trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.3579168123033904,
  "eval_steps": 500,
  "global_step": 16,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 1.9974521146102535e-05,
      "loss": 2.203,
      "step": 1
    },
    {
      "epoch": 0.11,
      "learning_rate": 1.936949724999762e-05,
      "loss": 2.2028,
      "step": 5
    },
    {
      "epoch": 0.22,
      "learning_rate": 1.7557495743542586e-05,
      "loss": 2.182,
      "step": 10
    },
    {
      "epoch": 0.34,
      "learning_rate": 1.479248986720057e-05,
      "loss": 2.1612,
      "step": 15
    },
    {
      "epoch": 0.36,
      "eval_loss": 2.144601583480835,
      "eval_runtime": 1290.2046,
      "eval_samples_per_second": 15.766,
      "eval_steps_per_second": 0.246,
      "step": 16
    },
    {
      "epoch": 0.36,
      "step": 16,
      "total_flos": 4.891029434007552e+16,
      "train_loss": 2.1917571127414703,
      "train_runtime": 41679.8817,
      "train_samples_per_second": 4.393,
      "train_steps_per_second": 0.001
    }
  ],
  "logging_steps": 5,
  "max_steps": 44,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 4.891029434007552e+16,
  "trial_name": null,
  "trial_params": null
}