checkpoint-2000/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.442340791738382,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.86,
      "learning_rate": 4.1394148020654047e-05,
      "loss": 2.054,
      "step": 500
    },
    {
      "epoch": 1.72,
      "learning_rate": 3.278829604130809e-05,
      "loss": 1.8443,
      "step": 1000
    },
    {
      "epoch": 2.58,
      "learning_rate": 2.418244406196214e-05,
      "loss": 1.7694,
      "step": 1500
    },
    {
      "epoch": 3.44,
      "learning_rate": 1.557659208261618e-05,
      "loss": 1.7264,
      "step": 2000
    }
  ],
  "max_steps": 2905,
  "num_train_epochs": 5,
  "total_flos": 3057659302772736.0,
  "trial_name": null,
  "trial_params": null
}