Vanilla Group Equivariant Vision Transformer: Simple and Effective

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Fu, Jiahong, Xie, Qi, Meng, Deyu, Xu, Zongben
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908820525023232
author Fu, Jiahong
Xie, Qi
Meng, Deyu
Xu, Zongben
author_facet Fu, Jiahong
Xie, Qi
Meng, Deyu
Xu, Zongben
contents Incorporating symmetry priors as inductive biases to design equivariant Vision Transformers (ViTs) has emerged as a promising avenue for enhancing their performance. However, existing equivariant ViTs often struggle to balance performance with equivariance, primarily due to the challenge of achieving holistic equivariant modifications across the diverse modules in ViTs-particularly in harmonizing the Self-Attention mechanism with Patch Embedding. To address this, we propose a straightforward framework that systematically renders key ViT components, including patch embedding, self-attention, positional encodings, and Down/Up-Sampling, equivariant, thereby constructing ViTs with guaranteed equivariance. The resulting architecture serves as a plug-and-play replacement that is both theoretically grounded and practically versatile, scaling seamlessly even to Swin Transformers. Extensive experiments demonstrate that our equivariant ViTs consistently improve performance and data efficiency across a wide spectrum of vision tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2602_08047
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Vanilla Group Equivariant Vision Transformer: Simple and Effective
Fu, Jiahong
Xie, Qi
Meng, Deyu
Xu, Zongben
Computer Vision and Pattern Recognition
Incorporating symmetry priors as inductive biases to design equivariant Vision Transformers (ViTs) has emerged as a promising avenue for enhancing their performance. However, existing equivariant ViTs often struggle to balance performance with equivariance, primarily due to the challenge of achieving holistic equivariant modifications across the diverse modules in ViTs-particularly in harmonizing the Self-Attention mechanism with Patch Embedding. To address this, we propose a straightforward framework that systematically renders key ViT components, including patch embedding, self-attention, positional encodings, and Down/Up-Sampling, equivariant, thereby constructing ViTs with guaranteed equivariance. The resulting architecture serves as a plug-and-play replacement that is both theoretically grounded and practically versatile, scaling seamlessly even to Swin Transformers. Extensive experiments demonstrate that our equivariant ViTs consistently improve performance and data efficiency across a wide spectrum of vision tasks.
title Vanilla Group Equivariant Vision Transformer: Simple and Effective
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2602.08047