A complete Llama2 inference engine that fits in 1356 bytes of x86 assembly #1 Post by monax » Tue, May 05, 2026, 6:53 AM UTC A complete Llama2 inference engine that fits in 1356 bytes of x86 assemblygithub.com