%0 Generic %1 dosovitskiy2020image %A Dosovitskiy, Alexey %A Beyer, Lucas %A Kolesnikov, Alexander %A Weissenborn, Dirk %A Zhai, Xiaohua %A Unterthiner, Thomas %A Dehghani, Mostafa %A Minderer, Matthias %A Heigold, Georg %A Gelly, Sylvain %A Uszkoreit, Jakob %A Houlsby, Neil %D 2020 %K cs.AI cs.CV cs.LG %T An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale %U http://arxiv.org/abs/2010.11929 %X While the Transformer architecture has become the de-facto standard for natural language processing tasks, its applications to computer vision remain limited. In vision, attention is either applied in conjunction with convolutional networks, or used to replace certain components of convolutional networks while keeping their overall structure in place. We show that this reliance on CNNs is not necessary and a pure transformer applied directly to sequences of image patches can perform very well on image classification tasks. When pre-trained on large amounts of data and transferred to multiple mid-sized or small image recognition benchmarks (ImageNet, CIFAR-100, VTAB, etc.), Vision Transformer (ViT) attains excellent results compared to state-of-the-art convolutional networks while requiring substantially fewer computational resources to train.

@misc{dosovitskiy2020image, abstract = {While the Transformer architecture has become the de-facto standard for natural language processing tasks, its applications to computer vision remain limited. In vision, attention is either applied in conjunction with convolutional networks, or used to replace certain components of convolutional networks while keeping their overall structure in place. We show that this reliance on CNNs is not necessary and a pure transformer applied directly to sequences of image patches can perform very well on image classification tasks. When pre-trained on large amounts of data and transferred to multiple mid-sized or small image recognition benchmarks (ImageNet, CIFAR-100, VTAB, etc.), Vision Transformer (ViT) attains excellent results compared to state-of-the-art convolutional networks while requiring substantially fewer computational resources to train.}, added-at = {2021-08-23T03:30:17.000+0200}, author = {Dosovitskiy, Alexey and Beyer, Lucas and Kolesnikov, Alexander and Weissenborn, Dirk and Zhai, Xiaohua and Unterthiner, Thomas and Dehghani, Mostafa and Minderer, Matthias and Heigold, Georg and Gelly, Sylvain and Uszkoreit, Jakob and Houlsby, Neil}, biburl = {https://www.bibsonomy.org/bibtex/2ed4c1d5b3a4068d9d621d3bbee93b968/aerover}, description = {An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale}, interhash = {f05240270962a2049a0aaec1db748975}, intrahash = {ed4c1d5b3a4068d9d621d3bbee93b968}, keywords = {cs.AI cs.CV cs.LG}, note = {cite arxiv:2010.11929Comment: Fine-tuning code and pre-trained models are available at https://github.com/google-research/vision_transformer. ICLR camera-ready version with 2 small modifications: 1) Added a discussion of CLS vs GAP classifier in the appendix, 2) Fixed an error in exaFLOPs computation in Figure 5 and Table 6 (relative performance of models is basically not affected)}, timestamp = {2021-08-23T03:30:17.000+0200}, title = {An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale}, url = {http://arxiv.org/abs/2010.11929}, year = 2020 }