VSA: Learning Varied-Size Window Attention in Vision Transformers.

Qiming Zhang, Yufei Xu, Jing Zhang, Dacheng Tao

Browse the full ECCV paper archive.