{
  "best_metric": 0.9208715596330275,
  "best_model_checkpoint": "electra-distilled-sst\\run-13\\checkpoint-1581",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 2108,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 22.835023880004883,
      "learning_rate": 4.3163581221778226e-05,
      "loss": 0.8573,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9162844036697247,
      "eval_loss": 0.5491362810134888,
      "eval_runtime": 1.4473,
      "eval_samples_per_second": 602.517,
      "eval_steps_per_second": 4.837,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 38.72591018676758,
      "learning_rate": 3.7769413020716535e-05,
      "loss": 0.4496,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9105504587155964,
      "eval_loss": 0.5399537682533264,
      "eval_runtime": 1.4499,
      "eval_samples_per_second": 601.429,
      "eval_steps_per_second": 4.828,
      "step": 1054
    },
    {
      "epoch": 3.0,
      "grad_norm": 6.218717575073242,
      "learning_rate": 3.237524481965485e-05,
      "loss": 0.3331,
      "step": 1581
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9208715596330275,
      "eval_loss": 0.5268839001655579,
      "eval_runtime": 1.4424,
      "eval_samples_per_second": 604.54,
      "eval_steps_per_second": 4.853,
      "step": 1581
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.019476890563965,
      "learning_rate": 2.6981076618593168e-05,
      "loss": 0.2593,
      "step": 2108
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9174311926605505,
      "eval_loss": 0.5509608387947083,
      "eval_runtime": 1.453,
      "eval_samples_per_second": 600.123,
      "eval_steps_per_second": 4.818,
      "step": 2108
    }
  ],
  "logging_steps": 500,
  "max_steps": 4743,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 500,
  "total_flos": 748818980000436.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8585651118406669,
    "learning_rate": 4.854751380955516e-05,
    "num_train_epochs": 9,
    "temperature": 23
  }
}