llama.cpp

28ba975a - Check the existence of f16_model_path_base in quantize.py (#574)