mlsanityData to 7 Oct 2026

PPT: Token Pruning and Pooling for Efficient Vision Transformers

arXiv 2310.01812 ↗First posted 3 Oct 2023cs.CVcited by 37

On the map: Efficient LLM inference and serving → Token pruning and merging for ViTs

Abstract (arXiv)

Vision Transformers (ViTs) have emerged as powerful models in the field of computer vision, delivering superior performance across various vision tasks. However, the high computational complexity poses a significant barrier to their practical applications in real-world scenarios. Motivated by the fact that not all tokens contribute equally to the final predictions and fewer tokens bring less computational cost, reducing redundant tokens has become a prevailing paradigm for accelerating vision transformers. However, we argue that it is not optimal to either only reduce inattentive redundancy by token pruning, or only reduce duplicative redundancy by token merging. To this end, in this paper we propose a novel acceleration framework, namely token Pruning & Pooling Transformers (PPT), to adaptively tackle these two types of redundancy in different layers. By heuristically integrating both token pruning and token pooling techniques in ViTs without additional trainable parameters, PPT effectively reduces the model complexity while maintaining its predictive accuracy. For example, PPT reduces over 37% FLOPs and improves the throughput by over 45% for DeiT-S without any accuracy drop on the ImageNet dataset. The code is available at https://github.com/xjwu1024/PPT and https://github.com/mindspore-lab/models/

Key facts. Cited by 37 papers in the map (data through 7 Oct 2026). Placed in Token pruning and merging for ViTs, one of 12 directions in Efficient LLM inference and serving. Builds on 31 of its 38 references that are in the map. First posted on arXiv 3 Oct 2023.

Builds on

31 of 38 references.

Closest ancestors · 3

Same direction, newest first

  1. 2023Joint Token Pruning and Squeezing Towards More Aggressive Compression of Vision Transformers21 Apr
  2. 2022Token Merging: Your ViT But Faster17 Oct
  3. 2021Evo-ViT: Slow-Fast Token Evolution for Dynamic Vision Transformer3 Aug

Other references · 28

  1. 2017Attention Is All You Needcited by 59,932
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2020Language Models are Few-Shot Learnerscited by 32,199
  4. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  5. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,082
  6. 2023Segment Anythingcited by 8,158
  7. 2020End-to-End Object Detection with Transformerscited by 7,384
  8. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,893
  9. 2020Training data-efficient image transformers & distillation through attentioncited by 4,560
  10. 2020Deformable DETR: Deformable Transformers for End-to-End Object Detectioncited by 3,410
  11. 2021Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutionscited by 1,960
  12. 2021Tokens-to-Token ViT: Training Vision Transformers from Scratch on ImageNetcited by 1,096
16 more
  1. 2020Designing Network Design Spacescited by 954
  2. 2021CvT: Introducing Convolutions to Vision Transformerscited by 946
  3. 2021DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsificationcited by 849
  4. 2021CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classificationcited by 806
  5. 2022Exploring Plain Vision Transformer Backbones for Object Detectioncited by 723
  6. 2021Transformer in Transformercited by 632
  7. 2021Rethinking Spatial Dimensions of Vision Transformerscited by 361
  8. 2021AdaViT: Adaptive Tokens for Efficient Vision Transformercited by 360
  9. 2022Not All Patches are What You Need: Expediting Vision Transformers via Token Reorganizationscited by 359
  10. 2021Co-Scale Conv-Attentional Image Transformerscited by 224
  11. 2021Adaptive Token Sampling For Efficient Vision Transformerscited by 181
  12. 2021All Tokens Matter: Token Labeling for Training Better Vision Transformerscited by 144
  13. 2021Patch Slimming for Efficient Vision Transformerscited by 131
  14. 2021IA-RED2: Interpretability-Aware Redundancy Reduction for Vision Transformerscited by 127
  15. 2021Token Pooling in Vision Transformerscited by 68
  16. 2021Vision Transformer with Progressive Samplingcited by 44

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 15; 2025: 16; 2026: 6.

37 papers.

