We can add an integration to llama.cpp, please open an issue on the repo if you’re interested!