A complete Llama2 inference engine that fits in 1356 bytes of x86 assembly

 (github.com)

27 points | by monax 3 days ago ago

No comments yet.