# Dynamic Token Pruning in Plain Vision Transformers for Semantic Segmentation

arXiv: https://arxiv.org/abs/2308.01045 · HTML: https://mlsanity.com/paper/2308.01045/

## Key facts

- Cited by 62 papers in the map (data through 14 Oct 2026).
- Placed in Token pruning and merging for ViTs, one of 12 directions in Efficient LLM inference and serving.
- Builds on 29 of its 35 references that are in the map.
- First posted on arXiv 2 Aug 2023.

## Placed in

- Efficient LLM inference and serving / [Token pruning and merging for ViTs](https://mlsanity.com/inference/token-pruning-and-merging-for-vits/)

## Builds on (most cited first, 29 of 29)

- [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)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [Layer Normalization](https://mlsanity.com/paper/1607.06450/) (2016)
- [Fully Convolutional Networks for Semantic Segmentation](https://mlsanity.com/paper/1411.4038/) (2014)
- [A ConvNet for the 2020s](https://mlsanity.com/paper/2201.03545/) (2022)
- [Gaussian Error Linear Units (GELUs)](https://mlsanity.com/paper/1606.08415/) (2016)
- [Deformable DETR: Deformable Transformers for End-to-End Object Detection](https://mlsanity.com/paper/2010.04159/) (2020)
- [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://mlsanity.com/paper/2105.15203/) (2021)
- [Pyramid Scene Parsing Network](https://mlsanity.com/paper/1612.01105/) (2016)
- [A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networks](https://mlsanity.com/paper/1610.02136/) (2016)
- [Masked-attention Mask Transformer for Universal Image Segmentation](https://mlsanity.com/paper/2112.01527/) (2021)
- [Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformers](https://mlsanity.com/paper/2012.15840/) (2020)
- [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://mlsanity.com/paper/2107.06278/) (2021)
- [Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNet](https://mlsanity.com/paper/2101.11986/) (2021)
- [COCO-Stuff: Thing and Stuff Classes in Context](https://mlsanity.com/paper/1612.03716/) (2016)
- [DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification](https://mlsanity.com/paper/2106.02034/) (2021)
- [Segmenter: Transformer for Semantic Segmentation](https://mlsanity.com/paper/2105.05633/) (2021)
- [AdaViT: Adaptive Tokens for Efficient Vision Transformer](https://mlsanity.com/paper/2112.07658/) (2021)
- [Not All Images are Worth 16x16 Words: Dynamic Transformers for Efficient Image Recognition](https://mlsanity.com/paper/2105.15075/) (2021)
- [Context Prior for Scene Segmentation](https://mlsanity.com/paper/2004.01547/) (2020)
- [Learning a Discriminative Feature Network for Semantic Segmentation](https://mlsanity.com/paper/1804.09337/) (2018)
- [SegViT: Semantic Segmentation with Plain Vision Transformers](https://mlsanity.com/paper/2210.05844/) (2022)
- [SparseViT: Revisiting Activation Sparsity for Efficient High-Resolution Vision Transformer](https://mlsanity.com/paper/2303.17605/) (2023)
- [Not All Pixels Are Equal: Difficulty-aware Semantic Segmentation via Deep Layer Cascade](https://mlsanity.com/paper/1704.01344/) (2017)
- [Expediting Large-Scale Vision Transformer for Dense Prediction without Fine-tuning](https://mlsanity.com/paper/2210.01035/) (2022)
- [Content-aware Token Sharing for Efficient Semantic Segmentation with Vision Transformers](https://mlsanity.com/paper/2306.02095/) (2023)

