نسخة أولية وصول مفتوح
REViT-v2: Hierarchical Windowed Roto-reflection Equivariant ViT for Equivariant Feature Extraction
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 prac …