Diffusion distillation is central to accelerating image and video generation, yet existing methods are fundamentally limited by the denoising process, where step reduction has largely saturated. Partial timestep low-resolution generation can further accelerate inference, but it suffers noticeable quality degradation due to cross-resolution distribution gaps. We propose Cross-Resolution Distribution Matching Distillation (RMD), a novel distillation framework that bridges cross-resolution distribution gaps for high-fidelity, few-step multi-resolution cascaded inference. Specifically, RMD divides the timestep intervals for each resolution using logarithmic signal-to-noise ratio (logSNR) curves, and introduces logSNR-based mapping to compensate for resolution-induced shifts. Distribution matching is conducted along resolution trajectories to reduce the gap between low-resolution generator distributions and the teacher's high-resolution distribution. In addition, a predicted-noise re-injection mechanism is incorporated during upsampling to stabilize training and improve synthesis quality. Quantitative and qualitative results show that RMD preserves high-fidelity generation while accelerating inference across various backbones. Notably, RMD achieves up to 33.4X speedup on SDXL and 25.6X on Wan2.1-14B, while preserving high visual fidelity.
Adversarial Distribution Matching for Diffusion Distillation Towards Efficient Image and Video Synthesis
Sun Yat-Sen University（中山大学）；ByteDance Seed Vision（字节跳动种子视觉）；Guangdong Provincial Key Laboratory of Information Security Technology（广东省信息安全技术重点实验室）；Key Laboratory of Machine Intelligence and Advanced Computing, Ministry of Education, China（教育部机器智能与高级计算重点实验室）；Pazhou Lab (HuangPu), Guangzhou, China（琶洲实验室（黄埔），广州，中国）
Joint Alignment and Distillation for Video Generation via Sample-Guided Distribution Matching
Tsinghua University（清华大学）；Kuaishou Technology（快手科技）；Harbin Institute of Technology (ShenZhen)（哈尔滨工业大学（深圳））；Beijing University of Posts and Telecommunications（北京邮电大学）
Teacher-Feature Drifting: One-Step Diffusion Distillation with Pretrained Diffusion Representations
JD Explore Academy（京东探索研究院）；Peking University（北京大学）；Tsinghua University（清华大学）；The Hong Kong University of Science and Technology（香港科技大学）；Beijing Institute of Technology（北京理工大学）
Although diffusion model has shown great potential for generating higher quality images than GANs, slow sampling speed hinders its wide application in practice. Progressive distillation is thus proposed for fast sampling by progressively aligning output images of $N$-step teacher sampler with $N/2$-step student sampler. In this paper, we argue that this distillation-based accelerating method can be further improved, especially for few-step samplers, with our proposed \textbf{C}lassifier-based \textbf{F}eature \textbf{D}istillation (CFD). Instead of aligning output images, we distill teacher's …
D$^4$M: Dataset Distillation via Disentangled Diffusion Model
School of Computer Science and Technology, UCAS（中国科学院大学计算机科学与技术学院）；Research Center on Fictitious Economy and Data Science, CAS（中国科学院虚拟经济与数据科学研究中心）；Sino-Danish College, UCAS（中国科学院大学中丹学院）；Department of Computer Science, NCSU（北卡罗来纳州立大学计算机科学系）；School of Economics and Management, UCAS（中国科学院大学经济与管理学院）；Institute of Computing Technology, CAS（中国科学院计算技术研究所）
SenseFlow: Scaling Distribution Matching for Flow-based Text-to-Image Distillation
The Hong Kong University of Science and Technology（香港科学与技术大学）；SenseTime Research（商汤科技研究院）；Vivix AI（维维克斯人工智能）；Institute for AI Industry Research, Tsinghua University（清华大学人工智能产业研究所）
Efficient Diffusion Distillation via Embedding Loss
School of Big Data and Artificial Intelligence（大数据与人工智能学院）；Guangdong University of Finance and Economics（广东金融学院）；School of Computer Science（计算机科学学院）；Shandong University（山东大学）
Generative Dataset Distillation using Min-Max Diffusion Model
School of Electrical and Electronic Engineering（电气与电子工程学院）；School of Mechanical and Aerospace Engineering（机械与航空航天工程学院）；Nanyang Technological University（南洋理工大学）
Korea Advanced Institute of Science and Technology (KAIST)（韩国科学技术院）；Bio and Brain Engineering（生物与脑工程系）；Kim Jaechul Graduate School of AI（金在哲人工智能研究生院）；Brain and Cognitive Sciences（脑与认知科学系）
Denoising Diffusion models have demonstrated their proficiency for generative sampling. However, generating good samples often requires many iterations. Consequently, techniques such as binary time-distillation (BTD) have been proposed to reduce the number of network calls for a fixed architecture. In this paper, we introduce TRAnsitive Closure Time-distillation (TRACT), a new method that extends BTD. For single step diffusion,TRACT improves FID by up to 2.4x on the same architecture, and achieves new single-step Denoising Diffusion Implicit Models (DDIM) state-of-the-art FID (7.4 for ImageNet…
Skolkovo Institute of Science and Technology（斯克尔克沃科学与技术学院）；Yandex Research（Yandex研究）；HSE University（俄罗斯高等经济大学）；Artificial Intelligence Research Institute（人工智能研究 institute）；Moscow Institute of Physics and Technology（莫斯科物理技术学院）
Continuous-Time Distribution Matching for Few-Step Diffusion Distillation
VCIP, College of Computer Science, Nankai University（南开大学计算机科学学院VCIP）；Alibaba Group（阿里巴巴集团）；College of Artificial Intelligence, Jilin University（吉林大学人工智能学院）