## Built on by (most cited first, 30 of 62)

- [LOOK-M: Look-Once Optimization in KV Cache for Efficient Multimodal Long-Context Inference](https://mlsanity.com/paper/2406.18139/) (2024)
- [AIM: Adaptive Inference of Multi-Modal LLMs via Token Merging and Pruning](https://mlsanity.com/paper/2412.03248/) (2024)
- [SegMAN: Omni-scale Context Modeling with State Space Models and Local Attention for Semantic Segmentation](https://mlsanity.com/paper/2412.11890/) (2024)
- [ALGM: Adaptive Local-then-Global Token Merging for Efficient Semantic Segmentation with Plain Vision Transformers](https://mlsanity.com/paper/2406.09936/) (2024)
- [Token Cropr: Faster ViTs for Quite a Few Tasks](https://mlsanity.com/paper/2412.00965/) (2024)
- [Hierarchical Graph Interaction Transformer with Dynamic Token Clustering for Camouflaged Object Detection](https://mlsanity.com/paper/2408.15020/) (2024)
- [SLAB: Efficient Transformers with Simplified Linear Attention and Progressive Re-parameterized Batch Normalization](https://mlsanity.com/paper/2405.11582/) (2024)
- [SparseVILA: Decoupling Visual Sparsity for Efficient VLM Inference](https://mlsanity.com/paper/2510.17777/) (2025)
- [Embedding-Free Transformer with Inference Spatial Reduction for Efficient Semantic Segmentation](https://mlsanity.com/paper/2407.17261/) (2024)
- [VLTP: Vision-Language Guided Token Pruning for Task-Oriented Segmentation](https://mlsanity.com/paper/2409.08464/) (2024)
- [The Model Knows Which Tokens Matter: Automatic Token Selection via Noise Gating](https://mlsanity.com/paper/2603.07135/) (2026)
- [Co-Me: Confidence-Guided Token Merging for Visual Geometric Transformers](https://mlsanity.com/paper/2511.14751/) (2025)
- [Efficient Partitioning Vision Transformer on Edge Devices for Distributed Inference](https://mlsanity.com/paper/2410.11650/) (2024)
- [A Survey on Dynamic Neural Networks: from Computer Vision to Multi-modal Sensor Fusion](https://mlsanity.com/paper/2501.07451/) (2025)
- [Redundant Queries in DETR-Based 3D Detection Methods: Unnecessary and Prunable](https://mlsanity.com/paper/2412.02054/) (2024)
- [AutoPrune: Each Complexity Deserves a Pruning Policy](https://mlsanity.com/paper/2509.23931/) (2025)
- [Generative Event Pretraining with Foundation Model Alignment](https://mlsanity.com/paper/2603.23032/) (2026)
- [Automatic Channel Pruning for Multi-Head Attention](https://mlsanity.com/paper/2405.20867/) (2024)
- [DUDA: Distilled Unsupervised Domain Adaptation for Lightweight Semantic Segmentation](https://mlsanity.com/paper/2504.09814/) (2025)
- [S2AFormer: Strip Self-Attention for Efficient Vision Transformer](https://mlsanity.com/paper/2505.22195/) (2025)
- [Purge-Gate: Backpropagation-Free Test-Time Adaptation for Point Clouds Classification via Token Purging](https://mlsanity.com/paper/2509.09785/) (2025)
- [Fast SAM2 with Text-Driven Token Pruning](https://mlsanity.com/paper/2512.21333/) (2025)
- [Faster or Stronger: Towards Flexible Visual Place Recognition via Weighted Aggregation and Token Pruning](https://mlsanity.com/paper/2605.20551/) (2026)
- [SinkPruner: Sink-Free Visual Token Pruning for Multimodal Large Language Models](https://mlsanity.com/paper/2609.01004/) (2026)
- [Neural Slot Interpreters: Grounding Object Semantics in Emergent Slot Representations](https://mlsanity.com/paper/2403.07887/) (2024)
- [When Less is Enough: Adaptive Token Reduction for Efficient Image Representation](https://mlsanity.com/paper/2503.16660/) (2025)
- [StructSAM: Structure- and Spectrum-Preserving Token Merging for Segment Anything Models](https://mlsanity.com/paper/2603.07307/) (2026)
- [ARTA: Adaptive Mixed-Resolution Token Allocation for Efficient Dense Feature Extraction](https://mlsanity.com/paper/2603.26258/) (2026)
- [Fusion: A Framework for Unified Sequential Token AdaptatIon in VisiOn TraNsformers](https://mlsanity.com/paper/2607.02612/) (2026)
- [HCPM: Hierarchical Candidates Pruning for Efficient Detector-Free Matching](https://mlsanity.com/paper/2403.12543/) (2024)

## Abstract (arXiv)

Vision transformers have achieved leading performance on various visual tasks yet still suffer from high computational complexity. The situation deteriorates in dense prediction tasks like semantic segmentation, as high-resolution inputs and outputs usually imply more tokens involved in computations. Directly removing the less attentive tokens has been discussed for the image classification task but can not be extended to semantic segmentation since a dense prediction is required for every patch. To this end, this work introduces a Dynamic Token Pruning (DToP) method based on the early exit of tokens for semantic segmentation. Motivated by the coarse-to-fine segmentation process by humans, we naturally split the widely adopted auxiliary-loss-based network architecture into several stages, where each auxiliary block grades every token's difficulty level. We can finalize the prediction of easy tokens in advance without completing the entire forward pass. Moreover, we keep k highest confidence tokens for each semantic category to uphold the representative context information. Thus, computational complexity will change with the difficulty of the input, akin to the way humans do segmentation. Experiments suggest that the proposed DToP architecture reduces on average 20% - 35% of computational cost for current semantic segmentation methods based on plain vision transformers without accuracy degradation.