I supported the development of a shift equivariant Vision Transformer. By making the vision transformer shift equivariant, it would potentially better classify images that are shifted from their original counterparts. I trained several different iterations of vision transformers and tested them with different types of adversarial attacks.