All Tokens Matter: Token Labeling for Training Better Vision Transformers

Jiang, Zihang, Hou, Qibin, Yuan, Li · arXiv · 2021