view article Article Performant local mixture-of-experts CPU inference with GPU acceleration in llama.cpp 3 days ago โข 9