Transformers supports loading models quantized with the llm-awq and autoawq libraries.