REVIEW 37 cited by
Fast Sampling of Diffusion Models with Exponential Integrator
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
Fast Sampling of Diffusion Models with Exponential Integrator
abstract
The past few years have witnessed the great success of Diffusion models~(DMs) in generating high-fidelity samples in generative modeling tasks. A major limitation of the DM is its notoriously slow sampling procedure which normally requires hundreds to thousands of time discretization steps of the learned diffusion process to reach the desired accuracy. Our goal is to develop a fast sampling method for DMs with a much less number of steps while retaining high sample quality. To this end, we systematically analyze the sampling procedure in DMs and identify key factors that affect the sample quality, among which the method of discretization is most crucial. By carefully examining the learned diffusion process, we propose Diffusion Exponential Integrator Sampler~(DEIS). It is based on the Exponential Integrator designed for discretizing ordinary differential equations (ODEs) and leverages a semilinear structure of the learned diffusion process to reduce the discretization error. The proposed method can be applied to any DMs and can generate high-fidelity samples in as few as 10 steps. In our experiments, it takes about 3 minutes on one A6000 GPU to generate $50k$ images from CIFAR10. Moreover, by directly using pre-trained DMs, we achieve the state-of-art sampling performance when the number of score function evaluation~(NFE) is limited, e.g., 4.17 FID with 10 NFEs, 3.37 FID, and 9.74 IS with only 15 NFEs on CIFAR10. Code is available at https://github.com/qsh-zh/deis
Forward citations
Cited by 37 Pith papers
-
Query Lower Bounds for Diffusion Sampling
Diffusion sampling from d-dimensional distributions requires at least ~sqrt(d) adaptive score queries when score estimates have polynomial accuracy.
-
Consistency Models
Consistency models achieve fast one-step generation with SOTA FID of 3.55 on CIFAR-10 and 6.20 on ImageNet 64x64 by directly mapping noise to data, outperforming prior distillation techniques.
-
Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow
Rectified flow learns straight-path neural ODEs for distribution transport, yielding efficient generative models and domain transfers that work well even with a single simulation step.
-
Asymptotic Preservation and Uniform Accuracy of Diffusion and Flow-Matching Samplers
DDIM (σ-clock Euler) is the unique layer-exact fixed-step sampler; deterministic residual budgets stay O(1) with no log(1/σ_min), while stochastic path-KL scales as Λ²/N from the Itô term alone.
-
CoDMD: Copula-aware Distribution Matching Distillation for Fast Video Generation
CoDMD adds a copula-matching regularizer to DMD for distilling 50-step video diffusion models to 4 steps, reporting VBench scores of 84.46/84.87 on 1.3B/14B Wan-2.1-T2V models.
-
Dual-Constrained Diffusion Image Compression for Operational Rate-Distortion-Perception Optimization
DCIC uses dual constraints on a diffusion decoder to realize adjustable RDP operating points in neural image compression without extra rate cost.
-
Proximal-Based Generative Modeling for Bayesian Inverse Problems
PGM replaces the intractable likelihood score in diffusion models with a closed-form Moreau score computed via proximal operators, enabling non-asymptotic sampling for inverse problems trained only on prior data.
-
Proximal-Based Generative Modeling for Bayesian Inverse Problems
PGM framework links diffusion to proximal regularization for closed-form Moreau-score sampling in Bayesian inverse problems, learned only from prior samples.
-
Watch Your Step: Information Injection in Diffusion Models via Shadow Timestep Embedding
Timestep embeddings in diffusion models function as a separable side channel that can carry dedicated information for adversarial injection or detection.
-
NI Sampling: Accelerating Discrete Diffusion Sampling by Token Order Optimization
NI Sampling accelerates discrete diffusion language models up to 14.3 times by training a neural indicator to select which tokens to sample at each step using a trajectory-preserving objective.
-
Blade: A Derivative-free Bayesian Inversion Method using Diffusion Priors
Blade produces well-calibrated derivative-free Bayesian posterior samples by combining split Gibbs sampling, ensemble statistical linearization, and diffusion priors.
-
DiffusionNFT: Online Diffusion Reinforcement with Forward Process
DiffusionNFT performs online RL for diffusion models on the forward process via flow matching and positive-negative contrasts, delivering up to 25x efficiency gains and rapid benchmark improvements over prior reverse-...
-
Lipschitz-Guided Design of Interpolation Schedules in Generative Models
Minimizing averaged squared Lipschitzness of the drift produces interpolation schedules that improve numerical accuracy and mitigate mode collapse in generative models, with closed-form optima for Gaussians and valida...
-
Elucidating the Design Space of Diffusion-Based Generative Models
Organizing diffusion model design choices yields SOTA FID of 1.79 on CIFAR-10 with only 35 network evaluations per image and similar gains on ImageNet-64.
-
Amortized Moment Matching for Visual Generation
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.
-
FARI: Robust One-Step Inversion for Watermarking in Diffusion Models
One-step adversarially LoRA-tuned inversion exploits low-curvature reverse trajectories to beat 50-step DDIM on watermark robustness after ~20 minutes of fine-tuning.
