The paper introduces an adaptive step schedule controller for text‑to‑image diffusion models, allowing the number of denoising steps to vary based on the complexity of the input prompt. By mixing step schedules of different sizes and monitoring error discrepancies at each timestep, the method switches schedules to maintain image quality while reducing inference time. Experiments on COCO and DiffusionDB demonstrate that this approach achieves faster generation without sacrificing visual fidelity.
By Kuluhan Binici, Cihan Acar, Shivam Aggarwal, Siying Liu, Tulika Mitra
The paper introduces Optimizing Your Sampling (OYS), a method that treats diffusion model timestep selection as a black-box optimization problem and uses Bayesian optimization to directly improve the target metric. OYS outperforms default schedules and the Align Your Steps approach on text-to-image generation, inpainting, and other image tasks, as shown by both quantitative and human evaluations. It works without additional training, applies to distilled models, and enhances both simple and sophisticated samplers, achieving 89–94% of a 50‑step schedule’s quality with only 5 steps and a tenfold reduction in inference cost.
By Travis Zhang, Christian Belardi, Justin Lovelace, Jin Peng Zhou, Saebyeol Shin, Carla P. Gomes, Kilian Q. Weinberger
The paper introduces adaptive second‑order solvers for diffusion models, employing a proportional‑integral (PI) step‑size controller based on a diffusion noise‑normalised error estimator. Unlike existing adaptive methods that react only to the current error, the PI solver incorporates past error information, producing smoother step adaptations. The authors also demonstrate that per‑sample trajectories share structure, allowing aggregation into a fixed schedule that largely preserves adaptive benefits. Experiments on natural‑image and language datasets show that the fixed schedule outperforms the commonly used EDM schedule for image sampling with the stochastic Heun sampler, and that the PI adaptive solver achieves lower FID than most baselines, though it does not surpass the EDM‑churn sampler at low neural‑network‑evaluation counts. In language diffusion, the PI solver improves perplexity over EDM and entropy schedules at low‑to‑medium NFE, albeit with reduced token entropy. The study concludes that per‑sample adaptivity is highly beneficial for simple 1D toy problems but only marginally useful for image and language data, where an average schedule can sometimes outperform the PI‑adaptive approach.
By Ella Kemperman, Luca Ambrogioni
arXiv:2606. 31026v1 Announce Type: cross Abstract: We propose OTCache, a training-free framework for accelerating diffusion sampling via caching schedule prediction.
By Huanlin Gao, Fang Zhao, Qiang Hui, Fuyuan Shi, Shaoan Zhao, Yantao Li, Chao Tan, Ting Lu, Yuren You, Kai Wang, Shiguo Lian
arXiv:2607. 00773v1 Announce Type: new Abstract: Discrete diffusion models are widely used for learning and generating discrete distributions.
By Yu Yao, Huanjian Zhou, Andi Han, Wei Huang, Masashi Sugiyama
arXiv:2602.12624v2 Announce Type: replace
Abstract: Diffusion-based generative models have achieved remarkable performance across various domains, yet their practical deployment is often limited by h...
By Sangwoo Jo, Sungjoon Choi
Diffusion models rely on numerical solvers requiring time-discretization, which has a large influence on the tradeoff between sampling cost and quality. However, the computational difficulty of the re...
arXiv:2605.16789v2 Announce Type: replace
Abstract: Diffusion and rectified flow (RF) models generate high-fidelity images and videos, but their iterative velocity-field evaluations are computational...
By Xiao Liu, Kai Liu, Naiyang Guan, Hongliang Lu, Zhixin Wang, Zhikai Chen, Renjing Pei, Yulun Zhang
GeoSPRINT is a training‑free framework that constructs non‑uniform sampling schedules for diffusion model inference by detecting geometrically redundant steps in denoising trajectories. It uses a hyperplanarity test in latent space, implemented via QR factorization, to allocate more steps to high‑curvature regions, and introduces the trajectory projection score α_traj as a model‑free diagnostic for flow quality. Across CIFAR‑10, LSUN Church, and Stable Diffusion v1.5, GeoSPRINT consistently outperforms uniform DDIM schedules at matched NFE budgets, improving FID scores by up to 1.93 points.
By Arpita Joshi
ChebBooster is a training‑free extrapolation framework that accelerates Diffusion Transformers (DiTs) by using Chebyshev polynomial theory. It employs a Barycentric formulation for numerically stable evaluation and separates the process into an offline weight precomputation phase and a lightweight online application stage. Experiments on DiT‑XL/2, PixArt‑Σ, and FLUX.1‑dev show consistent visual quality gains and up to 3.68× latency speedup and 5.12× FLOPs reduction compared to existing training‑free baselines.
By Chengjie Lu, Tianchi Deng, Zhengqi He, Chengwen Luo, Xueliang Li
arXiv:2506. 13058v2 Announce Type: replace-cross Abstract: Diffusion probabilistic models (DPMs) have demonstrated remarkable success in visual generation.
By Hu Yu, Hao Luo, Xueyang Fu, Jie Huang, Fan Wang, Feng Zhao
arXiv:2502. 10389v2 Announce Type: replace-cross Abstract: Diffusion models (DMs) have become the leading choice for generative tasks across diverse domains.
By Ziming Liu, Yifan Yang, Chengruidong Zhang, Yiqi Zhang, Lili Qiu, Yang You, Yuqing Yang