checkpoint-1000/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.5002501250625313,
  "eval_steps": 500,
  "global_step": 1000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5002501250625313,
      "grad_norm": 8.135814666748047,
      "learning_rate": 5.0075037518759384e-06,
      "loss": 0.2622,
      "step": 1000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 1999,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 0.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}