# Expediting Large-Scale Vision Transformer for Dense Prediction without Fine-tuning

arXiv: https://arxiv.org/abs/2210.01035 · HTML: https://mlsanity.com/paper/2210.01035/

## Key facts

- Cited by 33 papers in the map (data through 14 Oct 2026).
- Builds on 86 of its 124 references that are in the map.
- First posted on arXiv 3 Oct 2022.

## Builds on (most cited first, 30 of 86)

- [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)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [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)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://mlsanity.com/paper/1905.11946/) (2019)
- [The Cityscapes Dataset for Semantic Urban Scene Understanding](https://mlsanity.com/paper/1604.01685/) (2016)
- [Going Deeper with Convolutions](https://mlsanity.com/paper/1409.4842/) (2014)
- [Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision](https://mlsanity.com/paper/2102.05918/) (2021)
- [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://mlsanity.com/paper/2105.15203/) (2021)
- [BEiT: BERT Pre-Training of Image Transformers](https://mlsanity.com/paper/2106.08254/) (2021)
- [Masked-attention Mask Transformer for Universal Image Segmentation](https://mlsanity.com/paper/2112.01527/) (2021)
- [Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions](https://mlsanity.com/paper/2102.12122/) (2021)
- [The Kinetics Human Action Video Dataset](https://mlsanity.com/paper/1705.06950/) (2017)
- [SwinIR: Image Restoration Using Swin Transformer](https://mlsanity.com/paper/2108.10257/) (2021)
- [Vision Transformers for Dense Prediction](https://mlsanity.com/paper/2103.13413/) (2021)
- [Reformer: The Efficient Transformer](https://mlsanity.com/paper/2001.04451/) (2020)
- [Rethinking Attention with Performers](https://mlsanity.com/paper/2009.14794/) (2020)
- [Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time](https://mlsanity.com/paper/2203.05482/) (2022)
- [DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection](https://mlsanity.com/paper/2203.03605/) (2022)
- [Swin Transformer V2: Scaling Up Capacity and Resolution](https://mlsanity.com/paper/2111.09883/) (2021)
- [LVIS: A Dataset for Large Vocabulary Instance Segmentation](https://mlsanity.com/paper/1908.03195/) (2019)
- [Grounded Language-Image Pre-training](https://mlsanity.com/paper/2112.03857/) (2021)
- [SimMIM: A Simple Framework for Masked Image Modeling](https://mlsanity.com/paper/2111.09886/) (2021)
- [Video Swin Transformer](https://mlsanity.com/paper/2106.13230/) (2021)
- [Semantic Understanding of Scenes through the ADE20K Dataset](https://mlsanity.com/paper/1608.05442/) (2016)

