Live data from Hacker News

Sinkhorn: Make LLMs even smaller through quantisation while maintaining accuracy

github.com

1–2 of 2 posts