{
  "best_metric": 0.9208715596330275,
  "best_model_checkpoint": "electra-distilled-sst\\run-32\\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": 47.12026596069336,
      "learning_rate": 4.2569562606597884e-05,
      "loss": 2.7545,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8990825688073395,
      "eval_loss": 1.4578064680099487,
      "eval_runtime": 1.4519,
      "eval_samples_per_second": 600.58,
      "eval_steps_per_second": 4.821,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 78.47111511230469,
      "learning_rate": 3.1942296148200405e-05,
      "loss": 1.283,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9071100917431193,
      "eval_loss": 1.441101312637329,
      "eval_runtime": 1.4585,
      "eval_samples_per_second": 597.892,
      "eval_steps_per_second": 4.8,
      "step": 1054
    },
    {
      "epoch": 3.0,
      "grad_norm": 58.518775939941406,
      "learning_rate": 2.1315029689802918e-05,
      "loss": 0.9314,
      "step": 1581
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9208715596330275,
      "eval_loss": 1.2692312002182007,
      "eval_runtime": 1.4446,
      "eval_samples_per_second": 603.643,
      "eval_steps_per_second": 4.846,
      "step": 1581
    },
    {
      "epoch": 4.0,
      "grad_norm": 19.554431915283203,
      "learning_rate": 1.0687763231405437e-05,
      "loss": 0.7341,
      "step": 2108
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9151376146788991,
      "eval_loss": 1.3720850944519043,
      "eval_runtime": 1.4488,
      "eval_samples_per_second": 601.895,
      "eval_steps_per_second": 4.832,
      "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": 0.2744072380603728,
    "learning_rate": 5.313633229198741e-05,
    "num_train_epochs": 5,
    "temperature": 16
  }
}