mit-han-lab / TinyChatEngine

TinyChatEngine: On-Device LLM Inference Library
https://mit-han-lab.github.io/TinyChatEngine/
MIT License
715 stars 68 forks source link

Upload model checkpoints on the Hugging Face Hub. #100

Open Vaibhavs10 opened 6 months ago

Vaibhavs10 commented 6 months ago

Hi there,

Congrats on creating such a brilliant library and for also releasing AWQ kernels that work w/ metal as well. I'm VB I lead the advocacy efforts for Open Source ML at Hugging Face. I saw that the model checkpoints are currently hosted on a Dropbox. It'd be great if you could upload and share the models on the Hugging Face Hub!

The model checkpoints could reside under your org: https://huggingface.co/mit-han-lab

In general this ensures better discovery of both your library and the checkpoints.

In the process we can also integrate it with the huggingface_hub python library to programmatically download the checkpoints as well. This would ensure seamless caching of the models as well.

Happy to help you with this process.

Congrats, and I'm looking forward to these checkpoints on the Hub.

Cheers! VB

Vaibhavs10 commented 6 months ago

(happy to also take this conversation over on the slack channel too)