AI Paper Atlas

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

DRM: Diffusion-based Reward Model With Step-wise Guidance

Jaxon Zhang, Binxin Yang, Hubery Yin, Chen Li, Jing LYU

CVPR 2026 · Image generation › Text-to-image diffusion · approach: Preference optimization for diffusion · 1 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
initialize DRM from pre-trained diffusion weights instead of random (ablation)DRM · HPDv3 test preference accuracyaccuracy (%): 59.3 -> 74.0Improved
initialize from pre-trained weights rather than 3 epochs random training (ablation)DRM · HPDv3 test preference accuracyaccuracy (%): 63.0 -> 71.9Improved
raise training resolution from 256 to 512 (ablation)DRM · HPDv3 test preference accuracyaccuracy (%): 71.9 -> 74.0Improved
raise training resolution from 256 to 512 (ablation)DRM · PickScore test preference accuracyaccuracy (%): 72.1 -> 73.4Improved
evaluate DRM on noisier latents (timestep 750 vs 0) (ablation)DRM · HPSv3 testaccuracy (%): 74.0 -> 65.11Hurt
use Step-GRPO instead of standard GRPO with DRM (ablation)DRM · RL fine-tuning, HPSv3HPSv3: 10.07 -> 10.28Improved
use Step-GRPO instead of standard GRPO with DRM (ablation)DRM · RL fine-tuning, ImageRewardImageReward: 1.14 -> 1.17Improved
use DRM instead of HPSv3 as reward under GRPO (ablation)GRPO · RL fine-tuning, HPSv3HPSv3: 9.71 -> 10.07Improved
use DRM instead of HPSv3 as reward under GRPO (ablation)GRPO · RL fine-tuning, ImageRewardImageReward: 1.15 -> 1.14No clear effect
add Step-wise Sampling with k=6 candidates (ablation)SD3.5-Medium · Step-wise Sampling, HPSv3HPSv3: 8.95 -> 9.49Improved
add Step-wise Sampling with k=6 candidates (time cost) (ablation)SD3.5-Medium · Step-wise Samplingtime (s): 2.88 -> 9.83Hurt
increase Step-wise Sampling candidates from 2 to 6 (ablation)SD3.5-Medium · Step-wise Sampling, HPSv3HPSv3: 9.02 -> 9.49Improved

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