Attention Sparsity is Input-Stable: Training-Free Sparse Attention for Video Generation via Offline Sparsity Profiling and Online QK Co-Clustering

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Luo, Jiayi, Chen, Jiayu, Wang, Jiankun, Wang, Cong, Zhu, Hanxin, Sun, Qingyun, Gao, Chen, Chen, Zhibo, Li, Jianxin
Format:	Preprint
Published:	2026
Subjects:	Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2603.18636