llama.cpp/requirements-hf-to-gguf.txt