AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Sortblock: Similarity-Aware Feature Reuse for Diffusion Model

Hanqi Chen, Xu Zhang, Xiaoliu Guan, Lielin Jiang, Guanzhong Wang, Zeyu Chen, Yi Liu

AAAI 2026 · Image generation › Fast diffusion sampling · approach: Training-free diffusion acceleration · 2 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
apply SortBlock caching (K=5)Flux.1-dev · COCO text-to-imageFID: 70.59 -> 70.47; latency 23.23 -> 11.60 sNo clear effect
increase refresh interval K from 5 to 9 (ablation)SortBlock on Flux.1-dev · COCO text-to-imagePSNR: 31.49 -> 27.27; speedup 2.00x -> 2.39xMixed
use SortBlock instead of TeaCacheFlux.1-dev · COCO text-to-imagePSNR: 22.11 -> 31.49Improved
use SortBlock instead of TaylorSeerHunyuanVideo · VBench text-to-videoVBench: 79.14 -> 78.29; PSNR 20.98 -> 28.29Mixed
use SortBlock instead of TeaCacheWan2.1 · VBench text-to-videoPSNR: 28.09 -> 29.79Improved
apply SortBlock only in early denoising stage (0-30%) (ablation)SortBlock · COCO text-to-imagequalitative structureHurt
lower recomputation scale beta from 1 to 0.1 (ablation)SortBlock · COCO text-to-imagequalitative detailHurt

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