Exploring whether attention is necessary for vision transformers - View it on GitHub