mlsanityData to 7 Oct 2026

Phased Consistency Models

arXiv 2405.18407 ↗First posted 28 May 2024cs.LG cs.CVcited by 83

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Consistency Models (CMs) have made significant progress in accelerating the generation of diffusion models. However, their application to high-resolution, text-conditioned image generation in the latent space remains unsatisfactory. In this paper, we identify three key flaws in the current design of Latent Consistency Models (LCMs). We investigate the reasons behind these limitations and propose Phased Consistency Models (PCMs), which generalize the design space and address the identified limitations. Our evaluations demonstrate that PCMs outperform LCMs across 1--16 step generation settings. While PCMs are specifically designed for multi-step refinement, they achieve comparable 1-step generation results to previously state-of-the-art specifically designed 1-step methods. Furthermore, we show the methodology of PCMs is versatile and applicable to video generation, enabling us to train the state-of-the-art few-step text-to-video generator. Our code is available at https://github.com/G-U-N/Phased-Consistency-Model.

Key facts. Cited by 83 papers in the map (data through 7 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 28 of its 75 references that are in the map. First posted on arXiv 28 May 2024.

Builds on

28 of 75 references.

Closest ancestors · 20

Same direction, newest first

  1. 2024Distilling Diffusion Models into Conditional GANs9 May
  2. 2024Lumina-T2X: Transforming Text into Any Modality, Resolution, and Duration via Flow-based Large Diffusion Transformers9 May
  3. 2024Hyper-SD: Trajectory Segmented Consistency Model for Efficient Image Synthesis21 Apr
  4. 2024Self-Rectifying Diffusion Sampling with Perturbed-Attention Guidance26 Mar
  5. 2024Multistep Consistency Models11 Mar
  6. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesis5 Mar
  7. 2024Trajectory Consistency Distillation: Improved Latent Consistency Distillation by Semi-Linear Consistency Function with Trajectory Mapping29 Feb
  8. 2024SDXL-Lightning: Progressive Adversarial Diffusion Distillation21 Feb
  9. 2023Adversarial Diffusion Distillation28 Nov
  10. 2023Improved Techniques for Training Consistency Models22 Oct
  11. 2023Latent Consistency Models: Synthesizing High-Resolution Images with Few-Step Inference6 Oct
  12. 2023Consistency Trajectory Models: Learning Probability Flow ODE Trajectory of Diffusion1 Oct
8 more
  1. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis4 Jul
  2. 2023TRACT: Denoising Diffusion Models with Transitive Closure Time-Distillation7 Mar
  3. 2023Consistency Models2 Mar
  4. 2023Stable Target Field for Reduced Variance Score Estimation in Diffusion Models1 Feb
  5. 2022Classifier-Free Diffusion Guidance26 Jul
  6. 2022Progressive Distillation for Fast Sampling of Diffusion Models1 Feb
  7. 2020Score-Based Generative Modeling through Stochastic Differential Equations26 Nov
  8. 2020Denoising Diffusion Implicit Models6 Oct

Other references · 8

  1. 2022Flow Matching for Generative Modelingcited by 5,426
  2. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,550
  3. 2022Make-A-Video: Text-to-Video Generation without Text-Video Datacited by 1,725
  4. 2023Human Preference Score v2: A Solid Benchmark for Evaluating Human Preferences of Text-to-Image Synthesiscited by 847
  5. 2023Flow Matching on General Geometriescited by 260
  6. 2023Gen-L-Video: Multi-Text to Long Video Generation via Temporal Co-Denoisingcited by 135
  7. 2024CoMat: Aligning Text-to-Image Diffusion Model with Image-to-Text Concept Matchingcited by 60
  8. 2024AnimateLCM: Computation-Efficient Personalized Style Video Generation without Personalized Video Datacited by 42

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 7; 2025: 33; 2026: 43.

83 papers.

Sort:

By area:

  1. 2024Navigation World Modelscited by 262
  2. 2025Self-Forcing++: Towards Minute-Scale High-Quality Video Generationcited by 176
  3. 2025Diffusion Adversarial Post-Training for One-Step Video Generationcited by 142
  4. 2025Hunyuan-GameCraft: High-dynamic Interactive Game Video Generation with Hybrid History Conditioncited by 92
  5. 2025Align Your Flow: Scaling Continuous-Time Flow Map Distillationcited by 81
  6. 2025Large Scale Diffusion Distillation via Score-Regularized Continuous-Time Consistencycited by 81
  7. 2025LongCat-Video Technical Reportcited by 73
  8. 2025DiffuEraser: A Diffusion Model for Video Inpaintingcited by 69
  9. 2025Kling-Omni Technical Reportcited by 60
  10. 2026Context Forcing: Consistent Autoregressive Video Generation with Long Contextcited by 47
  11. 2025MiniMax-Remover: Taming Bad Noise Helps Video Object Removalcited by 42
  12. 2025AlphaFlow: Understanding and Improving MeanFlow Modelscited by 40
38 more
  1. 2025Decoupled DMD: CFG Augmentation as the Spear, Distribution Matching as the Shieldcited by 38
  2. 2025Transition Models: Rethinking the Generative Learning Objectivecited by 32
  3. 2025SenseFlow: Scaling Distribution Matching for Flow-based Text-to-Image Distillationcited by 29
  4. 2025Unleashing Vecset Diffusion Model for Fast Shape Generationcited by 25
  5. 2025VORTA: Efficient Video Diffusion via Routing Sparse Attentioncited by 23
  6. 2026AnyFlow: Any-Step Video Diffusion Model with On-Policy Flow Map Distillationcited by 23
  7. 2025Phased DMD: Few-step Distribution Matching Distillation via Score Matching within Subintervalscited by 21
  8. 2026Diversity-Preserved Distribution Matching Distillation for Fast Visual Synthesiscited by 20
  9. 2025Flash-DMD: Towards High-Fidelity Few-Step Image Generation with Efficient Distillation and Joint Reinforcement Learningcited by 17
  10. 2024NitroFusion: High-Fidelity Single-Step Diffusion through Dynamic Adversarial Trainingcited by 16
  11. 2024AsymRnR: Video Diffusion Transformers Acceleration with Asymmetric Reduction and Restorationcited by 15
  12. 2025FreqPolicy: Efficient Flow-based Visuomotor Policy via Frequency Consistencycited by 13
  13. 2026Salt: Self-Consistent Distribution Matching with Cache-Aware Training for Fast Video Generationcited by 11
  14. 2024Target-Driven Distillation: Consistency Distillation with Target Timestep Selection and Decoupled Guidancecited by 10
  15. 2025Adversarial Flow Modelscited by 10
  16. 2026Sparse Video Generation Propels Real-World Beyond-the-View Vision-Language Navigationcited by 10
  17. 2024Stable Consistency Tuning: Understanding and Improving Consistency Modelscited by 9
  18. 2026Language-Conditioned World Modeling for Visual Navigationcited by 9
  19. 2025SwiftVideo: A Unified Framework for Few-Step Video Generation through Trajectory-Distribution Alignmentcited by 8
  20. 2025Stream-DiffVSR: Low-Latency Streamable Video Super-Resolution via Auto-Regressive Diffusioncited by 7
  21. 2026MWM: Mobile World Models for Action-Conditioned Consistent Predictioncited by 7
  22. 2026Adaptive Video Distillation: Mitigating Oversaturation and Temporal Collapse in Few-Step Generationcited by 7
  23. 2026Fast Image Super-Resolution via Consistency Rectified Flowcited by 7
  24. 2026Data-Forcing Distillation: Restoring Diversity and Fidelity in Few-Step Video Generationcited by 7
  25. 2026Reinforcing Few-step Generators via Reward-Tilted Distribution Matchingcited by 6
  26. 2025ProReflow: Progressive Reflow with Decomposed Velocitycited by 5
  27. 2026Denoising as Path Planning: Training-Free Acceleration of Diffusion Models with DPCachecited by 5
  28. 2025VDOT: Efficient Unified Video Creation via Optimal Transport Distillationcited by 4
  29. 2026DUO-VSR: Dual-Stream Distillation for One-Step Video Super-Resolutioncited by 4
  30. 2026Continuous Adversarial Flow Modelscited by 4
  31. 2026Continuous-Time Distribution Matching for Few-Step Diffusion Distillationcited by 4
  32. 2026Vidu S1: A Real-Time Interactive Video Generation Modelcited by 4
  33. 2025Sig-DEG for Distillation: Making Diffusion Models Faster and Lightercited by 3
  34. 2025Shortcutting Pre-trained Flow Matching Diffusion Models is Almost Free Lunchcited by 3
  35. 2025Understanding, Accelerating, and Improving MeanFlow Trainingcited by 3
  36. 2025Self-Evaluation Unlocks Any-Step Text-to-Image Generationcited by 3
  37. 2026Formalizing the Sampling Design Space of Diffusion-Based Generative Models via Adaptive Solvers and Wasserstein-Bounded Timestepscited by 3
  38. 2026CollectionLoRA: Collecting 50 Effects in 1 LoRA via Multi-Teacher On-Policy Distillationcited by 3

50 most cited of 83 All 83 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2023VideoLCM: Video Latent Consistency Model
  2. 2024Simplifying, Stabilizing and Scaling Continuous-Time Consistency Models
  3. 2024TLCM: Training-efficient Latent Consistency Model for Image Generation with 2-8 Steps
  4. 2025MotionPCM: Real-Time Motion Synthesis with Phased Consistency Model
  5. 2023LCM-LoRA: A Universal Stable-Diffusion Acceleration Module
  6. 2024Provable Statistical Rates for Consistency Diffusion Models
  7. 2025Improved Training Technique for Latent Consistency Models
  8. 2024Reward Guided Latent Consistency Distillation