mlsanityData to 7 Oct 2026

Patch Similarity Aware Data-Free Quantization for Vision Transformers

arXiv 2203.02250 ↗First posted 4 Mar 2022cs.CVcited by 55

Not on the map. Why?

Abstract (arXiv)

Vision transformers have recently gained great success on various computer vision tasks; nevertheless, their high model complexity makes it challenging to deploy on resource-constrained devices. Quantization is an effective approach to reduce model complexity, and data-free quantization, which can address data privacy and security concerns during model deployment, has received widespread interest. Unfortunately, all existing methods, such as BN regularization, were designed for convolutional neural networks and cannot be applied to vision transformers with significantly different model architectures. In this paper, we propose PSAQ-ViT, a Patch Similarity Aware data-free Quantization framework for Vision Transformers, to enable the generation of "realistic" samples based on the vision transformer's unique properties for calibrating the quantization parameters. Specifically, we analyze the self-attention module's properties and reveal a general difference (patch similarity) in its processing of Gaussian noise and real images. The above insights guide us to design a relative value metric to optimize the Gaussian noise to approximate the real images, which are then utilized to calibrate the quantization parameters. Extensive experiments and ablation studies are conducted on various benchmarks to validate the effectiveness of PSAQ-ViT, which can even outperform the real-data-driven methods. Code is available at: https://github.com/zkkli/PSAQ-ViT.

Key facts. Cited by 55 papers in the map (data through 7 Oct 2026). Builds on 37 of its 43 references that are in the map. First posted on arXiv 4 Mar 2022.

Builds on

37 of 43 references.

References · 37

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  2. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,082
  3. 2020End-to-End Object Detection with Transformerscited by 7,384
  4. 2020Training data-efficient image transformers & distillation through attentioncited by 4,560
  5. 2020Deformable DETR: Deformable Transformers for End-to-End Object Detectioncited by 3,410
  6. 2013Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computationcited by 2,031
  7. 2021ViViT: A Video Vision Transformercited by 1,610
  8. 2017Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inferencecited by 1,548
  9. 2021Transformers in Vision: A Surveycited by 1,009
  10. 2020Pre-Trained Image Processing Transformercited by 891
  11. 2016XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networkscited by 829
  12. 2021A Survey of Quantization Methods for Efficient Neural Network Inferencecited by 798
25 more
  1. 2021Transformer in Transformercited by 632
  2. 2019Learned Step Size Quantizationcited by 604
  3. 2020Up or Down? Adaptive Rounding for Post-Training Quantizationcited by 600
  4. 2018Quantizing deep convolutional networks for efficient inference: A whitepapercited by 507
  5. 2021BRECQ: Pushing the Limit of Post-Training Quantization by Block Reconstructioncited by 464
  6. 2018PACT: Parameterized Clipping Activation for Quantized Neural Networkscited by 456
  7. 2019Dreaming to Distill: Data-free Knowledge Transfer via DeepInversioncited by 419
  8. 2021Transformer Trackingcited by 412
  9. 2021DeepViT: Towards Deeper Vision Transformercited by 288
  10. 2021Post-Training Quantization for Vision Transformercited by 274
  11. 2018LQ-Nets: Learned Quantization for Highly Accurate and Compact Deep Neural Networkscited by 272
  12. 2021Video Transformer Networkcited by 271
  13. 2020ZeroQ: A Novel Zero Shot Quantization Frameworkcited by 265
  14. 2021Rethinking and Improving Relative Position Encoding for Vision Transformercited by 206
  15. 2019Low-bit Quantization of Neural Networks for Efficient Inferencecited by 185
  16. 2021PTQ4ViT: Post-training quantization for vision transformers with twin uniform quantizationcited by 173
  17. 2019Additive Powers-of-Two Quantization: An Efficient Non-uniform Discretization for Neural Networkscited by 143
  18. 2021Patch Slimming for Efficient Vision Transformerscited by 131
  19. 2020Generative Low-bitwidth Data Free Quantizationcited by 87
  20. 2021Diversifying Sample Generation for Accurate Data-Free Quantizationcited by 63
  21. 2021Learning Efficient Vision Transformers via Fine-Grained Manifold Distillationcited by 60
  22. 2020EasyQuant: Post-training Quantization via Scale Optimizationcited by 57
  23. 2021IntraQ: Learning Synthetic Images with Intra-Class Heterogeneity for Zero-Shot Network Quantizationcited by 43
  24. 2020One Weight Bitwidth to Rule Them Allcited by 13
  25. 2020WaveQ: Gradient-Based Deep Quantization of Neural Networks through Sinusoidal Adaptive Regularizationcited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 3; 2023: 10; 2024: 15; 2025: 17; 2026: 10.

55 papers.

