mlsanityData to 13 Oct 2026

Decay Pruning Method: Smooth Pruning With a Self-Rectifying Procedure

arXiv 2406.03879 ↗First posted 6 Jun 2024cs.LG cs.CVcited by 1

Not on the map. Why?

Abstract (arXiv)

Current structured pruning methods often result in considerable accuracy drops due to abrupt network changes and loss of information from pruned structures. To address these issues, we introduce the Decay Pruning Method (DPM), a novel smooth pruning approach with a self-rectifying mechanism. DPM consists of two key components: (i) Smooth Pruning: It converts conventional single-step pruning into multi-step smooth pruning, gradually reducing redundant structures to zero over N steps with ongoing optimization. (ii) Self-Rectifying: This procedure further enhances the aforementioned process by rectifying sub-optimal pruning based on gradient information. Our approach demonstrates strong generalizability and can be easily integrated with various existing pruning methods. We validate the effectiveness of DPM by integrating it with three popular pruning methods: OTOv2, Depgraph, and Gate Decorator. Experimental results show consistent improvements in performance compared to the original pruning methods, along with further reductions of FLOPs in most scenarios.

Key facts. Cited by 1 paper in the map (data through 13 Oct 2026). Builds on 44 of its 59 references that are in the map. First posted on arXiv 6 Jun 2024.

Builds on

44 of 59 references.

References · 44

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,950
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,656
  3. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,321
  4. 2016Pruning Filters for Efficient ConvNetscited by 1,138
  5. 2017Channel Pruning for Accelerating Very Deep Neural Networkscited by 701
  6. 2017Learning Efficient Convolutional Networks through Network Slimmingcited by 675
  7. 2024YOLO-World: Real-Time Open-Vocabulary Object Detectioncited by 500
  8. 2019Rigging the Lottery: Making All Tickets Winnerscited by 493
  9. 2018AMC: AutoML for Model Compression and Acceleration on Mobile Devicescited by 377
  10. 2018Filter Pruning via Geometric Median for Deep Convolutional Neural Networks Accelerationcited by 367
  11. 2017Scalable Training of Artificial Neural Networks with Adaptive Sparse Connectivity inspired by Network Sciencecited by 326
  12. 2018Soft Filter Pruning for Accelerating Deep Convolutional Neural Networkscited by 315
32 more
  1. 2016Network Trimming: A Data-Driven Neuron Pruning Approach towards Efficient Deep Architecturescited by 221
  2. 2019Sparse Networks from Scratch: Faster Training without Losing Performancecited by 212
  3. 2023DepGraph: Towards Any Structural Pruningcited by 203
  4. 2023A Survey on Deep Neural Network Pruning-Taxonomy, Comparison, Analysis, and Recommendationscited by 201
  5. 2023Tracking Everything Everywhere All at Oncecited by 191
  6. 2017BlockDrop: Dynamic Inference Paths in Residual Networkscited by 187
  7. 2017Data-Driven Sparse Structure Selection for Deep Neural Networkscited by 181
  8. 2021Reviving Iterative Training with Mask Guidance for Interactive Segmentationcited by 163
  9. 2017Deep Rewiring: Training very sparse deep networkscited by 137
  10. 2020Dynamic Model Pruning with Feedbackcited by 125
  11. 2019Gate Decorator: Global Filter Pruning Method for Accelerating Deep Convolutional Neural Networkscited by 118
  12. 2017Bayesian Compression for Deep Learningcited by 116
  13. 2019EigenDamage: Structured Pruning in the Kronecker-Factored Eigenbasiscited by 98
  14. 2020Neural Pruning via Growing Regularizationcited by 97
  15. 2020ResRep: Lossless CNN Pruning via Decoupling Remembering and Forgettingcited by 85
  16. 2021Sparse Training via Boosting Pruning Plasticity with Neuroregenerationcited by 84
  17. 2020Operation-Aware Soft Channel Pruning using Differentiable Maskscited by 74
  18. 2020Group Sparsity: The Hinge Between Filter Pruning and Decomposition for Network Compressioncited by 72
  19. 2021Only Train Once: A One-Shot Neural Network Training And Pruning Frameworkcited by 70
  20. 2024An Open and Comprehensive Pipeline for Unified Object Grounding and Detectioncited by 64
  21. 2023YOLOv6 v3.0: A Full-Scale Reloadingcited by 58
  22. 2019DeepHoyer: Learning Sparser Neural Network with Differentiable Scale-Invariant Sparsity Measurescited by 56
  23. 2019PruneTrain: Fast Neural Network Training by Dynamic Sparse Model Reconfigurationcited by 41
  24. 2017Structured Bayesian Pruning via Log-Normal Multiplicative Noisecited by 39
  25. 2017An Entropy-based Pruning Method for CNN Compressioncited by 37
  26. 2022SPDY: Accurate Pruning with Speedup Guaranteescited by 34
  27. 2023OTOV2: Automatic, Generic, User-Friendlycited by 32
  28. 2018Asymptotic Soft Filter Pruning for Deep Convolutional Neural Networkscited by 27
  29. 2018Accelerate CNN via Recursive Bayesian Pruningcited by 27
  30. 2024Auto-Train-Once: Controller Network Guided Automatic Network Pruning from Scratchcited by 20
  31. 2022N2NSkip: Learning Highly Sparse Networks using Neuron-to-Neuron Skip Connectionscited by 9
  32. 2022Ensemble Knowledge Guided Sub-network Search and Fine-tuning for Filter Pruningcited by 5

Built on by

Cited by year

Citing papers per year: 2025: 1.

1 paper.