checkpoint-1275/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.127858002406739,
  "eval_steps": 500,
  "global_step": 1275,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.07,
      "learning_rate": 1.9839679358717435e-05,
      "loss": 6.8032,
      "step": 664
    }
  ],
  "logging_steps": 664,
  "max_steps": 9972,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1.113158711771136e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}