{
  "best_metric": 0.6877562975981254,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-5/checkpoint-428",
  "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": 3.334259033203125,
      "learning_rate": 0.0007289392500735427,
      "loss": 0.5536,
      "step": 214
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_f1": 0.0,
      "eval_loss": 0.5510665774345398,
      "eval_mcc": 0.0,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 3.1386,
      "eval_samples_per_second": 543.88,
      "eval_steps_per_second": 17.205,
      "step": 214
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.9964511394500732,
      "learning_rate": 0.0006074493750612857,
      "loss": 0.5347,
      "step": 428
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6877562975981254,
      "eval_f1": 0.17873651771956856,
      "eval_loss": 0.5295280814170837,
      "eval_mcc": 0.18423582600040866,
      "eval_precision": 0.725,
      "eval_recall": 0.10193321616871705,
      "eval_runtime": 3.1376,
      "eval_samples_per_second": 544.038,
      "eval_steps_per_second": 17.21,
      "step": 428
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.638342022895813,
      "learning_rate": 0.0004859595000490285,
      "loss": 0.5214,
      "step": 642
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6736965436438196,
      "eval_f1": 0.08837970540098199,
      "eval_loss": 0.5322084426879883,
      "eval_mcc": 0.10428403940757605,
      "eval_precision": 0.6428571428571429,
      "eval_recall": 0.04745166959578207,
      "eval_runtime": 3.131,
      "eval_samples_per_second": 545.187,
      "eval_steps_per_second": 17.247,
      "step": 642
    }
  ],
  "logging_steps": 500,
  "max_steps": 1498,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 1574326993320.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8578412183256148,
    "learning_rate": 0.0008504291250857999,
    "num_train_epochs": 7,
    "temperature": 7
  }
}