Skip to content
Better HN
A complete Llama2 inference engine that fits in 1356 bytes of x86 assembly | Better HN