A complete Llama2 inference engine that fits in 1356 bytes of x86 assembly(github.com)27 points by monax 13 days ago | 0 commentsNo comments yet