trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.1616842105263158,
  "eval_steps": 500,
  "global_step": 3,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.05,
      "learning_rate": 1.9848077530122083e-05,
      "loss": 2.0569,
      "step": 1
    },
    {
      "epoch": 0.16,
      "eval_loss": 2.04089093208313,
      "eval_runtime": 14.9506,
      "eval_samples_per_second": 33.444,
      "eval_steps_per_second": 4.214,
      "step": 3
    },
    {
      "epoch": 0.16,
      "step": 3,
      "total_flos": 1.3928862408769536e+17,
      "train_loss": 2.1085566679636636,
      "train_runtime": 937.9718,
      "train_samples_per_second": 10.128,
      "train_steps_per_second": 0.019
    }
  ],
  "logging_steps": 5,
  "max_steps": 18,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 1.3928862408769536e+17,
  "trial_name": null,
  "trial_params": null
}