Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
jsulz 
posted an update Oct 10
Post
1638
The Hugging Face Hub hosts over 1.5M Model, Dataset, and Space repositories. To scale to 10M+, the XetHub team (https://huggingface.co/xet-team) is replacing Git LFS with a new technology that improves storage and transfer capabilities with some future developer experience benefits to boot.

Thanks to @yuchenglow and @port8080 (for their analysis covering LFS usage from March 2022–Sept 2024), we now have insights into what we’re storing. Check out the Gradio app to explore:
- Storage growth over time
- File types over all repositories
- Some simple optimizations we're investigating

xet-team/lfs-analysis
In this post