AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

From Reusing to Forecasting: Accelerating Diffusion Models with TaylorSeers

Jiacheng Liu, Chang Zou, Yuanhuiyi Lyu, Junjie Chen, Linfeng Zhang

ICCV 2025 · Image generation › Fast diffusion sampling · approach: Training-free diffusion acceleration · 98 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 first-order Taylor forecasting (O=1) instead of direct reuse (O=0) (ablation)TaylorSeer N=7 · DiT-XL/2 ImageNetFID: 12.67 -> 4.29Improved
raise Taylor order from 1 to 2 (ablation)TaylorSeer N=7 · DiT-XL/2 ImageNetFID: 4.29 -> 3.66Improved
raise Taylor order from 2 to 3 (ablation)TaylorSeer N=7 · DiT-XL/2 ImageNetFID: 3.66 -> 3.63Improved
raise Taylor order from 3 to 4 (ablation)TaylorSeer N=3 · DiT-XL/2 ImageNetFID: 2.34 -> 2.35No clear effect
add Taylor forecasting (O=1) instead of direct reuse (ablation)TaylorSeer N=3 · DiT-XL/2 ImageNetFID: 3.55 -> 2.49Improved
add Taylor forecasting (O=1) instead of direct reuse (ablation)TaylorSeer N=5 · DiT-XL/2 ImageNetFID: 6.58 -> 3.05Improved
lengthen caching interval from N=3 to N=7 with O=3 (ablation)TaylorSeer · DiT-XL/2 ImageNetFID: 2.34 -> 3.63Hurt
replace direct reuse caching with TaylorSeer at about 4.5x speedupFORA · DiT-XL/2 ImageNetFID: 6.58 -> 2.65Improved
replace TeaCache with TaylorSeer at about 5x speedupTeaCache · HunyuanVideo VBenchVBench score: 79.36 -> 79.93Improved
use TaylorSeer at low speedup ratioFLUX.1-dev · DrawBenchImageReward: 0.9898 -> 1.0181Improved
use TaylorSeer at low speedup ratioHunyuanVideo · VBenchVBench score: 80.66 -> 80.74Improved
replace TeaCache with TaylorSeer at about 4x speedupTeaCache · HiDream ImageRewardImageReward: 0.9849 -> 1.0833Improved

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