[arxiv]
Vision transformer has achieved impressive performance for many vision tasks. However, it may suffer from high redundancy in capturing local features for shallow layers. Local self-attention or early-stage convolutions are thus utilized, which sacrifice the capacity to capture long-range dependency. A challenge then arises: can we access efficient and effective global context modeling at the early stages of a neural network? To address this issue, we draw inspiration from the design of superpixels, which reduces the number of image primitives in subsequent processing, and introduce super tokens into vision transformer. Super tokens attempt to provide a semantically meaningful tessellation of visual content, thus reducing the token number in self-attention as well as preserving global modeling. Specifically, we propose a simple yet strong super token attention (STA) mechanism with three steps: the first samples super tokens from visual tokens via sparse association learning, the second performs self-attention on super tokens, and the last maps them back to the original token space. STA decomposes vanilla global attention into multiplications of a sparse association map and a low-dimensional attention, leading to high efficiency in capturing global dependencies. Based on STA, we develop a hierarchical vision transformer. Extensive experiments demonstrate its strong performance on various vision tasks. In particular, without any extra training data or label, it achieves 86.4 top-1 accuracy on ImageNet-1K with less than 100M parameters. It also achieves 53.9 box AP and 46.8 mask AP on the COCO detection task, and 51.9 mIOU on the ADE20K semantic segmentation task.
The pretrained weights will be released soon.
| Model | Input-size | Params | FLOPs | Acc |
|---|---|---|---|---|
| SViT-S | 224x224 | 25M | 4.4G | 83.6% |
| SViT-S | 384x384 | 25M | 14.1G | 85.0% |
| SViT-B | 224x224 | 52M | 9.9G | 84.8% |
| SViT-B | 384x384 | 52M | 32.5G | 86.0% |
| SViT-L | 224x224 | 95M | 15.6G | 85.3% |
| SViT-L | 384x384 | 95M | 49.7G | 86.4% |
@article{huang2022stvit,
title={Vision Transformer with Super Token Sampling},
author={Huang, Huaibo and Zhou, Xiaoqiang and Cao, Jie and He, Ran and Tan, Tieniu},
journal={arXiv:2211.11167},
year={2022}
}