MetaFormer is actually what you need for vision

• Main Authors: YU, Weihao, LUO, Mi, ZHOU, Pan, SI, Chenyang, ZHOU, Yichen, WANG, Xinchao, FENG, Jiashi, YAN, Shuicheng
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2022
• Subjects: Computer vision; Computational modeling; Focusing; Computer architecture; Transformers; Pattern recognition; Task analysis; Graphics and Human Computer Interfaces
• Online Access: https://ink.library.smu.edu.sg/sis_research/8983; https://ink.library.smu.edu.sg/context/sis_research/article/9986/viewcontent/2022_CVPR_MetaFormer.pdf
• Institution: Singapore Management University

Transformers have shown great potential in computer vision tasks. A common belief is their attention-based token mixer module contributes most to their competence. However, recent works show the attention-based module in transformers can be replaced by spatial MLPs and the resulted models still perf...