trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 100.0,
  "eval_steps": 500,
  "global_step": 800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 62.5,
      "grad_norm": 1.1140620708465576,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 0.7145,
      "step": 500
    },
    {
      "epoch": 100.0,
      "step": 800,
      "total_flos": 4353076050246000.0,
      "train_loss": 0.49458239078521726,
      "train_runtime": 511.4461,
      "train_samples_per_second": 47.317,
      "train_steps_per_second": 1.564
    }
  ],
  "logging_steps": 500,
  "max_steps": 800,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 100,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4353076050246000.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}