mlsanityData to 14 Oct 2026

Structural Pruning for Diffusion Models

arXiv 2305.10924 ↗First posted 18 May 2023cs.LG cs.AI cs.CVcited by 195

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Generative modeling has recently undergone remarkable advancements, primarily propelled by the transformative implications of Diffusion Probabilistic Models (DPMs). The impressive capability of these models, however, often entails significant computational overhead during both training and inference. To tackle this challenge, we present Diff-Pruning, an efficient compression method tailored for learning lightweight diffusion models from pre-existing ones, without the need for extensive re-training. The essence of Diff-Pruning is encapsulated in a Taylor expansion over pruned timesteps, a process that disregards non-contributory diffusion steps and ensembles informative gradients to identify important weights. Our empirical assessment, undertaken across several datasets highlights two primary benefits of our proposed method: 1) Efficiency: it enables approximately a 50% reduction in FLOPs at a mere 10% to 20% of the original training expenditure; 2) Consistency: the pruned diffusion models inherently preserve generative behavior congruent with their pre-trained models. Code is available at urlhttps://github.com/VainF/Diff-Pruning.

Key facts. Cited by 195 papers in the map (data through 14 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 45 of its 60 references that are in the map. First posted on arXiv 18 May 2023.

Builds on

45 of 60 references.

Closest ancestors · 12

Same direction, newest first

  1. 2023Patch Diffusion: Faster and More Data-Efficient Training of Diffusion Models25 Apr
  2. 2023Efficient Diffusion Training via Min-SNR Weighting Strategy16 Mar
  3. 2022Wavelet Diffusion Models are fast and scalable Image Generators29 Nov
  4. 2022Diffusion Probabilistic Model Made Slim27 Nov
  5. 2022Accelerating Diffusion Models via Early Stop of the Diffusion Process25 May
  6. 2022Subspace Diffusion Generative Models3 May
  7. 2022Pseudo Numerical Methods for Diffusion Models on Manifolds20 Feb
  8. 2021High-Resolution Image Synthesis with Latent Diffusion Models20 Dec
  9. 2021UNIT-DDPM: UNpaired Image Translation with Denoising Diffusion Probabilistic Models12 Apr
  10. 2021Improved Denoising Diffusion Probabilistic Models18 Feb
  11. 2020Denoising Diffusion Implicit Models6 Oct
  12. 2020Denoising Diffusion Probabilistic Models19 Jun

