view article Article Performant local mixture-of-experts CPU inference with GPU acceleration in llama.cpp Doctor-Shotgun • Jan 30 • 25