checkpoint-2180/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 2180,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.15,
      "learning_rate": 4.9800000000000004e-05,
      "loss": 0.0354,
      "step": 500
    },
    {
      "epoch": 2.29,
      "learning_rate": 3.992063035456259e-05,
      "loss": 0.0576,
      "step": 1000
    },
    {
      "epoch": 3.44,
      "learning_rate": 1.7720517816018433e-05,
      "loss": 0.0357,
      "step": 1500
    },
    {
      "epoch": 4.59,
      "learning_rate": 1.4339627226955392e-06,
      "loss": 0.017,
      "step": 2000
    }
  ],
  "logging_steps": 500,
  "max_steps": 2180,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 1.84692670576128e+16,
  "trial_name": null,
  "trial_params": null
}