When Vision Transformers Outperform ResNets without Pre-training or Strong Data Augmentations.

Xiangning Chen, Cho-Jui Hsieh, Boqing Gong

Browse the full ICLR paper archive.