llama.cpp

Port of Facebook's LLaMA model in C/C++ for efficient CPU inference.