run-23/checkpoint-160/trainer_state.json

{
  "best_metric": 0.6666666666666666,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-23/checkpoint-160",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 160,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.446049451828003,
      "learning_rate": 0.0005691296983851737,
      "loss": 0.6049,
      "step": 160
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6666666666666666,
      "eval_f1": 0.0,
      "eval_loss": 0.5720294713973999,
      "eval_mcc": 0.0,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 1.9517,
      "eval_samples_per_second": 653.261,
      "eval_steps_per_second": 20.494,
      "step": 160
    }
  ],
  "logging_steps": 500,
  "max_steps": 1120,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 291755437920.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9539104497371855,
    "learning_rate": 0.000663984648116036,
    "num_train_epochs": 7,
    "temperature": 6
  }
}