trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 5159,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9998061639852684,
      "grad_norm": 1.3203125,
      "learning_rate": 0.00016,
      "loss": 0.7513,
      "step": 2579
    },
    {
      "epoch": 0.9998061639852684,
      "eval_loss": 0.7057055234909058,
      "eval_runtime": 107.7734,
      "eval_samples_per_second": 10.643,
      "eval_steps_per_second": 5.326,
      "step": 2579
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.9765625,
      "learning_rate": 0.0001199844901124467,
      "loss": 0.5953,
      "step": 5159
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.6852993369102478,
      "eval_runtime": 107.6973,
      "eval_samples_per_second": 10.65,
      "eval_steps_per_second": 5.33,
      "step": 5159
    }
  ],
  "logging_steps": 500,
  "max_steps": 12895,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.4602958088568013e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}