trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.32,
  "eval_steps": 10,
  "global_step": 20,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.16,
      "learning_rate": 1.6129032258064517e-05,
      "loss": 1.5807,
      "step": 10
    },
    {
      "epoch": 0.16,
      "eval_accuracy": 0.889,
      "eval_loss": 1.6238535642623901,
      "eval_runtime": 15.4204,
      "eval_samples_per_second": 64.849,
      "eval_steps_per_second": 4.085,
      "step": 10
    },
    {
      "epoch": 0.32,
      "learning_rate": 3.2258064516129034e-05,
      "loss": 1.5408,
      "step": 20
    },
    {
      "epoch": 0.32,
      "eval_accuracy": 0.896,
      "eval_loss": 1.56419038772583,
      "eval_runtime": 15.4357,
      "eval_samples_per_second": 64.785,
      "eval_steps_per_second": 4.081,
      "step": 20
    }
  ],
  "logging_steps": 10,
  "max_steps": 310,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 9.9277758332928e+16,
  "trial_name": null,
  "trial_params": null
}