Grainger, et al.. Learning Patch-to-cluster Attention in Vision Transformer. 22 Mar. 2022.