run-4/checkpoint-856/trainer_state.json

{
  "best_metric": 0.7065026362038664,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-4/checkpoint-856",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 856,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 5.794419765472412,
      "learning_rate": 3.974483565917368e-05,
      "loss": 0.5313,
      "step": 214
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_f1": 0.0035026269702276708,
      "eval_loss": 0.49643149971961975,
      "eval_mcc": 0.01210898699241207,
      "eval_precision": 0.5,
      "eval_recall": 0.0017574692442882249,
      "eval_runtime": 3.1332,
      "eval_samples_per_second": 544.813,
      "eval_steps_per_second": 17.235,
      "step": 214
    },
    {
      "epoch": 2.0,
      "grad_norm": 3.522770643234253,
      "learning_rate": 3.3120696382644734e-05,
      "loss": 0.493,
      "step": 428
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7047451669595782,
      "eval_f1": 0.48884381338742394,
      "eval_loss": 0.48239898681640625,
      "eval_mcc": 0.2950146249100624,
      "eval_precision": 0.5779376498800959,
      "eval_recall": 0.4235500878734622,
      "eval_runtime": 3.1182,
      "eval_samples_per_second": 547.428,
      "eval_steps_per_second": 17.318,
      "step": 428
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.776003122329712,
      "learning_rate": 2.6496557106115785e-05,
      "loss": 0.4796,
      "step": 642
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7012302284710018,
      "eval_f1": 0.45278969957081544,
      "eval_loss": 0.47861993312835693,
      "eval_mcc": 0.2733353610467977,
      "eval_precision": 0.581267217630854,
      "eval_recall": 0.37082601054481545,
      "eval_runtime": 3.1521,
      "eval_samples_per_second": 541.541,
      "eval_steps_per_second": 17.131,
      "step": 642
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.881204843521118,
      "learning_rate": 1.987241782958684e-05,
      "loss": 0.4758,
      "step": 856
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7065026362038664,
      "eval_f1": 0.47318611987381703,
      "eval_loss": 0.47547900676727295,
      "eval_mcc": 0.29121450441681707,
      "eval_precision": 0.5890052356020943,
      "eval_recall": 0.3954305799648506,
      "eval_runtime": 3.1289,
      "eval_samples_per_second": 545.555,
      "eval_steps_per_second": 17.258,
      "step": 856
    }
  ],
  "logging_steps": 500,
  "max_steps": 1498,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 2099102657760.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8274705871724447,
    "learning_rate": 4.6368974935702624e-05,
    "num_train_epochs": 7,
    "temperature": 26
  }
}