MorphMLP: A Self-Attention Free, MLP-Like Backbone for Image and Video - 42Papers