Training data-efficient image transformers & distillation through attention


Recently, neural networks purely based on attention were shown to address image understanding tasks such as image classification. However, these visual transformers are pre-trained with hundreds of millions of images using an expensive infrastructure, thereby limiting their adoption. In this work, we produce a competitive convolution-free transformer by training on Imagenet only. We train them on a single computer in less than 3 days. Our reference vision transformer (86M parameters) achieves top-1 accuracy of 83.1% (single-crop evaluation) on ImageNet with no external data. More importantly, we introduce a teacher-student strategy specific to transformers. It relies on a distillation token ensuring that the student learns from the teacher through attention. We show the interest of this token-based distillation, especially when using a convnet as a teacher. This leads us to report results competitive with convnets for both Imagenet (where we obtain up to 85.2% accuracy) and when transferring to other tasks. We share our code and models.

Results and models


The teacher of the distilled version DeiT is RegNetY-16GF.

Model Pretrain Params(M) Flops(G) Top-1 (%) Top-5 (%) Config Download
DeiT-tiny From scratch 5.72 1.08 74.50 92.24 config model | log
DeiT-tiny distilled* From scratch 5.72 1.08 74.51 91.90 config model
DeiT-small From scratch 22.05 4.24 80.69 95.06 config model | log
DeiT-small distilled* From scratch 22.05 4.24 81.17 95.40 config model
DeiT-base From scratch 86.57 16.86 81.76 95.81 config model | log
DeiT-base* From scratch 86.57 16.86 81.79 95.59 config model
DeiT-base distilled* From scratch 86.57 16.86 83.33 96.49 config model
DeiT-base 384px* ImageNet-1k 86.86 49.37 83.04 96.31 config model
DeiT-base distilled 384px* ImageNet-1k 86.86 49.37 85.55 97.35 config model

Models with * are converted from the official repo. The config files of these models are only for validation. We don’t ensure these config files’ training accuracy and welcome you to contribute your reproduction results.


MMClassification doesn’t support training the distilled version DeiT. And we provide distilled version checkpoints for inference only.


  title =     {Training data-efficient image transformers & distillation through attention},
  author =    {Touvron, Hugo and Cord, Matthieu and Douze, Matthijs and Massa, Francisco and Sablayrolles, Alexandre and Jegou, Herve},
  booktitle = {International Conference on Machine Learning},
  pages =     {10347--10357},
  year =      {2021},
  volume =    {139},
  month =     {July}
Read the Docs v: latest
On Read the Docs
Project Home

Free document hosting provided by Read the Docs.