# REDI: Corpus Aware Patch Ranking for DINOv3 Token Reduction

arXiv: https://arxiv.org/abs/2606.31676 · HTML: https://mlsanity.com/paper/2606.31676/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 39 references that are in the map.
- First posted on arXiv 30 Jun 2026.

## Builds on (most cited first, 25 of 25)

- [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)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [Token Merging: Your ViT But Faster](https://mlsanity.com/paper/2210.09461/) (2022)
- [Vision Transformers Need Registers](https://mlsanity.com/paper/2309.16588/) (2023)
- [Quantifying Attention Flow in Transformers](https://mlsanity.com/paper/2005.00928/) (2020)
- [DeiT III: Revenge of the ViT](https://mlsanity.com/paper/2204.07118/) (2022)
- [Not All Patches are What You Need: Expediting Vision Transformers via Token Reorganizations](https://mlsanity.com/paper/2202.07800/) (2022)
- [AdaViT: Adaptive Tokens for Efficient Vision Transformer](https://mlsanity.com/paper/2112.07658/) (2021)
- [AdaViT: Adaptive Vision Transformers for Efficient Image Recognition](https://mlsanity.com/paper/2111.15668/) (2021)
- [Adaptive Token Sampling For Efficient Vision Transformers](https://mlsanity.com/paper/2111.15667/) (2021)
- [Evo-ViT: Slow-Fast Token Evolution for Dynamic Vision Transformer](https://mlsanity.com/paper/2108.01390/) (2021)
- [SPViT: Enabling Faster Vision Transformers via Soft Token Pruning](https://mlsanity.com/paper/2112.13890/) (2021)
- [Patch Slimming for Efficient Vision Transformers](https://mlsanity.com/paper/2106.02852/) (2021)
- [Federated Unlearning via Class-Discriminative Pruning](https://mlsanity.com/paper/2110.11794/) (2021)
- [Learning to Merge Tokens in Vision Transformers](https://mlsanity.com/paper/2202.12015/) (2022)
- [Beyond Attentive Tokens: Incorporating Token Importance and Diversity for Efficient Vision Transformers](https://mlsanity.com/paper/2211.11315/) (2022)
- [Teaching Matters: Investigating the Role of Supervision in Vision Transformers](https://mlsanity.com/paper/2212.03862/) (2022)
- [Robustifying Token Attention for Vision Transformers](https://mlsanity.com/paper/2303.11126/) (2023)
- [Faster or Stronger: Towards Flexible Visual Place Recognition via Weighted Aggregation and Token Pruning](https://mlsanity.com/paper/2605.20551/) (2026)
- [Visual-Word Tokenizer: Beyond Fixed Sets of Tokens in Vision Transformers](https://mlsanity.com/paper/2411.15397/) (2024)
- [Visual Words Meet BM25: Sparse Auto-Encoder Visual Word Scoring for Image Retrieval](https://mlsanity.com/paper/2603.05781/) (2026)
- [Structural Assessment for Understanding and Guiding Dataset Distillation in Discrete Token Space](https://mlsanity.com/paper/2606.21705/) (2026)
- [Rethinking Prototype-based Similarity Learning for Few-Shot Object Detection](https://mlsanity.com/paper/2606.23069/) (2026)

## Abstract (arXiv)

Most token reduction methods for Vision Transformers seek favorable tradeoffs between accuracy and efficiency by pruning, merging, or pooling patch tokens. REDI (Relevance for DINOv3 Token Reduction) studies this question through a controlled supervised reference: how should a fixed token budget be allocated across patches for image classification? REDI quantizes final block DINOv3 patch representations into a visual vocabulary and derives class conditioned corpus scores using supervised TF-IDF over visual words. For each validation image, the ground truth class selects a row of the TF-IDF table, and four transformed views produce a TF-IDF map aligned to a reference center crop. A separate dense pass on the same crop provides an attention map. After independent min max normalization, their elementwise product defines the REDI score. A fixed keep, merge, and compress operator then uses score rank to assign patch roles and score magnitude to weight merging and compression. With precomputed REDI scores, a frozen DINOv3 ViT-B/16 backbone, and the same linear classifier used for dense evaluation, the operator reduces the sequence length from 201 to 107 tokens, a 46.8% sequence reduction. The REDI variant based on incoming attention mass achieves 84.706% Top-1 accuracy on ImageNet-1K, compared with 83.514% for the dense baseline, 82.634% for incoming attention mass alone, and 81.796% for supervised TF-IDF alone. The same corpus term also improves reduced classification for three alternative attention formulations relative to their attention only counterparts. Together, these controlled comparisons indicate that class specific corpus statistics and image specific attention provide complementary signals for patch ranking in this setting.