checkpoint-600/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.2631578947368425,
  "eval_steps": 500,
  "global_step": 600,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 4.385964912280702,
      "grad_norm": 9.875,
      "learning_rate": 2.5e-05,
      "loss": 2.8556,
      "step": 500
    }
  ],
  "logging_steps": 500,
  "max_steps": 4560,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 150,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.05082540900352e+16,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}