AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Laplacian Multi-scale Flow Matching for Generative Modeling

Zelin Zhao, Petr Molodyk, Haotian Xue, Yongxin Chen

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Flow matching distillation · 5 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 MoT shared model instead of separate per-scale models (ablation)LapFlow · CelebA-HQ 256x256GFLOPs / FID: GFLOPs 38.9 -> 16.5; FID 3.60 -> 3.53Improved
add causal attention mask across scales (vs none) (ablation)LapFlow · CelebA-HQ 256x256FID: 3.91 -> 3.53Improved
use causal mask instead of self mask (ablation)LapFlow · CelebA-HQ 256x256FID: 5.19 -> 3.53Improved
set critical time T to 0.5 rather than 0.1 or 0.9 (ablation)LapFlow · CelebA-HQ 256x256FID: 5.12 (T=0.1), 4.92 (T=0.9) -> 3.53 (T=0.5)Improved
use EQVAE instead of SDVAE (ablation)LapFlow · CelebA-HQ 256x256FID: 4.37 -> 3.53Improved
use EQVAE instead of SDVAE (ablation)LFM baseline · CelebA-HQ 256x256FID: 5.26 -> 7.77Hurt
linear noise schedule instead of GVP (ablation)LapFlow · CelebA-HQ 256x256FID: 4.10 -> 3.53Improved
linear noise decay instead of cubic decay (ablation)LapFlow · CelebA-HQ 256x256FID: 4.25 -> 3.53Improved
two Laplacian scales instead of four (ablation)LapFlow · CelebA-HQ 256x256FID: 5.12 -> 3.53Improved
three scales instead of two at 256x256 (ablation)LapFlow · CelebA-HQ 256x256FID: 3.53 -> 3.59No clear effect
three scales instead of two at 512x512 (ablation)LapFlow · CelebA-HQ 512x512FID: 5.45 -> 4.04Improved
three scales instead of two at 1024x1024 (ablation)LapFlow · CelebA-HQ 1024x1024FID: 6.62 -> 5.51Improved

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