## Built on by (most cited first, 30 of 33)

- [Transformer-Based Visual Segmentation: A Survey](https://mlsanity.com/paper/2304.09854/) (2023)
- [VLT: Vision-Language Transformer and Query Generation for Referring Segmentation](https://mlsanity.com/paper/2210.15871/) (2022)
- [Convolution Meets LoRA: Parameter Efficient Finetuning for Segment Anything Model](https://mlsanity.com/paper/2401.17868/) (2024)
- [Dynamic Token Pruning in Plain Vision Transformers for Semantic Segmentation](https://mlsanity.com/paper/2308.01045/) (2023)
- [Token Reduction Should Go Beyond Efficiency in Generative Models -- From Vision, Language to Multimodality](https://mlsanity.com/paper/2505.18227/) (2025)
- [CrossGET: Cross-Guided Ensemble of Tokens for Accelerating Vision-Language Transformers](https://mlsanity.com/paper/2305.17455/) (2023)
- [ALGM: Adaptive Local-then-Global Token Merging for Efficient Semantic Segmentation with Plain Vision Transformers](https://mlsanity.com/paper/2406.09936/) (2024)
- [SlimSAM: 0.1% Data Makes Segment Anything Slim](https://mlsanity.com/paper/2312.05284/) (2023)
- [Parameter Efficient Fine-tuning of Self-supervised ViTs without Catastrophic Forgetting](https://mlsanity.com/paper/2404.17245/) (2024)
- [On Efficient Variants of Segment Anything Model: A Survey](https://mlsanity.com/paper/2410.04960/) (2024)
- [Self-Regularized Prototypical Network for Few-Shot Semantic Segmentation](https://mlsanity.com/paper/2210.16829/) (2022)
- [TCFormer: Visual Recognition via Token Clustering Transformer](https://mlsanity.com/paper/2407.11321/) (2024)
- [AiluRus: A Scalable ViT Framework for Dense Prediction](https://mlsanity.com/paper/2311.01197/) (2023)
- [Token Transforming: A Unified and Training-Free Token Compression Framework for Vision Transformer Acceleration](https://mlsanity.com/paper/2506.05709/) (2025)
- [ClusterFormer: Clustering As A Universal Visual Learner](https://mlsanity.com/paper/2309.13196/) (2023)
- [ProMotion: Prototypes As Motion Learners](https://mlsanity.com/paper/2406.04999/) (2024)
- [Efficient Token Compression for Vision Transformer with Spatial Information Preserved](https://mlsanity.com/paper/2503.23455/) (2025)
- [Fast SAM 3D Body: Accelerating SAM 3D Body for Real-Time Full-Body Human Mesh Recovery](https://mlsanity.com/paper/2603.15603/) (2026)
- [Training-Free Acceleration of ViTs with Delayed Spatial Merging](https://mlsanity.com/paper/2303.02331/) (2023)
- [Mask Frozen-DETR: High Quality Instance Segmentation with One GPU](https://mlsanity.com/paper/2308.03747/) (2023)
- [Efficient Data Driven Mixture-of-Expert Extraction from Trained Networks](https://mlsanity.com/paper/2505.15414/) (2025)
- [TORE: Token Recycling in Vision Transformers for Efficient Active Visual Exploration](https://mlsanity.com/paper/2311.15335/) (2023)
- [Expediting Contrastive Language-Image Pretraining via Self-distilled Encoders](https://mlsanity.com/paper/2312.12659/) (2023)
- [StructSAM: Structure- and Spectrum-Preserving Token Merging for Segment Anything Models](https://mlsanity.com/paper/2603.07307/) (2026)
- [MPM: Mutual Pair Merging for Efficient Vision Transformers](https://mlsanity.com/paper/2604.05718/) (2026)
- [GasTwinFormer: A Hybrid Vision Transformer for Livestock Methane Emission Segmentation and Dietary Classification in Optical Gas Imaging](https://mlsanity.com/paper/2508.15057/) (2025)
- [PicoSAM3: Real-Time In-Sensor Region-of-Interest Segmentation](https://mlsanity.com/paper/2603.11917/) (2026)
- [SparseSAM: Structured Sparsification of Activations in Segment Anything Models](https://mlsanity.com/paper/2605.17633/) (2026)
- [Predicting Token Impact Towards Efficient Vision Transformer](https://mlsanity.com/paper/2305.14840/) (2023)
- [Temporal Cluster Assignment for Efficient Real-Time Video Segmentation](https://mlsanity.com/paper/2508.05851/) (2025)

## Abstract (arXiv)

Vision transformers have recently achieved competitive results across various vision tasks but still suffer from heavy computation costs when processing a large number of tokens. Many advanced approaches have been developed to reduce the total number of tokens in large-scale vision transformers, especially for image classification tasks. Typically, they select a small group of essential tokens according to their relevance with the class token, then fine-tune the weights of the vision transformer. Such fine-tuning is less practical for dense prediction due to the much heavier computation and GPU memory cost than image classification. In this paper, we focus on a more challenging problem, i.e., accelerating large-scale vision transformers for dense prediction without any additional re-training or fine-tuning. In response to the fact that high-resolution representations are necessary for dense prediction, we present two non-parametric operators, a token clustering layer to decrease the number of tokens and a token reconstruction layer to increase the number of tokens. The following steps are performed to achieve this: (i) we use the token clustering layer to cluster the neighboring tokens together, resulting in low-resolution representations that maintain the spatial structures; (ii) we apply the following transformer layers only to these low-resolution representations or clustered tokens; and (iii) we use the token reconstruction layer to re-create the high-resolution representations from the refined low-resolution representations. The results obtained by our method are promising on five dense prediction tasks, including object detection, semantic segmentation, panoptic segmentation, instance segmentation, and depth estimation.