# DeiT-LT Distillation Strikes Back for Vision Transformer Training on Long-Tailed Datasets

arXiv: https://arxiv.org/abs/2404.02900 · HTML: https://mlsanity.com/paper/2404.02900/

## Key facts

- Cited by 14 papers in the map (data through 14 Oct 2026).
- Builds on 55 of its 66 references that are in the map.
- First posted on arXiv 3 Apr 2024.

## Builds on (most cited first, 30 of 55)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [mixup: Beyond Empirical Risk Minimization](https://mlsanity.com/paper/1710.09412/) (2017)
- [Training data-efficient image transformers & distillation through attention](https://mlsanity.com/paper/2012.12877/) (2020)
- [CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features](https://mlsanity.com/paper/1905.04899/) (2019)
- [MLP-Mixer: An all-MLP Architecture for Vision](https://mlsanity.com/paper/2105.01601/) (2021)
- [Sharpness-Aware Minimization for Efficiently Improving Generalization](https://mlsanity.com/paper/2010.01412/) (2020)
- [Class-Balanced Loss Based on Effective Number of Samples](https://mlsanity.com/paper/1901.05555/) (2019)
- [LVIS: A Dataset for Large Vocabulary Instance Segmentation](https://mlsanity.com/paper/1908.03195/) (2019)
- [Learning Imbalanced Datasets with Label-Distribution-Aware Margin Loss](https://mlsanity.com/paper/1906.07413/) (2019)
- [Designing Network Design Spaces](https://mlsanity.com/paper/2003.13678/) (2020)
- [Segmenter: Transformer for Semantic Segmentation](https://mlsanity.com/paper/2105.05633/) (2021)
- [Decoupling Representation and Classifier for Long-Tailed Recognition](https://mlsanity.com/paper/1910.09217/) (2019)
- [Quantifying Attention Flow in Transformers](https://mlsanity.com/paper/2005.00928/) (2020)
- [Going deeper with Image Transformers](https://mlsanity.com/paper/2103.17239/) (2021)
- [Large-Scale Long-Tailed Recognition in an Open World](https://mlsanity.com/paper/1904.05160/) (2019)
- [Long-tail learning via logit adjustment](https://mlsanity.com/paper/2007.07314/) (2020)
- [DeiT III: Revenge of the ViT](https://mlsanity.com/paper/2204.07118/) (2022)
- [BBN: Bilateral-Branch Network with Cumulative Learning for Long-Tailed Visual Recognition](https://mlsanity.com/paper/1912.02413/) (2019)
- [Balanced Meta-Softmax for Long-Tailed Visual Recognition](https://mlsanity.com/paper/2007.10740/) (2020)
- [Long-tailed Recognition by Routing Diverse Distribution-Aware Experts](https://mlsanity.com/paper/2010.01809/) (2020)
- [Demystifying CLIP Data](https://mlsanity.com/paper/2309.16671/) (2023)
- [Parametric Contrastive Learning](https://mlsanity.com/paper/2107.12028/) (2021)
- [Improving Calibration for Long-Tailed Recognition](https://mlsanity.com/paper/2104.00466/) (2021)
- [Distribution Alignment: A Unified Framework for Long-tail Visual Recognition](https://mlsanity.com/paper/2103.16370/) (2021)
- [Disentangling Label Distribution for Long-tailed Visual Recognition](https://mlsanity.com/paper/2012.00321/) (2020)
- [Rethinking Class-Balanced Methods for Long-Tailed Visual Recognition from a Domain Adaptation Perspective](https://mlsanity.com/paper/2003.10780/) (2020)

## Built on by (most cited first, 14 of 14)

- [A Systematic Review on Long-Tailed Learning](https://mlsanity.com/paper/2408.00483/) (2024)
- [Enhancing Features in Long-tailed Data Using Large Vision Model](https://mlsanity.com/paper/2504.10852/) (2025)
- [Empowering Vision Transformers with Multi-Scale Causal Intervention for Long-Tailed Image Classification](https://mlsanity.com/paper/2505.08173/) (2025)
- [Parameter Reduction Improves Vision Transformers: A Comparative Study of Sharing and Width Reduction](https://mlsanity.com/paper/2512.01059/) (2025)
- [Dataset Awareness is not Enough: Implementing Sample-level Tail Encouragement in Long-tailed Self-supervised Learning](https://mlsanity.com/paper/2410.22883/) (2024)
- [Deep Minimax Classifiers for Imbalanced Datasets with a Small Number of Minority Samples](https://mlsanity.com/paper/2502.16948/) (2025)
- [Model compression using knowledge distillation with integrated gradients](https://mlsanity.com/paper/2506.14440/) (2025)
- [Semantics-Aware Generative Latent Data Augmentation for Learning in Low-Resource Domains](https://mlsanity.com/paper/2602.02841/) (2026)
- [Decompose, Mix, Adapt: A Unified Framework for Parameter-Efficient Neural Network Recombination and Compression](https://mlsanity.com/paper/2603.27383/) (2026)
- [Weak-to-Strong Knowledge Distillation Accelerates Visual Learning](https://mlsanity.com/paper/2604.15451/) (2026)
- [Autonomous Drift Learning in Data Streams: A Unified Perspective](https://mlsanity.com/paper/2605.01295/) (2026)
- [OSDTW: Optimal Shared Depth and Task Weighting for Long-Tailed Recognition](https://mlsanity.com/paper/2605.24969/) (2026)
- [Cross-Modal Knowledge Distillation for Acoustic Pedestrian Detection](https://mlsanity.com/paper/2609.31785/) (2026)
- [OFBD: Object-Focused Background Debiasing for Long-Tailed Learning](https://mlsanity.com/paper/2609.37331/) (2026)

## Abstract (arXiv)

Vision Transformer (ViT) has emerged as a prominent architecture for various computer vision tasks. In ViT, we divide the input image into patch tokens and process them through a stack of self attention blocks. However, unlike Convolutional Neural Networks (CNN), ViTs simple architecture has no informative inductive bias (e.g., locality,etc. ). Due to this, ViT requires a large amount of data for pre-training. Various data efficient approaches (DeiT) have been proposed to train ViT on balanced datasets effectively. However, limited literature discusses the use of ViT for datasets with long-tailed imbalances. In this work, we introduce DeiT-LT to tackle the problem of training ViTs from scratch on long-tailed datasets. In DeiT-LT, we introduce an efficient and effective way of distillation from CNN via distillation DIST token by using out-of-distribution images and re-weighting the distillation loss to enhance focus on tail classes. This leads to the learning of local CNN-like features in early ViT blocks, improving generalization for tail classes. Further, to mitigate overfitting, we propose distilling from a flat CNN teacher, which leads to learning low-rank generalizable features for DIST tokens across all ViT blocks. With the proposed DeiT-LT scheme, the distillation DIST token becomes an expert on the tail classes, and the classifier CLS token becomes an expert on the head classes. The experts help to effectively learn features corresponding to both the majority and minority classes using a distinct set of tokens within the same ViT architecture. We show the effectiveness of DeiT-LT for training ViT from scratch on datasets ranging from small-scale CIFAR-10 LT to large-scale iNaturalist-2018.