VPTQ: Extreme Low-bit Vector Post-Training Quantization for Large Language Models Paper • 2409.17066 • Published Sep 25 • 27
