run-42/checkpoint-48/trainer_state.json

{
  "best_metric": 0.7553816046966731,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-42/checkpoint-48",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 48,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.022854804992676,
      "learning_rate": 0.000339891740397472,
      "loss": 0.5757,
      "step": 48
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7553816046966731,
      "eval_f1": 0.795751633986928,
      "eval_loss": 0.4926829934120178,
      "eval_precision": 0.6830294530154277,
      "eval_recall": 0.9530332681017613,
      "eval_runtime": 27.49,
      "eval_samples_per_second": 37.177,
      "eval_steps_per_second": 0.582,
      "step": 48
    }
  ],
  "logging_steps": 500,
  "max_steps": 144,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 235695197280.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9597520656622972,
    "learning_rate": 0.000509837610596208,
    "num_train_epochs": 3,
    "temperature": 28
  }
}