": 540.508,
      "eval_steps_per_second": 17.125,
      "step": 891
    }
  ],
  "logging_steps": 500,
  "max_steps": 891,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 2192103797940.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.37059488996882817,
    "learning_rate": 0.0007628738364663827,
    "num_train_epochs": 3,
    "temperature": 5
  }
}