Convolutional Design

In the field of video recognition, achieving high performance usually entails significant computational costs. The proposed Video-FocalNets effectively merges the efficiency of convolutional designs with the global context modeling of…