-
Asymptotic Preservation and Uniform Accuracy of Diffusion and Flow-Matching Samplers
A diffusion sampler's terminal rule and grid, not its update formula alone, determine its zero-noise behavior; one fitted terminal map restores asymptotic preservation and new Hermite maps raise the provable endpoint ...
-
Function-Space Diffusion for Motion Planning
FSD-MP performs diffusion over continuous trajectory functions with Matérn noise and a DST-FNO reverse model, enabling zero-shot multi-resolution robot motion planning.
-
DanceOPD: On-Policy Generative Field Distillation
Hard-routed, single low-noise on-policy velocity matching composes conflicting image-generation capabilities into one flow student better than joint training, merging, or dense OPD baselines.
-
Colored Noise Diffusion Sampling
CNS is a plug-and-play stochastic sampler for diffusion models that uses timestep- and frequency-dependent colored noise to allocate energy to unresolved bands, producing lower FID scores than standard ODE/SDE baselin...
-
LIVEditor-14B: Lightning Unified Video Editing via In-Context Sparse Attention
LIVEditor-14B applies a new sparse attention method (ISA) that prunes context and uses query-sharpness routing to cut attention latency ~60% with no loss in editing quality on standard benchmarks.
-
Efficient Diffusion Distillation via Embedding Loss
Embedding Loss aligns feature distributions via MMD in random network embeddings to boost one-step diffusion distillation, reaching SOTA FID of 1.475 on CIFAR-10 unconditional generation.
-
MPDiT: Multi-Patch Global-to-Local Transformer Architecture For Efficient Flow Matching and Diffusion Model
MPDiT uses a hierarchical multi-patch design in transformers to lower computation in diffusion models by handling coarse global features first then fine local details, plus faster-converging embeddings.
-
RSTR: Reducing SpatioTemporal Redundancy in Diffusion Transformers
OUSAC uses evolutionary search to skip most classifier-free-guidance passes and adaptive low-rank caching to recover quality, cutting diffusion transformer compute by about 50–60% with equal or better FID.
-
Large Scale Diffusion Distillation via Score-Regularized Continuous-Time Consistency
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.
-
2ndMatch: Finetuning Pruned Diffusion Models via Second-Order Jacobian Matching
2ndMatch finetunes pruned diffusion models via second-order Jacobian matching inspired by Finite-Time Lyapunov Exponents to reduce the quality gap with dense models on image generation tasks.
-
Improved Techniques for Training Consistency Models
Improved consistency training techniques achieve FID scores of 2.51 on CIFAR-10 and 3.25 on ImageNet 64x64 in one sampling step, outperforming prior consistency training and distillation methods.
-
eDiff-I: Text-to-Image Diffusion Models with an Ensemble of Expert Denoisers
An ensemble of stage-specialized text-to-image diffusion models improves prompt alignment over single shared-parameter models while preserving visual quality and inference speed.
-
DPM-Solver++: Fast Solver for Guided Sampling of Diffusion Probabilistic Models
DPM-Solver++ enables high-quality guided sampling of diffusion models in 15-20 steps via data-prediction ODE solving and multistep stabilization.
-
DanceOPD: On-Policy Generative Field Distillation
DanceOPD routes samples across capability velocity fields in flow-matching models and trains via on-policy student-induced states to compose T2I, local editing, and global editing without mutual interference.
-
TEMPO-Diffusion: Temporally Exposed Malicious Poisoning of Diffusion Models
TEMPO-Diffusion is a targeted backdoor attack framework for diffusion models that uses time-conditioned triggers to poison class-specific synthetic data, achieving high attack success in downstream classifiers.
-
ARIA: Adaptive Region-Based Importance Allocation for Conditional Diffusion Distillation
ARIA adaptively focuses distillation updates on conditioning-space regions with high ongoing teacher-student misalignment while preserving the base objective.
-
Precise: SDE-Consistent Stochastic Sampling for RL Post-Training of Flow-Matching Models
Precise is a new SDE-consistent stochastic sampler that balances exploration and stability for RL post-training of flow-matching models via a novel posterior-mean approximation.
-
Divergence-Suppressing Couplings for Rectified Flow
Divergence-suppressing couplings attenuate the divergent part of the velocity field when generating training couplings for Rectified Flow, yielding straighter paths and better generation quality at no extra inference cost.
-
LIVEditor-14B: Lightning Unified Video Editing via In-Context Sparse Attention
ISA prunes low-saliency context tokens and routes queries by sharpness to either full or 0-th order Taylor sparse attention, enabling LIVEditor to cut attention latency ~60% while beating prior video editing methods o...
-
Accelerating Redshift-Conditioned Galaxy Image Synthesis with One-step Generative Modeling
One-step pixel-MeanFlow models recover key galaxy morphology statistics at orders-of-magnitude lower computational cost than standard DDPM sampling while remaining weaker on fine-grained structure.
-
FSampler: Training Free Acceleration of Diffusion Sampling via Epsilon Extrapolation
FSampler accelerates diffusion sampling by substituting finite-difference extrapolations of epsilon for model calls on selected steps, reducing NFE by 15-25% at SSIM 0.95-0.99.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.