trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500.0,
  "global_step": 768,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0013020833333333333,
      "grad_norm": 576.0,
      "learning_rate": 1.5384615384615385e-07,
      "loss": 12.40298843383789,
      "step": 1,
      "token_acc": 0.265625
    },
    {
      "epoch": 0.6510416666666666,
      "grad_norm": 4.46875,
      "learning_rate": 1.7880630644520184e-06,
      "loss": 0.548660308899048,
      "step": 500,
      "token_acc": 0.9504947394789579
    }
  ],
  "logging_steps": 500,
  "max_steps": 768,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 2000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 4.891798036021248e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}