AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Decoupled DMD: CFG Augmentation as the Spear, Distribution Matching as the Shield

Dongyang Liu, Peng Gao, David Liu, Ruoyi Du, Zhen Li, Qilong Wu, Xin Jin, Sihan Cao, Shifeng Zhang, Steven HOI, Hongsheng Li

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Fast diffusion ODE solvers · 43 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
decouple re-noising schedules but keep both global (ablation)DMD · DPG-Bench (Lumina-Image-2.0)DPG overall: 83.90 -> 83.77No clear effect
constrain CA re-noising to tau > t (ablation)DMD (Lumina-Image-2.0) · DPG-BenchDPG overall: 83.77 -> 85.64Improved
keep DM regularizer over full noise range while constraining CA (hybrid) (ablation)Decoupled-Constrained · HPS v2.1HPS v2.1 overall: 31.71 -> 32.29Improved
keep DM regularizer over full noise range while constraining CA (hybrid) (ablation)Decoupled-Constrained · HPS v3HPS v3 overall: 11.08 -> 11.59Improved
replace DMD2 re-noising schedule with decoupled-hybridDMD2 · 4-step SDXLFID: 18.95 -> 17.80Improved
replace DMD2 re-noising schedule with decoupled-hybridDMD2 · 4-step SDXLHPS v2.1: 30.64 -> 30.34Hurt
replace DMD2 re-noising schedule with decoupled-hybridDMD2 · 4-step SDXLImageReward: 71.01 -> 78.61Improved
train with CFG Augmentation only (no DM) (ablation)DMD · 4-step SDXLstability / image qualityMixed
replace DM with mean-variance KL regularizer (ablation)CA engine · 4-step SDXLstability / qualityMixed
replace DM with GAN regularizer (ablation)CA engine · 4-step SDXLtraining stabilityHurt
restrict CA re-noising to clean timesteps [0.7, 1.0] (ablation)CA-only single-step generator · image qualitytraining stabilityHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy