mlsanityData to 14 Oct 2026

One Step Diffusion via Shortcut Models

arXiv 2410.12557 ↗First posted 16 Oct 2024cs.LG cs.CVcited by 391

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Diffusion models and flow-matching models have enabled generating diverse and realistic images by learning to transfer noise to data. However, sampling from these models involves iterative denoising over many neural network passes, making generation slow and expensive. Previous approaches for speeding up sampling require complex training regimes, such as multiple training phases, multiple networks, or fragile scheduling. We introduce shortcut models, a family of generative models that use a single network and training phase to produce high-quality samples in a single or multiple sampling steps. Shortcut models condition the network not only on the current noise level but also on the desired step size, allowing the model to skip ahead in the generation process. Across a wide range of sampling step budgets, shortcut models consistently produce higher quality samples than previous approaches, such as consistency models and reflow. Compared to distillation, shortcut models reduce complexity to a single network and training phase and additionally allow varying step budgets at inference time.

Key facts. Cited by 391 papers in the map (data through 14 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 37 of its 42 references that are in the map. First posted on arXiv 16 Oct 2024.

Builds on

37 of 42 references.

Closest ancestors · 21

Same direction, newest first

  1. 2024Consistency Models Made Easy20 Jun
  2. 2024EM Distillation for One-step Diffusion Models27 May
  3. 2024Improved Distribution Matching Distillation for Fast Image Synthesis23 May
  4. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesis5 Mar
  5. 2023One-step Diffusion with Distribution Matching Distillation30 Nov
  6. 2023Adversarial Diffusion Distillation28 Nov
  7. 2023Improved Techniques for Training Consistency Models22 Oct
  8. 2023Latent Consistency Models: Synthesizing High-Resolution Images with Few-Step Inference6 Oct
  9. 2023BOOT: Data-free Distillation of Denoising Diffusion Models with Bootstrapping8 Jun
  10. 2023A Comprehensive Survey on Knowledge Distillation of Diffusion Models9 Apr
  11. 2023TRACT: Denoising Diffusion Models with Transitive Closure Time-Distillation7 Mar
  12. 2023Consistency Models2 Mar
9 more
  1. 2023Understanding Diffusion Objectives as the ELBO with Simple Data Augmentation1 Mar
  2. 2022Scalable Diffusion Models with Transformers19 Dec
  3. 2022On Distillation of Guided Diffusion Models6 Oct
  4. 2022Classifier-Free Diffusion Guidance26 Jul
  5. 2021High-Resolution Image Synthesis with Latent Diffusion Models20 Dec
  6. 2021Diffusion Models Beat GANs on Image Synthesis11 May
  7. 2021Knowledge Distillation in Iterative Generative Models for Improved Sampling Speed7 Jan
  8. 2020Denoising Diffusion Implicit Models6 Oct
  9. 2020Denoising Diffusion Probabilistic Models19 Jun

Other references · 16

  1. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
  2. 2022Flow Matching for Generative Modelingcited by 5,702
  3. 2022Generative Adversarial Networkscited by 5,627
  4. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,700
  5. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusioncited by 2,838
  6. 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,626
  7. 2022Video Diffusion Modelscited by 2,305
  8. 2020DiffWave: A Versatile Diffusion Model for Audio Synthesiscited by 1,354
  9. 2021What Matters in Learning from Offline Human Demonstrations for Robot Manipulationcited by 725
  10. 2022StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasetscited by 492
  11. 2024Lumiere: A Space-Time Diffusion Model for Video Generationcited by 455
  12. 2021Implicit Behavioral Cloningcited by 452
4 more
  1. 2022Behavior Transformers: Cloning k modes with one stonecited by 304
  2. 2023The Dormant Neuron Phenomenon in Deep Reinforcement Learningcited by 174
  3. 2020Implicit Under-Parameterization Inhibits Data-Efficient Deep Reinforcement Learningcited by 134
  4. 2024Mixtures of Experts Unlock Parameter Scaling for Deep RLcited by 78

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 3; 2025: 158; 2026: 230.

391 papers.

Sort:

By area:

  1. 2025Mean Flows for One-step Generative Modelingcited by 561
  2. 2025MAGI-1: Autoregressive Video Generation at Scalecited by 276
  3. 2025SAM 3D: 3Dfy Anything in Imagescited by 247
  4. 2025Real-Time Execution of Action Chunking Flow Policiescited by 216
  5. 2025In-Context Edit: Enabling Instructional Image Editing with In-Context Generation in Large Scale Diffusion Transformercited by 204
  6. 2025Flow Q-Learningcited by 184
  7. 2025WorldPlay: Towards Long-Term Geometric Consistency for Real-Time Interactive World Modelingcited by 170
  8. 2025Training Agents Inside of Scalable World Modelscited by 136
  9. 2025The Diffusion Dualitycited by 135
  10. 2026Generative Modeling via Driftingcited by 116
  11. 2025Inductive Moment Matchingcited by 113
  12. 2025Improved Mean Flows: On the Challenges of Fastforward Generative Modelscited by 110
38 more
  1. 2025How to build a consistency model: Learning flow maps via self-distillationcited by 103
  2. 2025ReinFlow: Fine-tuning Flow Matching Policy with Online Reinforcement Learningcited by 97
  3. 2025Align Your Flow: Scaling Continuous-Time Flow Map Distillationcited by 87
  4. 2025Diffuse and Disperse: Image Generation with Representation Regularizationcited by 70
  5. 2025VideoVLA: Video Generators Can Be Generalizable Robot Manipulatorscited by 64
  6. 2026LoopFormer: Elastic-Depth Looped Transformers for Latent Reasoning via Shortcut Modulationcited by 61
  7. 2025Learning Few-Step Diffusion Models by Trajectory Distribution Matchingcited by 57
  8. 2026One-step Latent-free Image Generation with Pixel Mean Flowscited by 53
  9. 2025Distribution Matching Distillation Meets Reinforcement Learningcited by 50
  10. 2025ManiFlow: A General Robot Manipulation Policy via Consistency Flow Trainingcited by 49
  11. 2025Decoupled DMD: CFG Augmentation as the Spear, Distribution Matching as the Shieldcited by 45
  12. 2025GestureLSM: Latent Shortcut based Co-Speech Gesture Generation with Spatial-Temporal Modelingcited by 44
  13. 2025Diffusion Models for Robotic Manipulation: A Surveycited by 41
  14. 2025AlphaFlow: Understanding and Improving MeanFlow Modelscited by 41
  15. 2026Meta Flow Maps enable scalable reward alignmentcited by 37
  16. 2025Much Ado About Noising: Dispelling the Myths of Generative Robotic Controlcited by 36
  17. 2025Transition Models: Rethinking the Generative Learning Objectivecited by 34
  18. 2025Diffusion Models without Classifier-free Guidancecited by 33
  19. 2025GraspGen: A Diffusion-based Framework for 6-DOF Grasping with On-Generator Trainingcited by 32
  20. 2025MoFlow: One-Step Flow Matching for Human Trajectory Forecasting via Implicit Maximum Likelihood Estimation based Distillationcited by 31
  21. 2026Transition Matching Distillation for Fast Video Generationcited by 30
  22. 2025pi-Flow: Policy-Based Few-Step Generation via Imitation Distillationcited by 29
  23. 2026Diamond Maps: Efficient Reward Alignment via Stochastic Flow Mapscited by 29
  24. 2025Scaling Offline RL via Efficient and Expressive Shortcut Modelscited by 27
  25. 2025SplitMeanFlow: Interval Splitting Consistency in Few-Step Generative Modelingcited by 26
  26. 2025Temporal Difference Flowscited by 24
  27. 2025There is No VAE: End-to-End Pixel-Space Generative Modeling via Self-Supervised Pre-trainingcited by 24
  28. 2025Phased DMD: Few-step Distribution Matching Distillation via Score Matching within Subintervalscited by 24
  29. 2026One-Step Generative Modeling via Wasserstein Gradient Flowscited by 24
  30. 2025Flow Matching-Based Autonomous Driving Planning with Advanced Interactive Behavior Modelingcited by 22
  31. 2025One-Step Flow Policy Mirror Descentcited by 21
  32. 2026FASTER: Rethinking Real-Time Flow VLAscited by 21
  33. 2025Joint Velocity-Growth Flow Matching for Single-Cell Dynamics Modelingcited by 20
  34. 2025Decoupled MeanFlow: Turning Flow Models into Flow Maps for Accelerated Samplingcited by 20
  35. 2025Terminal Velocity Matchingcited by 20
  36. 2026Flash-WAM: Modality-Aware Distillation for World Action Modelscited by 19
  37. 2025CMT: Mid-Training for Efficient Learning of Consistency, Mean Flow, and Flow Map Modelscited by 18
  38. 2025One-Step Generative Policies with Q-Learning: A Reformulation of MeanFlowcited by 18

50 most cited of 391 All 391 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025You Only Look One Step: Accelerating Backpropagation in Diffusion Sampling with Gradient Shortcuts
  2. 2023Parallel Sampling of Diffusion Models
  3. 2022Speed up the inference of diffusion models via shortcut MCMC sampling
  4. 2022Progressive Distillation for Fast Sampling of Diffusion Models
  5. 2026Setting-Matched and Semantics-Scaled Benchmarking of One-Step Generative Models Against Multistep Diffusion and Flow Models
  6. 2024Multistep Distillation of Diffusion Models via Moment Matching
  7. 2025Pretrained Diffusion Models Are Inherently Skipped-Step Samplers
  8. 2025Fewer Denoising Steps or Cheaper Per-Step Inference: Towards Compute-Optimal Diffusion Model Deployment