ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders Friday February 17th, 2023 佐藤なつ dls-2023,