1
ARCQuant: Boosting NVFP4 Quantization with Augmented Residual Channels for LLMs
新型量化方法ARCQuant通过增强残差通道，显著提升NVFP4在大语言模型上的精度表现，为低比特量化开辟新路径。
arXiv:2601.07475v2 Announce Type: replace-cross Abstract: The emergence of fine-grained numerical formats like NVFP4 presents new opportunities for ef…