run-13/checkpoint-500/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.5723270440251573,
  "eval_steps": 500,
  "global_step": 500,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6780645161290323,
      "eval_loss": 0.40519121289253235,
      "eval_runtime": 1.9988,
      "eval_samples_per_second": 1550.899,
      "eval_steps_per_second": 32.519,
      "step": 318
    },
    {
      "epoch": 1.57,
      "grad_norm": 0.8578532934188843,
      "learning_rate": 1.650593990216632e-05,
      "loss": 0.6413,
      "step": 500
    }
  ],
  "logging_steps": 500,
  "max_steps": 2862,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 500,
  "total_flos": 130072209152340.0,
  "train_batch_size": 48,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8018818773384359,
    "num_train_epochs": 9,
    "temperature": 2
  }
}