trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.0,
  "global_step": 3724,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.27,
      "learning_rate": 2.597207303974221e-05,
      "loss": 2.5132,
      "step": 500
    },
    {
      "epoch": 0.54,
      "learning_rate": 2.1944146079484428e-05,
      "loss": 1.3194,
      "step": 1000
    },
    {
      "epoch": 0.81,
      "learning_rate": 1.7916219119226638e-05,
      "loss": 1.0847,
      "step": 1500
    },
    {
      "epoch": 1.07,
      "learning_rate": 1.388829215896885e-05,
      "loss": 0.9563,
      "step": 2000
    },
    {
      "epoch": 1.34,
      "learning_rate": 9.860365198711064e-06,
      "loss": 0.8287,
      "step": 2500
    },
    {
      "epoch": 1.61,
      "learning_rate": 5.832438238453276e-06,
      "loss": 0.8071,
      "step": 3000
    },
    {
      "epoch": 1.88,
      "learning_rate": 1.8045112781954887e-06,
      "loss": 0.7936,
      "step": 3500
    },
    {
      "epoch": 2.0,
      "step": 3724,
      "total_flos": 3.502304143748813e+16,
      "train_loss": 0.500451392185035,
      "train_runtime": 1495.419,
      "train_samples_per_second": 119.508,
      "train_steps_per_second": 2.49
    }
  ],
  "max_steps": 3724,
  "num_train_epochs": 2,
  "total_flos": 3.502304143748813e+16,
  "trial_name": null,
  "trial_params": null
}