REViT-v2: Hierarchical Windowed Roto-reflection Equivariant ViT for Equivariant Feature Extraction

Abstract:We propose a scalable roto-reflection-group-equivariant vision transformer based on windowed group-convolutional self-attention and a hierarchical feature architecture. We demonstrate that our approach can be scaled to group-equivariant vision transformers (ViTs) with millions of parameters and large datasets with practically sized images, i.e., ImageNet. The code and pretrained weights for the proposed Hierarchical Windowed Roto-reflection Equivariant ViTs (REViT-v2) are available at this https URL.

Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)

arXiv:2610.07585 [cs.CV]

(or arXiv:2610.07585v1 [cs.CV] for this version)

https://doi.org/10.48550/arXiv.2610.07585 Focus to learn more arXiv-issued DOI via DataCite (pending registration)

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.