llama.cpp introduces optimization for prompt lookup drafting

A new optimization for prompt lookup drafting has been announced for llama.cpp, aiming to speed up this speculative decoding technique. The update focuses on improving the efficiency of the prompt lookup process within the library.