[Official] prima.cpp: Fast 30-70B LLM inference on heterogeneous and everyday home devices
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).