{
  "best_metric": 0.7457527826596367,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-28/checkpoint-642",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 642,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 4.335964679718018,
      "learning_rate": 0.0004134351696083348,
      "loss": 0.5585,
      "step": 214
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7246631517281781,
      "eval_f1": 0.41831683168316836,
      "eval_loss": 0.5304129123687744,
      "eval_mcc": 0.31993210092179336,
      "eval_precision": 0.7071129707112971,
      "eval_recall": 0.29701230228471004,
      "eval_runtime": 3.165,
      "eval_samples_per_second": 539.333,
      "eval_steps_per_second": 17.061,
      "step": 214
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.162600994110107,
      "learning_rate": 0.00035437300252142985,
      "loss": 0.5111,
      "step": 428
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7375512595196251,
      "eval_f1": 0.5400410677618069,
      "eval_loss": 0.4986778795719147,
      "eval_mcc": 0.37392389834639567,
      "eval_precision": 0.6493827160493827,
      "eval_recall": 0.46221441124780316,
      "eval_runtime": 3.2765,
      "eval_samples_per_second": 520.988,
      "eval_steps_per_second": 16.481,
      "step": 428
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.091866970062256,
      "learning_rate": 0.00029531083543452485,
      "loss": 0.4873,
      "step": 642
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7457527826596367,
      "eval_f1": 0.5363247863247863,
      "eval_loss": 0.5050870776176453,
      "eval_mcc": 0.38921258590221086,
      "eval_precision": 0.6839237057220708,
      "eval_recall": 0.44112478031634444,
      "eval_runtime": 3.1601,
      "eval_samples_per_second": 540.17,
      "eval_steps_per_second": 17.088,
      "step": 642
    }
  ],
  "logging_steps": 500,
  "max_steps": 1712,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 1574326993320.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9125783768108973,
    "learning_rate": 0.0004724973366952398,
    "num_train_epochs": 8,
    "temperature": 13
  }
}