1 paper
Jiaxuan Li, Qing Xu, Xiangjian He +6
Masked Autoencoders (MAE) achieve self-supervised learning of image representations by randomly removing a portion of visual tokens and reconstructing the original image as a prete…