checkpoint-2000/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.9840637450199203,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 4.00398406374502e-05,
      "loss": 3.6087,
      "step": 500
    },
    {
      "epoch": 1.99,
      "learning_rate": 3.00796812749004e-05,
      "loss": 3.242,
      "step": 1000
    },
    {
      "epoch": 2.99,
      "learning_rate": 2.01195219123506e-05,
      "loss": 3.0879,
      "step": 1500
    },
    {
      "epoch": 3.98,
      "learning_rate": 1.0159362549800798e-05,
      "loss": 2.9902,
      "step": 2000
    }
  ],
  "max_steps": 2510,
  "num_train_epochs": 5,
  "total_flos": 3055939046866944.0,
  "trial_name": null,
  "trial_params": null
}