trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 56.94915254237288,
  "eval_steps": 1000,
  "global_step": 420,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 13.56,
      "learning_rate": 3.821428571428572e-05,
      "loss": 0.554,
      "step": 100
    },
    {
      "epoch": 27.12,
      "learning_rate": 2.6309523809523813e-05,
      "loss": 0.0396,
      "step": 200
    },
    {
      "epoch": 40.68,
      "learning_rate": 1.4404761904761905e-05,
      "loss": 0.0312,
      "step": 300
    },
    {
      "epoch": 54.24,
      "learning_rate": 2.5e-06,
      "loss": 0.0308,
      "step": 400
    },
    {
      "epoch": 56.95,
      "step": 420,
      "total_flos": 8.064772262536032e+17,
      "train_loss": 0.15708597316628412,
      "train_runtime": 790.3622,
      "train_samples_per_second": 35.376,
      "train_steps_per_second": 0.531
    },
    {
      "epoch": 56.95,
      "eval_accuracy": 0.8709677457809448,
      "eval_loss": 0.4480999708175659,
      "eval_runtime": 1.4774,
      "eval_samples_per_second": 62.95,
      "eval_steps_per_second": 8.123,
      "step": 420
    }
  ],
  "logging_steps": 100,
  "max_steps": 420,
  "num_train_epochs": 60,
  "save_steps": 2000,
  "total_flos": 8.064772262536032e+17,
  "trial_name": null,
  "trial_params": null
}