SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models

Xiao, Guangxuan, Lin, Ji, Seznec, Mickael · arXiv · 2022