trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 1250,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.4,
      "grad_norm": 13.731385231018066,
      "learning_rate": 1.2048000000000002e-05,
      "loss": 0.6602,
      "step": 500
    },
    {
      "epoch": 0.8,
      "grad_norm": 28.441198348999023,
      "learning_rate": 4.048e-06,
      "loss": 0.419,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9027631578947368,
      "eval_loss": 0.3514924645423889,
      "eval_runtime": 19.6631,
      "eval_samples_per_second": 386.511,
      "eval_steps_per_second": 48.314,
      "step": 1250
    }
  ],
  "logging_steps": 500,
  "max_steps": 1250,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 657801246720000.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}