trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.4707820867415995,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 1.3674,
      "step": 1000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.0002,
      "loss": 39667.712,
      "step": 2000
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.0002,
      "loss": 0.0,
      "step": 3000
    },
    {
      "epoch": 0.47,
      "step": 3000,
      "total_flos": 4.834313788425339e+17,
      "train_loss": 13223.026456787109,
      "train_runtime": 91003.2011,
      "train_samples_per_second": 0.527,
      "train_steps_per_second": 0.033
    }
  ],
  "max_steps": 3000,
  "num_train_epochs": 1,
  "total_flos": 4.834313788425339e+17,
  "trial_name": null,
  "trial_params": null
}