ShiftAddViT: Mixture of Multiplication Primitives Towards Efficient Vision Transformer

Open in new window