English

Less is More: Pay Less Attention in Vision Transformers

Computer Vision and Pattern Recognition 2021-12-24 v4

Abstract

Transformers have become one of the dominant architectures in deep learning, particularly as a powerful alternative to convolutional neural networks (CNNs) in computer vision. However, Transformer training and inference in previous works can be prohibitively expensive due to the quadratic complexity of self-attention over a long sequence of representations, especially for high-resolution dense prediction tasks. To this end, we present a novel Less attention vIsion Transformer (LIT), building upon the fact that the early self-attention layers in Transformers still focus on local patterns and bring minor benefits in recent hierarchical vision Transformers. Specifically, we propose a hierarchical Transformer where we use pure multi-layer perceptrons (MLPs) to encode rich local patterns in the early stages while applying self-attention modules to capture longer dependencies in deeper layers. Moreover, we further propose a learned deformable token merging module to adaptively fuse informative patches in a non-uniform manner. The proposed LIT achieves promising performance on image recognition tasks, including image classification, object detection and instance segmentation, serving as a strong backbone for many vision tasks. Code is available at: https://github.com/zhuang-group/LIT

Keywords

Cite

@article{arxiv.2105.14217,
  title  = {Less is More: Pay Less Attention in Vision Transformers},
  author = {Zizheng Pan and Bohan Zhuang and Haoyu He and Jing Liu and Jianfei Cai},
  journal= {arXiv preprint arXiv:2105.14217},
  year   = {2021}
}

Comments

Accepted to AAAI 2022

R2 v1 2026-06-24T02:35:43.862Z