trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "eval_steps": 100,
  "global_step": 56,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.36,
      "learning_rate": 0.00016428571428571428,
      "loss": 3.6178,
      "step": 10
    },
    {
      "epoch": 0.71,
      "learning_rate": 0.00012857142857142858,
      "loss": 3.2104,
      "step": 20
    },
    {
      "epoch": 1.07,
      "learning_rate": 9.285714285714286e-05,
      "loss": 2.8772,
      "step": 30
    },
    {
      "epoch": 1.43,
      "learning_rate": 5.714285714285714e-05,
      "loss": 2.6572,
      "step": 40
    },
    {
      "epoch": 1.79,
      "learning_rate": 2.1428571428571428e-05,
      "loss": 2.554,
      "step": 50
    },
    {
      "epoch": 2.0,
      "step": 56,
      "total_flos": 6.762570523759411e+16,
      "train_loss": 2.928572859082903,
      "train_runtime": 1229.5876,
      "train_samples_per_second": 0.709,
      "train_steps_per_second": 0.046
    }
  ],
  "logging_steps": 10,
  "max_steps": 56,
  "num_train_epochs": 2,
  "save_steps": 100,
  "total_flos": 6.762570523759411e+16,
  "trial_name": null,
  "trial_params": null
}