Skip to content
Discussion options

You must be logged in to vote

This is a deliberate optimization. The script is configured to explicitly target Safetensors (.safetensors) weights and skip PyTorch .bin / .pth files.

Safetensors is a safer, faster-loading format that is fully supported by modern inference frameworks like Hugging Face Transformers, vLLM, and TGI. Excluding .bin prevents downloading duplicate weights and saves ~16GB of download bandwidth and disk space.

Replies: 1 comment

Comment options

You must be logged in to vote
0 replies
Answer selected by AhmadHassan-BTed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Category
Q&A
Labels
None yet
1 participant