🔘 Paper page: arxiv.org/abs/2108.08810v1 Abstract «Convolutional neural networks (CNNs) have so far been the de-facto model for visual ...