AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Budget-Constrained Step-Level Diffusion Caching

Mingkun Lei, Tong Zhao, Liangyu Yuan, Chi Zhang

ICML 2026 · Image generation › Fast diffusion sampling · approach: Training-free diffusion acceleration

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
search a static budgeted cache policy instead of threshold-triggered caching, same latencyBudCache vs TeaCache · DrawBench, FLUX.1-dev, 2.51xLPIPS: 0.3218 -> 0.1759Improved
use BudCache-ultra (3.00x) versus slower TeaCache (2.51x)BudCache-ultra · DrawBench, FLUX.1-devPSNR: 18.401 -> 22.163Improved
transfer searched cache mask to other solvers (iPNDM, DPM-Solver++) (ablation)BudCache-base · DrawBench, FLUX.1-devPSNR: 15.35 -> 23.80 (iPNDM vs TeaCache)Improved
transfer searched cache mask to 512x512, 768x1024 and CFG 5/7 (ablation)BudCache-base · DrawBench, FLUX.1-devLPIPS: 0.3154 -> 0.1167 (512x512)Improved
transfer cache policy to GenEval prompts (ablation)BudCache-base · GenEval, FLUX.1-devLPIPS: 0.3349 -> 0.1388Improved
re-search cache policy on a distilled few-step modelBudCache-base on Hyper-SD (8 steps) · DrawBenchLPIPS: 0.1982 -> 0.1417Improved
transfer mask searched on Wan2.1 1.3B to 14B (ablation)BudCache · Wan2.1-T2V-14B, 30 promptsPSNR: 26.70 -> 27.94Improved
apply cache-aware schedule alignment on top of the cache policy at 5 NFE (ablation)BudCache · DrawBench, 5 NFEImageReward: 0.7808 -> 0.7888Improved
apply schedule alignment alone on the original uniform solver (ablation)Original solver · DrawBench, 5/6/7 NFEImageReward: 0.5896 -> 0.5991 (5 NFE); 0.8519 -> 0.8099 (7 NFE)Mixed
use optimized cache policy instead of simply reducing step count (ablation)BudCache · DrawBench, 5 NFEImageReward: 0.5896 -> 0.7808Improved
increase SA search intensity (R1 SA20 to R2 SA100) (ablation)BudCache search, 9 NFE · DrawBenchLPIPS: 0.2020 -> 0.2036No clear effect
increase nominal solver steps K under fixed 10 NFE (ablation)BudCache · DrawBench, 10 NFEquality: best at K=30Mixed

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