Fetching the paper…

ResQ: Mixed-Precision Quantization of Large Language Models with Low-Rank Residuals · Around