checkpoint-2/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.971879096000318e-05,
  "global_step": 2,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 0.0001984126984126984,
      "loss": 10.8451,
      "theoretical_loss": 20.812814784551147,
      "tokens_seen": 65536
    },
    {
      "epoch": 0.0,
      "learning_rate": 0.0003968253968253968,
      "loss": 10.8162,
      "theoretical_loss": 17.56619808733264,
      "tokens_seen": 131072
    }
  ],
  "max_steps": 50354,
  "num_train_epochs": 9223372036854775807,
  "total_flos": 66890760192000.0,
  "trial_name": null,
  "trial_params": null
}