view article Article Performant local mixture-of-experts CPU inference with GPU acceleration in llama.cpp 1 day ago • 6