Diffusion Timestep Scheduling

Follow this topic Ask about this topic

Diffusion timestep scheduling studies when noise is added during training and where a sampler evaluates the denoising process at inference. Approaches include hand-designed cosine or log-SNR schedules, learned monotone splines, and optimization using numerical-error estimates, optimal control, reinforcement learning, or Bayesian search. Adaptive methods also allocate steps by input difficulty or task, and schedule sparse guidance or cached computation; they are evaluated across image, video, language, scientific, and discrete-generation settings. A central tension is whether schedules should be fixed and transferable or adapted to each model and sample, since benefits depend on the solver, denoiser, step budget, and quality measure.

Latest papers

64 papers10 papers in the last four weeks, up 67% on the four weeks before. 0.1% of all new papers.

Formalizing the Sampling Design Space of Diffusion-Based Generative Models via Adaptive Solvers and Wasserstein-Bounded Timesteps
Feb 13, 2026Sangwoo Jo, Sungjoon ChoiDiffusion Model Sampling Diffusion Sampling
Department of Artificial Intelligence, Korea University, Seoul, South Korea.
Improved Sampling Schedules for Discrete Diffusion Models
Feb 6, 2026Alberto Foresti, Mustapha Bounoua, Giulio Franzese +2Diffusion Sampling Diffusion Models
Department of Data Science, EURECOM, 450 Route des Chappes, 06410 Biot, France · Donders Institute for Brain, Cognition and Behaviour, Radboud University, Postbus 9102, 6500 HC Nijmegen, Netherlands
ART for Diffusion Sampling: A Reinforcement Learning Approach to Timestep Schedule
Jan 26, 2026Yilie Huang, Wenpin Tang, Xunyu ZhouDiffusion Sampling Adaptive Sampling
Department of Industrial Engineering and Operations Research, Columbia University, New York, NY 10027, USA. · Department of Industrial Engineering and Operations Research & Data Science Institute, Columbia University, New York, NY 10027, USA.
Disentangling Total-Variance and Signal-to-Noise-Ratio Improves Diffusion Models
Feb 12, 2025Khaled Kahouli, Winfried Ripken, Stefan Gugler +3Diffusion Model Sampling Diffusion Timestep Scheduling
BIFOLD – Berlin Institute for the Foundations of Learning and Data · Machine Learning Group, Technische Universität Berlin · Google DeepMind +3

Feb 13, 2026J/K move · Enter open · S save

Formalizing the Sampling Design Space of Diffusion-Based Generative Models via Adaptive Solvers and Wasserstein-Bounded Timesteps

Sangwoo Jo, Sungjoon Choi

Department of Artificial Intelligence, Korea University, Seoul, South Korea.

Diffusion Model Sampling Diffusion Sampling Adaptive Sampling Diffusion Timestep Scheduling Diffusion Model Acceleration

Open paper AskPDF arXiv

Diffusion-based generative models have achieved remarkable performance across various domains, yet their practical deployment is often limited by high sampling costs. While prior work focuses on training objectives or individual solvers, the broader sampling design problem, specifically solver selection and scheduling, remains largely governed by static heuristics. We propose SDM, a principled, training-free sampling framework that adapts both the numerical solver and the timestep schedule to the intrinsic properties of the diffusion trajectory. By analyzing the PF-ODE dynamics, we show that velocity variation is small in high-noise stages and increases near the data manifold, identifying intervals where solver order is most consequential. In parallel, we introduce an offline-calibrated adaptive scheduling method that explicitly controls the local Wasserstein discretization error and projects the calibrated trajectory to a prescribed NFE budget. We further extend the formulation to a mixed-transition Wasserstein error bound, providing a unified error-propagation view of adaptive scheduling and solver selection within the overall SDM framework. Across standard benchmarks, with extensions to modern ODE samplers, high-resolution synthesis, and text-to-image generation, SDM achieves improved sample quality compared to baseline methods, attaining an FID of 1.93 on CIFAR-10, 2.41 on FFHQ, and 1.98 on AFHQv2, with a reduced number of function evaluations compared to existing samplers. Our code is available at https://github.com/aiimaginglab/sdm.

Latest papers

Formalizing the Sampling Design Space of Diffusion-Based Generative Models via Adaptive Solvers and Wasserstein-Bounded Timesteps

Improved Sampling Schedules for Discrete Diffusion Models

ART for Diffusion Sampling: A Reinforcement Learning Approach to Timestep Schedule

Disentangling Total-Variance and Signal-to-Noise-Ratio Improves Diffusion Models