checkpoint-5/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.25,
  "eval_steps": 500,
  "global_step": 5,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.25,
      "learning_rate": 0.0001,
      "loss": 2.38,
      "step": 1
    },
    {
      "epoch": 1.25,
      "learning_rate": 0.0002,
      "loss": 2.38,
      "step": 2
    },
    {
      "epoch": 2.25,
      "learning_rate": 0.00018888888888888888,
      "loss": 2.2781,
      "step": 3
    },
    {
      "epoch": 3.25,
      "learning_rate": 0.00017777777777777779,
      "loss": 2.1161,
      "step": 4
    },
    {
      "epoch": 4.25,
      "learning_rate": 0.0001666666666666667,
      "loss": 1.9852,
      "step": 5
    }
  ],
  "logging_steps": 1,
  "max_steps": 20,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 52796817408000.0,
  "trial_name": null,
  "trial_params": null
}