Sort:

By area:

  1. 2024DiTFastAttn: Attention Compression for Diffusion Transformer Modelscited by 109
  2. 2025Token Reduction Should Go Beyond Efficiency in Generative Models -- From Vision, Language to Multimodalitycited by 49
  3. 2025Similarity-Aware Token Pruning: Your VLM but Fastercited by 32
  4. 2024D2O: Dynamic Discriminative Operations for Efficient Long-Context Inference of Large Language Modelscited by 30
  5. 2024ALGM: Adaptive Local-then-Global Token Merging for Efficient Semantic Segmentation with Plain Vision Transformerscited by 28
  6. 2024Importance-Based Token Merging for Efficient Image and Video Generationcited by 20
  7. 2024ST3: Accelerating Multimodal Large Language Model by Spatial-Temporal Visual Token Trimmingcited by 20
  8. 2024Token Cropr: Faster ViTs for Quite a Few Taskscited by 18
  9. 2025MambaIC: State Space Models for High-Performance Learned Image Compressioncited by 18
  10. 2024SLAB: Efficient Transformers with Simplified Linear Attention and Progressive Re-parameterized Batch Normalizationcited by 17
  11. 2024LLMCBench: Benchmarking Large Language Model Compression for Efficient Deploymentcited by 16
  12. 2024Turbo: Informativity-Driven Acceleration Plug-In for Vision-Language Large Modelscited by 13
25 more
  1. 2024Agglomerative Token Clusteringcited by 13
  2. 2025Grouping First, Attending Smartly: Training-Free Acceleration for Diffusion Transformerscited by 13
  3. 2025FlexiDiT: Your Diffusion Transformer Can Easily Generate High-Quality Samples with Less Computecited by 12
  4. 2025Accelerating Vision Transformers with Adaptive Patch Sizescited by 12
  5. 2025ToSA: Token Merging with Spatial Awarenesscited by 11
  6. 2025DyMU: Dynamic Merging and Virtual Unmerging for Efficient VLMscited by 10
  7. 2025Token Transforming: A Unified and Training-Free Token Compression Framework for Vision Transformer Accelerationcited by 9
  8. 2025Continual Learning for Generative AI: From LLMs to MLLMs and Beyondcited by 9
  9. 2024You Only Need Less Attention at Each Stage in Vision Transformerscited by 7
  10. 2026Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradientscited by 7
  11. 2024ExCP: Extreme LLM Checkpoint Compression via Weight-Momentum Joint Shrinkingcited by 5
  12. 2024Token Compensator: Altering Inference Cost of Vision Transformer without Re-Tuningcited by 5
  13. 2025TinyChemVL: Advancing Chemical Vision-Language Models via Efficient Visual Token Reduction and Complex Reaction Taskscited by 5
  14. 2024Efficient Sparse Attention needs Adaptive Token Releasecited by 4
  15. 2025Purge-Gate: Backpropagation-Free Test-Time Adaptation for Point Clouds Classification via Token Purgingcited by 3
  16. 2025GreedyPrune: Retenting Critical Visual Token Set for Large Vision Language Modelscited by 2
  17. 2026Collaborative Multi-Mode Pruning for Vision-Language Modelscited by 2
  18. 2026MPM: Mutual Pair Merging for Efficient Vision Transformerscited by 2
  19. 2024Sparsity Meets Similarity: Leveraging Long-Tail Distribution for Dynamic Optimized Token Representation in Multimodal Large Language Modelscited by 1
  20. 2025LFTR: Learning-Free Token Reduction for Multimodal Large Language Modelscited by 1
  21. 2025Towards Efficient Vision State Space Models via Token Mergingcited by 0
  22. 2025Where Do Tokens Go? Understanding Pruning Behaviors in STEP at High Resolutionscited by 0
  23. 2026Learning Dynamic Evidence Routes for Vision Transformer Probingcited by 0
  24. 2026ASAP: Attention Sink Anchored Pruningcited by 0
  25. 2026RS3-Prune: Read-Sparse, Store-Sparse Token Pruning for Video Object Segmentationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2021SPViT: Enabling Faster Vision Transformers via Soft Token Pruning
  2. 2025HEART-VIT: Hessian-Guided Efficient Dynamic Attention and Token Pruning in Vision Transformer
  3. 2023No Token Left Behind: Efficient Vision Transformer via Dynamic Token Idling
  4. 2026RAPID: Layer-Wise Redundancy-Aware Pruning and Importance-Driven Token Merging for Efficient ViT
  5. 2026MergeOver: Post-Training Token Merging for Recursive Vision Transformers
  6. 2026ToaSt: Token Channel Selection and Structured Pruning for Efficient ViT
  7. 2022SaiT: Sparse Vision Transformers through Adaptive Token Pruning
  8. 2023Token Fusion: Bridging the Gap between Token Pruning and Token Merging