Sort:

By area:

  1. 2022RepQ-ViT: Scale Reparameterization for Post-Training Quantization of Vision Transformerscited by 134
  2. 2022I-ViT: Integer-only Quantization for Efficient Vision Transformer Inferencecited by 93
  3. 2024Q-VLM: Post-training Quantization for Large Vision-Language Modelscited by 47
  4. 2023Towards Accurate Post-training Quantization for Diffusion Modelscited by 46
  5. 2024A Survey on Transformer Compressioncited by 44
  6. 2022PSAQ-ViT V2: Towards Accurate and General Data-Free Quantization for Vision Transformerscited by 42
  7. 2023A Survey of Techniques for Optimizing Transformer Inferencecited by 35
  8. 2024EDA-DM: Enhanced Distribution Alignment for Post-Training Quantization of Diffusion Modelscited by 23
  9. 2023PowerQuant: Automorphism Search for Non-Uniform Quantizationcited by 20
  10. 2023Boost Vision Transformer with GPU-Friendly Sparsity and Quantizationcited by 20
  11. 2023I&S-ViT: An Inclusive & Stable Method for Pushing the Limit of Post-Training ViTs Quantizationcited by 19
  12. 2024DopQ-ViT: Towards Distribution-Friendly and Outlier-Aware Post-Training Quantization for Vision Transformerscited by 19
38 more
  1. 2024CLAMP-ViT: Contrastive Data-Free Learning for Adaptive Post-Training Quantization of ViTscited by 17
  2. 2024K-Sort Arena: Efficient and Reliable Benchmarking for Generative Models via K-wise Human Preferencescited by 16
  3. 2024Model Quantization and Hardware Acceleration for Vision Transformers: A Comprehensive Surveycited by 13
  4. 2025CacheQuant: Comprehensively Accelerated Diffusion Modelscited by 13
  5. 2025Low-bit Model Quantization for Deep Neural Networks: A Surveycited by 11
  6. 2023Transformers in Single Object Tracking: An Experimental Surveycited by 10
  7. 2023QFT: Quantized Full-parameter Tuning of LLMs with Affordable Resourcescited by 10
  8. 2024RepQuant: Towards Accurate Post-Training Quantization of Large Transformer Models via Scale Reparameterizationcited by 10
  9. 2025Mix-QViT: Mixed-Precision Vision Transformer Quantization Driven by Layer Importance and Quantization Sensitivitycited by 9
  10. 2023GenQ: Quantization in Low Data Regimes with Generative Synthetic Datacited by 8
  11. 2024LRP-QViT: Mixed-Precision Vision Transformer Quantization via Layer-wise Relevance Propagationcited by 8
  12. 2023BinaryViT: Towards Efficient and Accurate Binary Vision Transformerscited by 7
  13. 2024MimiQ: Low-Bit Data-Free Quantization of Vision Transformers with Encouraging Inter-Head Attention Similaritycited by 7
  14. 2025Zero-shot Quantization: A Comprehensive Surveycited by 7
  15. 2024PQD: Post-training Quantization for Efficient Diffusion Modelscited by 6
  16. 2025OuroMamba: A Data-Free Quantization Framework for Vision Mambacited by 5
  17. 2023GSB: Group Superposition Binarization for Vision Transformer with Limited Training Samplescited by 4
  18. 2024Privacy-Preserving SAM Quantization for Efficient Edge Intelligence in Healthcarecited by 4
  19. 2025Sparse Model Inversion: Efficient Inversion of Vision Transformers for Data-Free Applicationscited by 4
  20. 2024MGRQ: Post-Training Quantization For Vision Transformer With Mixed Granularity Reconstructioncited by 3
  21. 2024Semantic Alignment and Reinforcement for Data-Free Quantization of Vision Transformerscited by 3
  22. 2025AHCQ-SAM: Toward Accurate and Hardware-Compatible Post-Training Segment Anything Model Quantizationcited by 3
  23. 2025SAQ-SAM: Semantically-Aligned Quantization for Segment Anything Modelcited by 3
  24. 2026PTQ4ARVG: Post-Training Quantization for AutoRegressive Visual Generation Modelscited by 3
  25. 2026K-Sort Eval: Efficient Preference Evaluation for Visual Generation via Corrected VLM-as-a-Judgecited by 3
  26. 2025DFQ-ViT: Data-Free Quantization for Vision Transformers without Fine-tuningcited by 2
  27. 2025LampQ: Towards Accurate Layer-wise Mixed Precision Quantization for Vision Transformerscited by 2
  28. 2026Efficient-SAM2: Accelerating SAM2 with Object-Aware Visual Encoding and Memory Retrievalcited by 2
  29. 2026OSAQ: Outlier Self-Absorption for Accurate Low-bit LLM Quantizationcited by 2
  30. 2025Enhancing Generalization in Data-free Quantization via Mixup-class Promptingcited by 1
  31. 2025Exploiting Information Redundancy in Attention Maps for Extreme Quantization of Vision Transformerscited by 1
  32. 2025IPTQ-ViT: Post-Training Quantization of Non-linear Functions for Integer-only Vision Transformerscited by 1
  33. 2026SynQ: Accurate Zero-shot Quantization by Synthesis-aware Fine-tuningcited by 1
  34. 2026Selective Coupling of Decoupled Informative Regions: Masked Attention Alignment for Data-Free Quantization of Vision Transformerscited by 1
  35. 2024DataFreeShield: Defending Adversarial Attacks without Training Datacited by 0
  36. 2025UAV-Assisted Real-Time Disaster Detection Using Optimized Transformer Modelcited by 0
  37. 2025Binary Quadratic Quantization: Beyond First-Order Quantization for Real-Valued Matrix Compressioncited by 0
  38. 2025D4C: Data-Free Quantization for Contrastive Language-Image Pre-training Modelscited by 0

50 most cited of 55 All 55 (JSON)