checkpoint-16/trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 16.0,
  "eval_steps": 500,
  "global_step": 16,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 5.48399443043752,
      "learning_rate": 1e-05,
      "loss": 0.5302,
      "mean_token_accuracy": 0.8446678519248962,
      "num_tokens": 39860.0,
      "step": 1
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.5189862450591725,
      "learning_rate": 1e-05,
      "loss": 0.1203,
      "mean_token_accuracy": 0.9620179865095351,
      "num_tokens": 398600.0,
      "step": 10
    }
  ],
  "logging_steps": 10,
  "max_steps": 32,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 32,
  "save_steps": 8,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1443804610560.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}