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

 (github.com)

9 points | by monax 7 hours ago ago

No comments yet.