Other references · 33

  1. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  2. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,585
  3. 2023Adding Conditional Control to Text-to-Image Diffusion Modelscited by 5,503
  4. 2018The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networkscited by 2,262
  5. 2018Group Normalizationcited by 1,524
  6. 2015LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loopcited by 1,355
  7. 2016Pruning Filters for Efficient ConvNetscited by 1,143
  8. 2017Channel Pruning for Accelerating Very Deep Neural Networkscited by 702
  9. 2017Learning Efficient Convolutional Networks through Network Slimmingcited by 676
  10. 2019Importance Estimation for Neural Network Pruningcited by 541
  11. 2017ThiNet: A Filter Level Pruning Method for Deep Neural Network Compressioncited by 442
  12. 2020Movement Pruning: Adaptive Sparsity by Fine-Tuningcited by 403
21 more
  1. 2018AMC: AutoML for Model Compression and Acceleration on Mobile Devicescited by 378
  2. 2018Filter Pruning via Geometric Median for Deep Convolutional Neural Networks Accelerationcited by 367
  3. 2022DiffusionDet: Diffusion Model for Object Detectioncited by 354
  4. 2021SegDiff: Image Segmentation with Diffusion Probabilistic Modelscited by 280
  5. 2017Learning to Prune Deep Neural Networks via Layer-wise Optimal Brain Surgeoncited by 221
  6. 2023DepGraph: Towards Any Structural Pruningcited by 203
  7. 2021Group Fisher Pruning for Practical Network Compressioncited by 131
  8. 2019Gate Decorator: Global Filter Pruning Method for Accelerating Deep Convolutional Neural Networkscited by 118
  9. 2022Dataset Distillation via Factorizationcited by 115
  10. 2018Rethinking the Smaller-Norm-Less-Informative Assumption in Channel Pruning of Convolution Layerscited by 112
  11. 2019A Signal Propagation Perspective for Pruning Neural Networks at Initializationcited by 96
  12. 2022The Unreasonable Effectiveness of Random Pruning: Return of the Most Naive Baseline for Sparse Trainingcited by 85
  13. 2022Deep Model Reassemblycited by 78
  14. 2019Centripetal SGD for Pruning Very Deep Convolutional Networks with Complicated Structurecited by 76
  15. 2019Towards Efficient Model Compression via Learned Global Rankingcited by 75
  16. 2023Diffusion Action Segmentationcited by 74
  17. 2020Lookahead: A Far-Sighted Alternative of Magnitude-based Pruningcited by 61
  18. 2022Factorizing Knowledge in Neural Networkscited by 59
  19. 2023OTOV2: Automatic, Generic, User-Friendlycited by 32
  20. 2021Joint-DetNAS: Upgrade Your Detector with NAS, Pruning and Dynamic Distillationcited by 23
  21. 2022PPCD-GAN: Progressive Pruning and Class-Aware Distillation for Large-Scale Conditional GANs Compressioncited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 18; 2024: 66; 2025: 68; 2026: 43.

195 papers.

Sort:

By area:

