run-8/checkpoint-766/trainer_state.json

{
  "best_metric": 0.6193737769080235,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-8/checkpoint-766",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 766,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.324540138244629,
      "learning_rate": 5.804695204967391e-05,
      "loss": 0.357,
      "step": 766
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6193737769080235,
      "eval_f1": 0.45133991537376583,
      "eval_loss": 0.33073100447654724,
      "eval_precision": 0.8080808080808081,
      "eval_recall": 0.3131115459882583,
      "eval_runtime": 132.7451,
      "eval_samples_per_second": 7.699,
      "eval_steps_per_second": 1.929,
      "step": 766
    }
  ],
  "logging_steps": 500,
  "max_steps": 2298,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 235695197280.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.49392332530271654,
    "learning_rate": 8.707042807451088e-05,
    "num_train_epochs": 3,
    "temperature": 24
  }
}