Hurwitz Quaternion Multiplicative Quantization for KV Cache Compression 文章

ArXiv CS.AI2026-05-28NEWSen作者: Kabir Swain, Sijie Han, Daniel Karl I. Weidele, Mauro Martino, David Cox, Antonio Torralba

Hurwitz Quaternion Multiplicative Quantization for KV Cache Compression · 相关技术