论文笔记-Masked Autoencoders Are Scalable Vision Learners
论文地址:https://arxiv.org/pdf/2111.06377.pdf代码:https://github.com/lucidrains/vit-pytorch#masked-autoencoder一句话简介:随机mask掉高比例imagepatch,输入一个非对称的autoencoder,encoder是ViT,decoder是Transformer,效果好。1.介绍现在的深度学习模型