checkpoint-2000/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.27427317608337903,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.07,
      "learning_rate": 2.897147558968733e-05,
      "loss": 3.8538,
      "step": 500
    },
    {
      "epoch": 0.14,
      "learning_rate": 2.794295117937466e-05,
      "loss": 2.9886,
      "step": 1000
    },
    {
      "epoch": 0.21,
      "learning_rate": 2.6914426769061988e-05,
      "loss": 2.685,
      "step": 1500
    },
    {
      "epoch": 0.27,
      "learning_rate": 2.5885902358749317e-05,
      "loss": 2.4911,
      "step": 2000
    }
  ],
  "max_steps": 14584,
  "num_train_epochs": 2,
  "total_flos": 252701945856000.0,
  "trial_name": null,
  "trial_params": null
}