GitLab

* [gguf] Refactor __torch_function__ to avoid unnecessary computation

This helps with torch.compile compilation lantency. Avoiding unnecessary
computation should also lead to a slightly improved eager latency.

* Apply style fixes

---------
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>