run-24/checkpoint-1054/trainer_state.json

{
  "best_metric": 0.9059633027522935,
  "best_model_checkpoint": "electra-distilled-sst\\run-24\\checkpoint-1054",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 1054,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 29.21723175048828,
      "learning_rate": 2.957608242808066e-05,
      "loss": 1.7353,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8967889908256881,
      "eval_loss": 0.9055101871490479,
      "eval_runtime": 1.4428,
      "eval_samples_per_second": 604.362,
      "eval_steps_per_second": 4.852,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 32.18894958496094,
      "learning_rate": 2.5360088022343416e-05,
      "loss": 0.855,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9059633027522935,
      "eval_loss": 0.8889522552490234,
      "eval_runtime": 1.4504,
      "eval_samples_per_second": 601.197,
      "eval_steps_per_second": 4.826,
      "step": 1054
    }
  ],
  "logging_steps": 500,
  "max_steps": 4216,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 373776249469380.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.6600028389648191,
    "learning_rate": 3.3792076833817904e-05,
    "num_train_epochs": 8,
    "temperature": 21
  }
}