AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

Flash-DMD: Towards High-Fidelity Few-Step Image Generation with Efficient Distillation and Joint Reinforcement Learning

Guanjie Chen, Shirui Huang, Yifu Sun, Kai Liu, Jianchen Zhu, Xiaoye Qu, Yu Cheng, Peng Chen

CVPR 2026 · Image generation › Text-to-image diffusion · approach: Inference-time T2I improvement · 13 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
add Pixel-GAN to DMD2 (ablation)DMD2 SDXL · COCO-10kImageReward: 0.8748 -> 0.8918Improved
use aggressive TTUR (lower score-estimator update ratio) in DMD2 (ablation)DMD2 SDXL · COCO-10kImageReward: 0.8748 -> 0.8871Improved
add timestep-aware loss decoupling on top of Pixel-GAN and aggressive TTUR (ablation)Flash-DMD SDXL · COCO-10kImageReward: 0.8748 -> 0.9509 (24k -> 1k steps)Improved
add joint reinforcement learning with distillation (ablation)Flash-DMD SDXL · COCO-10kImageReward: 0.9509 -> 1.0035Improved
increase training steps from 1k to 8k (ablation)Flash-DMD SDXL · COCO-10kImageReward: 0.9509 -> 0.9740Improved
use TTUR 5 instead of 1-2 in Flash-DMD (ablation)Flash-DMD SDXL · COCO-10kImageReward: 0.9426 (TTUR5-18k) vs 0.9509 (TTUR1-1k)No clear effect
use RL sampling over all noise timesteps instead of only high-noise ones (ablation)Flash-DMD phase 2 · COCO-10kImageReward: 1.0004 -> 0.9421Hurt
apply RL as separate post-training instead of jointly with distillation (ablation)Flash-DMD phase 2 · COCO-10kHPSv2: 0.2931 -> 0.2882Hurt
alternate RL and DM updates at 5:1 ratio rather than 1:1 (ablation)Flash-DMD phase 2 · COCO-10kImageReward: 0.9135 -> 0.9808Improved
raise RL:DM ratio further to 10:1 (ablation)Flash-DMD phase 2 · COCO-10kHPSv2: 0.2904 (5:1) -> 0.2874 (10:1)Hurt
add Pixel-GAN loss during RL stage (ablation)Flash-DMD phase 2 · COCO-10kImageReward: 0.9421 -> 0.9678Improved
use Flash-DMD instead of DMD2 at TTUR=2 (ablation)SDXL distillation · COCO-10ktraining stabilityImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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