{
  "best_metric": 0.9208715596330275,
  "best_model_checkpoint": "electra-distilled-sst\\run-48\\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": 8.276034355163574,
      "learning_rate": 4.888681242124731e-05,
      "loss": 0.3265,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9036697247706422,
      "eval_loss": 0.2659453749656677,
      "eval_runtime": 1.4511,
      "eval_samples_per_second": 600.924,
      "eval_steps_per_second": 4.824,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 7.019628524780273,
      "learning_rate": 3.6665109315935477e-05,
      "loss": 0.1866,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9059633027522935,
      "eval_loss": 0.29601141810417175,
      "eval_runtime": 1.4476,
      "eval_samples_per_second": 602.36,
      "eval_steps_per_second": 4.835,
      "step": 1054
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.9231150150299072,
      "learning_rate": 2.4443406210623653e-05,
      "loss": 0.1398,
      "step": 1581
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9208715596330275,
      "eval_loss": 0.27852392196655273,
      "eval_runtime": 1.4489,
      "eval_samples_per_second": 601.827,
      "eval_steps_per_second": 4.831,
      "step": 1581
    },
    {
      "epoch": 4.0,
      "grad_norm": 6.364706039428711,
      "learning_rate": 1.2221703105311827e-05,
      "loss": 0.1103,
      "step": 2108
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9162844036697247,
      "eval_loss": 0.31632736325263977,
      "eval_runtime": 1.4507,
      "eval_samples_per_second": 601.084,
      "eval_steps_per_second": 4.825,
      "step": 2108
    }
  ],
  "logging_steps": 500,
  "max_steps": 2635,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 748818980000436.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9947666589577905,
    "learning_rate": 6.110851552655913e-05,
    "num_train_epochs": 5,
    "temperature": 22
  }
}