{
  "best_metric": 0.8968592477704537,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-39/checkpoint-1290",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 1935,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.5754092931747437,
      "learning_rate": 0.0002792573994181373,
      "loss": 0.2921,
      "step": 645
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.893951143854207,
      "eval_f1": 0.21968616262482166,
      "eval_loss": 0.2762198746204376,
      "eval_mcc": 0.24469253589704112,
      "eval_precision": 0.5833333333333334,
      "eval_recall": 0.13532513181019332,
      "eval_runtime": 9.9239,
      "eval_samples_per_second": 519.755,
      "eval_steps_per_second": 16.324,
      "step": 645
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.5258708000183105,
      "learning_rate": 0.0002482287994827887,
      "loss": 0.2756,
      "step": 1290
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8968592477704537,
      "eval_f1": 0.18153846153846154,
      "eval_loss": 0.2713065445423126,
      "eval_mcc": 0.24920140552364634,
      "eval_precision": 0.7283950617283951,
      "eval_recall": 0.10369068541300527,
      "eval_runtime": 9.9253,
      "eval_samples_per_second": 519.679,
      "eval_steps_per_second": 16.322,
      "step": 1290
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.9706701040267944,
      "learning_rate": 0.0002172001995474401,
      "loss": 0.2683,
      "step": 1935
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8962776269872044,
      "eval_f1": 0.1272430668841762,
      "eval_loss": 0.26737910509109497,
      "eval_mcc": 0.2297743836439136,
      "eval_precision": 0.8863636363636364,
      "eval_recall": 0.06854130052724078,
      "eval_runtime": 9.9187,
      "eval_samples_per_second": 520.03,
      "eval_steps_per_second": 16.333,
      "step": 1935
    }
  ],
  "logging_steps": 500,
  "max_steps": 6450,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 4760827597800.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8141838638911323,
    "learning_rate": 0.00031028599935348586,
    "num_train_epochs": 10,
    "temperature": 13
  }
}