last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.00350938761186173,
  "eval_steps": 5,
  "global_step": 5,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0007018775223723461,
      "eval_loss": 10.378314971923828,
      "eval_runtime": 5.1789,
      "eval_samples_per_second": 463.422,
      "eval_steps_per_second": 57.928,
      "step": 1
    },
    {
      "epoch": 0.002105632567117038,
      "grad_norm": 0.07256367057561874,
      "learning_rate": 3e-05,
      "loss": 10.3792,
      "step": 3
    },
    {
      "epoch": 0.00350938761186173,
      "eval_loss": 10.377887725830078,
      "eval_runtime": 5.1938,
      "eval_samples_per_second": 462.089,
      "eval_steps_per_second": 57.761,
      "step": 5
    }
  ],
  "logging_steps": 3,
  "max_steps": 50,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 5,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 557753303040.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}