How to train your ViT? Data, Augmentation, and Regularization in Vision Transformers

Abstract

Vision Transformers (ViT) have been shown to attain highly competitive performance for a wide range of vision applications, such as image classification, object detection and semantic image segmentation. In comparison to convolutional neural networks, the Vision Transformer's weaker inductive bias is generally found to cause an increased reliance on model regularization or data augmentation ("AugReg" for short) when training on smaller training datasets. We conduct a systematic empirical study in order to better understand the interplay between the amount of training data, AugReg, model size and compute budget. As one result of this study we find that the combination of increased compute and AugReg can yield models with the same performance as models trained on an order of magnitude more training data: we train ViT models of various sizes on the public ImageNet-21k dataset which either match or outperform their counterparts trained on the larger, but not publicly available JFT-300M dataset.

Full paper on arXiv

Details

arXiv identifier

2106.10270

Published

2021-06-18

Authors

Andreas Steiner, Alexander Kolesnikov, Xiaohua Zhai, Ross Wightman, Jakob Uszkoreit, Lucas Beyer

Models That Cite This Paper

• Described byvit_small_patch16_224.augreg_in21k_ft_in1k
• Described byvit_base_patch16_224.augreg2_in21k_ft_in1k
• Described byvit_base_patch8_224.augreg2_in21k_ft_in1k
• Described byvit_tiny_patch16_224.augreg_in21k_ft_in1k
• Described byvit_base_patch16_384.augreg_in21k_ft_in1k
• Described byvit_base_patch16_224.augreg_in21k
• Described byvit_base_patch16_224.augreg_in21k_ft_in1k