Pith. sign in

REVIEW 19 cited by

Multistep Consistency Models

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2403.06807 v3 pith:L72AO57O submitted 2024-03-11 cs.LG cs.CVstat.ML

Multistep Consistency Models

classification cs.LG cs.CVstat.ML
keywords consistencymodelmodelsstepdiffusionsamplesgeneratemultistep
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
abstract

Diffusion models are relatively easy to train but require many steps to generate samples. Consistency models are far more difficult to train, but generate samples in a single step. In this paper we propose Multistep Consistency Models: A unification between Consistency Models (Song et al., 2023) and TRACT (Berthelot et al., 2023) that can interpolate between a consistency model and a diffusion model: a trade-off between sampling speed and sampling quality. Specifically, a 1-step consistency model is a conventional consistency model whereas a $\infty$-step consistency model is a diffusion model. Multistep Consistency Models work really well in practice. By increasing the sample budget from a single step to 2-8 steps, we can train models more easily that generate higher quality samples, while retaining much of the sampling speed benefits. Notable results are 1.4 FID on Imagenet 64 in 8 step and 2.1 FID on Imagenet128 in 8 steps with consistency distillation, using simple losses without adversarial training. We also show that our method scales to a text-to-image diffusion model, generating samples that are close to the quality of the original model.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 19 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance

    cs.LG 2026-04 unverdicted novelty 8.0

    FMRG reformulates guidance as deterministic optimal control, deriving a single-trajectory method using the flow map that matches or exceeds baselines on reward-guided generation and inverse problems with 3 NFEs at tex...

  2. Covariance-aware sampling for Diffusion Models

    stat.ML 2026-05 conditional novelty 7.0

    A covariance-aware extension of DDIM sampling for pixel-space diffusion models that uses Tweedie's formula and Fourier decomposition to model reverse-process covariance and improves sample quality at low NFE.

  3. How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance

    cs.LG 2026-04 unverdicted novelty 7.0

    FMRG is a training-free, single-trajectory guidance method for flow models derived from optimal control that achieves strong reward alignment with only 3 NFEs.

  4. Agentic World Modeling: Foundations, Capabilities, Laws, and Beyond

    cs.AI 2026-04 unverdicted novelty 7.0

    Proposes a levels x laws taxonomy for world models in AI agents, defining L1-L3 capabilities across physical, digital, social, and scientific regimes while reviewing over 400 works to outline a roadmap for advanced ag...

  5. Amortized Moment Matching for Visual Generation

    cs.LG 2026-07 accept novelty 6.0

    Amortized Fréchet Distance uses neural nets to match conditional means and covariances, yielding stronger one-step visual generators than explicit FD-loss or multi-step teachers.

  6. Diffusion Fine-tuning with Rewarded Moment Matching Distillation

    cs.LG 2026-06 unverdicted novelty 6.0

    RMMD simultaneously distills diffusion models and optimizes rewards, yielding better FID-reward trade-offs on ImageNet than DI++, DRaFT and HyperNoise, and a 7.5x faster GenCast model that beats its teacher on 93% of ...

  7. Consistent Diffusion Language Models

    cs.LG 2026-04 unverdicted novelty 6.0

    CDLM trains denoisers to be path-invariant across stochastic posterior bridges in discrete diffusion, unifying prior methods and achieving new SOTA few-step text generation performance.

  8. How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance

    cs.LG 2026-04 unverdicted novelty 6.0

    FMRG is a training-free single-trajectory guidance framework for flow-based models that matches or exceeds baselines on reward-guided tasks and inverse problems using as few as 3 NFEs.

  9. Extending One-Step Image Generation from Class Labels to Text via Discriminative Text Representation

    cs.CV 2026-04 unverdicted novelty 6.0

    By requiring and using highly discriminative LLM text features, the work enables the first effective one-step text-conditioned image generation with MeanFlow.

  10. Heterogeneous Decentralized Diffusion Models

    cs.LG 2026-03 conditional novelty 6.0

    A decentralized diffusion framework that mixes DDPM and Flow Matching experts, converts all predictions to velocity space at inference, and claims 16x less compute and 14x less data than prior DDM work.

  11. Consistency Deep Equilibrium Models

    cs.LG 2026-02 conditional novelty 6.0

    C-DEQ trains a consistency model to map intermediate solver states directly to the DEQ equilibrium, enabling accurate one-to-few-step inference for deep equilibrium models.

  12. Understanding, Accelerating, and Improving MeanFlow Training

    cs.CV 2025-11 conditional novelty 6.0

    Training MeanFlow by first forming instantaneous velocity and short-gap average velocity, then shifting to long gaps, improves 1-NFE ImageNet FID from 3.43 to 2.87 and speeds training by about 2.5x.

  13. Large Scale Diffusion Distillation via Score-Regularized Continuous-Time Consistency

    cs.CV 2025-10 conditional novelty 6.0

    The work introduces rCM, a score-regularized continuous-time consistency model that matches DMD2 quality on large models up to 14B parameters while improving diversity and enabling 1-4 step sampling.

  14. MACS: Measurement-Aware Consistency Sampling for Inverse Problems

    eess.IV 2025-10 conditional novelty 6.0

    MACS replaces the variance term in aDDIM with a measurement residual, yielding better FID/KID than fast ODE baselines in two-step consistency-model inverse problems.

  15. Trajectory-Consistent Flow Matching for Robust Visuomotor Policy Learning

    cs.RO 2026-05 unverdicted novelty 5.0

    Trajectory consistency training, smoothness regularization, and higher-order integration for flow matching policies deliver 60-70% success on long-horizon real-robot tasks where baselines achieve 0%.

  16. Consistent Diffusion Language Models

    cs.LG 2026-04 unverdicted novelty 5.0

    CDLM introduces MPDC training for discrete diffusion models, recovering prior methods as limits and claiming new SOTA text generation performance especially at low sampling budgets.

  17. Heterogeneous Decentralized Diffusion Models

    cs.LG 2026-03 conditional novelty 5.0

    Isolated experts trained with mixed DDPM and Flow Matching objectives can be fused at inference via deterministic conversion, cutting reported DDM compute ~16× while slightly improving FID and diversity over a homogen...

  18. A Continuous-Time Consistency Model for 3D Point Cloud Generation

    cs.CV 2025-09 reject novelty 5.0

    ConTiCoM-3D trains a continuous-time consistency-style model directly on raw 3D point clouds using flow matching plus Chamfer distance, with one- to two-step generation.

  19. Agentic World Modeling: Foundations, Capabilities, Laws, and Beyond

    cs.AI 2026-04 conditional novelty 4.0

    A survey proposing a three-level capability taxonomy (L1 Predictor, L2 Simulator, L3 Evolver) for world models across physical, digital, social, and scientific domains.