trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 5432,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5522827687776142,
      "grad_norm": 0.2080295830965042,
      "learning_rate": 3.986212489862125e-05,
      "loss": 1.6017,
      "step": 1500
    },
    {
      "epoch": 1.1045655375552283,
      "grad_norm": 0.15871605277061462,
      "learning_rate": 2.4655312246553124e-05,
      "loss": 0.2784,
      "step": 3000
    },
    {
      "epoch": 1.6568483063328423,
      "grad_norm": 0.21663200855255127,
      "learning_rate": 9.448499594484997e-06,
      "loss": 0.2594,
      "step": 4500
    }
  ],
  "logging_steps": 1500,
  "max_steps": 5432,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 1500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.038878105187123e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}