AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

OBS-Diff: Accurate Pruning For Diffusion Models in One-Shot

Junhan Zhu, Hesong Wang, Mingluo Su, Zefang Wang, Huan Wang

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Training-free diffusion acceleration · 15 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
use OBS-Diff instead of Wanda at 60% unstructured sparsitySD 3-medium · MS-COCO 2014FID: 170.33 -> 28.49Improved
use OBS-Diff instead of Wanda at 70% unstructured sparsityFlux 1.dev · MS-COCO 2014FID: 49.68 -> 39.79Improved
use OBS-Diff instead of Wanda at 50% unstructured sparsitySD 3.5-large · MS-COCO 2014FID: 27.49 -> 29.61Hurt
use OBS-Diff instead of Wanda at 2:4 sparsitySD 3.5-large · MS-COCO 2014FID: 32.08 -> 32.13No clear effect
use OBS-Diff instead of Wanda at 2:4 sparsitySD 3.5-large · MS-COCO 2014ImageReward: -0.1363 -> 0.4493Improved
use OBS-Diff instead of L1-norm at 30% structured sparsitySD 3.5-large · MS-COCO 2014FID: 327.48 -> 34.51Improved
use OBS-Diff instead of EcoDiff at 30% structured sparsitySDXL · MS-COCO 2014FID: 101.96 -> 29.75Improved
use OBS-Diff instead of Diff-PruningDDPM CIFAR-10 · CIFAR-10FID: 7.72 -> 7.55Improved
add timestep-aware Hessian weighting with logarithmic decrease (ablation)OBS-Diff · MS-COCO 2014 (1,000 prompts)ImageRewardImproved
increase the number of module packages (ablation)OBS-Diff · MS-COCO 2014 (1,000 prompts)ImageRewardNo clear effect
increase calibration prompts up to 100 (ablation)OBS-Diff · MS-COCO 2014 (1,000 prompts)ImageRewardImproved
calibrate on out-of-distribution GCC3M instead of MS-COCO train (ablation)OBS-Diff SD 3-medium 50% unstructured · MS-COCO 2014FID: 27.93 -> 27.20No clear effect

Later papers that compared with this method

1 later accepted paper reports OBS-Diff as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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