  1. 2023DeepCache: Accelerating Diffusion Models for Freecited by 359
  2. 2024Deep Compression Autoencoder for Efficient High-Resolution Diffusion Modelscited by 217
  3. 2023A Survey on Deep Neural Network Pruning-Taxonomy, Comparison, Analysis, and Recommendationscited by 202
  4. 2024Delta-DiT: A Training-Free Acceleration Method Tailored for Diffusion Transformerscited by 175
  5. 2024Accelerating Diffusion Transformers with Token-wise Feature Cachingcited by 156
  6. 2024Learning-to-Cache: Accelerating Diffusion Transformer via Layer Cachingcited by 142
  7. 2024FORA: Fast-Forward Caching in Diffusion Transformer Accelerationcited by 137
  8. 2025From Reusing to Forecasting: Accelerating Diffusion Models with TaylorSeerscited by 136
  9. 2024Is Sora a World Simulator? A Comprehensive Survey on General World Models and Beyondcited by 111
  10. 2023Diffusion Models for Image Restoration and Enhancement: A Comprehensive Surveycited by 109
  11. 2023Diffusion Model as Representation Learnercited by 95
  12. 2024Q-DiT: Accurate Post-Training Quantization for Diffusion Transformerscited by 94
38 more
  1. 2024Not All Noises Are Created Equally:Diffusion Noise Selection and Optimizationcited by 71
  2. 2023Priority-Centric Human Motion Generation in Discrete Latent Spacecited by 67
  3. 2024MindBridge: A Cross-Subject Brain Decoding Frameworkcited by 66
  4. 2024Dynamic Diffusion Transformercited by 62
  5. 2023On the Design Fundamentals of Diffusion Models: A Surveycited by 60
  6. 2024ConceptPrune: Concept Editing in Diffusion Models via Skilled Neuron Pruningcited by 55
  7. 2024Adversarial Diffusion Compression for Real-World Image Super-Resolutioncited by 52
  8. 2024TinyFusion: Diffusion Transformers Learned Shallowcited by 45
  9. 2024DiP-GO: A Diffusion Pruner via Few-step Gradient Optimizationcited by 43
  10. 2024Motion Consistency Model: Accelerating Video Diffusion with Disentangled Motion-Appearance Distillationcited by 39
  11. 2024Attention-Driven Training-Free Efficiency Enhancement of Diffusion Modelscited by 38
  12. 2024Finding NeMo: Localizing Neurons Responsible For Memorization in Diffusion Modelscited by 37
  13. 2025Efficient Diffusion Models: A Surveycited by 35
  14. 2025FastVAR: Linear Visual Autoregressive Modeling via Cached Token Pruningcited by 34
  15. 2025Sparse-vDiT: Unleashing the Power of Sparse Attention to Accelerate Video Diffusion Transformerscited by 34
  16. 2024T-Stitch: Accelerating Sampling in Pre-Trained Diffusion Models with Trajectory Stitchingcited by 33
  17. 2024Collaborative Decoding Makes Visual Auto-Regressive Modeling Efficientcited by 33
  18. 2023SG-Former: Self-guided Transformer with Evolving Token Reallocationcited by 32
  19. 2025U-REPA: Aligning Diffusion U-Nets to ViTscited by 32
  20. 2024SlimFlow: Training Smaller One-Step Diffusion Models with Rectified Flowcited by 30
  21. 2024SnapGen-V: Generating a Five-Second Video within Five Seconds on a Mobile Devicecited by 27
  22. 2023Denoising Diffusion Step-aware Modelscited by 24
  23. 2024SDXS: Real-Time One-Step Latent Diffusion Models with Image Conditionscited by 24
  24. 2025BlockDance: Reuse Structurally Similar Spatio-Temporal Features to Accelerate Diffusion Transformerscited by 24
  25. 2024EDA-DM: Enhanced Distribution Alignment for Post-Training Quantization of Diffusion Modelscited by 23
  26. 2024AsyncDiff: Parallelizing Diffusion Models by Asynchronous Denoisingcited by 22
  27. 2024DKDM: Data-Free Knowledge Distillation for Diffusion Models with Any Architecturecited by 22
  28. 2024Importance-Based Token Merging for Efficient Image and Video Generationcited by 22
  29. 2024Layer- and Timestep-Adaptive Differentiable Token Compression Ratios for Efficient Diffusion Transformerscited by 22
  30. 2023FRDiff : Feature Reuse for Universal Training-free Acceleration of Diffusion Modelscited by 20
  31. 2023SlimSAM: 0.1% Data Makes Segment Anything Slimcited by 20
  32. 2025FreqCa: Accelerating Diffusion Models via Frequency-Aware Cachingcited by 20
  33. 2024Hash3D: Training-free Acceleration for 3D Generationcited by 19
  34. 2024Beta Sampling is All You Need: Efficient Image Generation Strategy for Diffusion Models using Stepwise Spectral Analysiscited by 19
  35. 2025OBS-Diff: Accurate Pruning For Diffusion Models in One-Shotcited by 19
  36. 2024SparseDM: Toward Sparse Efficient Diffusion Modelscited by 18
  37. 2024Binarized Diffusion Model for Image Super-Resolutioncited by 18
  38. 2025OmniCache: A Trajectory-Oriented Global Perspective on Training-Free Cache Reuse for Diffusion Transformer Modelscited by 18

50 most cited of 195 All 195 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2023Diff-Instruct: A Universal Approach for Transferring Knowledge From Pre-trained Diffusion Models
  2. 2024I-Diff: Structural Regularization for High-Fidelity Diffusion Models
  3. 2024LD-Pruner: Efficient Pruning of Latent Diffusion Models using Task-Agnostic Insights
  4. 2026Forward-Learned Discrete Diffusion: Learning how to noise to denoise faster
  5. 2025Accelerating Diffusion Model Training under Minimal Budgets: A Condensation-Based Perspective
  6. 2025Selective Underfitting in Diffusion Models
  7. 2025Learning Diffusion Models with Flexible Representation Guidance
  8. 2023DiffFit: Unlocking Transferability of Large Diffusion Models via Simple Parameter-Efficient Fine-Tuning