Hacker Next
A complete Llama2 inference engine that fits in 1356 bytes of x86 assemblygithub.com
27 points by monax 25 days ago | 0 comments