trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.557544757033248,
  "eval_steps": 500,
  "global_step": 1000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.278772378516624,
      "grad_norm": 7.643950939178467,
      "learning_rate": 1.1474850809889173e-05,
      "loss": 0.3095,
      "step": 500
    },
    {
      "epoch": 2.557544757033248,
      "grad_norm": 4.580963611602783,
      "learning_rate": 2.949701619778346e-06,
      "loss": 0.1631,
      "step": 1000
    }
  ],
  "logging_steps": 500,
  "max_steps": 1173,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1.6722992037888e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}