CWCT: An Effective Vision Transformer using improved Cross-Window Self-Attention and CNN | IEEE Conference Publication | IEEE Xplore