Interested in adding a new quantization method to Transformers?