Lightning-AI / litgpt

20+ high-performance LLMs with recipes to pretrain, finetune and deploy at scale.
https://lightning.ai
Apache License 2.0
9.1k stars 910 forks source link

Deal with warnings during model download and conversion #1557

Closed Andrei-Aksionov closed 1 week ago

Andrei-Aksionov commented 3 weeks ago

When we download (and automatically convert) a model, we have these three warnings:

  1. .../huggingface_hub/file_download.py:1132: FutureWarning: resume_download is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use force_download=True.
  2. .../huggingface_hub/file_download.py:1194: UserWarning: local_dir_use_symlinks parameter is deprecated and will be ignored. The process to download files to a local folder has been updated and do not rely on symlinks anymore. You only need to pass a destination folder aslocal_dir. For more details, check out https://huggingface.co/docs/huggingface_hub/main/en/guides/download#download-files-to-local-folder.
  3. [W init.cpp:1436] Warning: write_record(): Passing Storage by data pointer is deprecated and will be an error in the future, please pass the Storage object instead. (function operator())

First two come during weights downloading, while the third - during conversion.

rasbt commented 3 weeks ago

Good call. I can take care of these next week

Andrei-Aksionov commented 3 weeks ago

It's not super-urgent, so maybe let's give the community a chance. I think it's a cool thing when there are a number of good first issue issues, so it's easier to start contributing. Other issues are quite technical, and it might be overwhelming for someone. I started contributing by just providing annotations 😉.

rasbt commented 3 weeks ago

Good point, I agree

bhargavyagnik commented 1 week ago

This Issue has been solved, I think we can close it.

rasbt commented 1 week ago

Good point. Thanks again for the contribution!

bhargavyagnik commented 1 week ago

Anytime! I would love to contribute more and will keep an eye on the issues page for more contributions! If you have any, let me know. Thanks