AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

SoFlow: Solution Flow Models for One-Step Generative Modeling

Tianze Luo, Haotian Yuan, Zhuang Liu

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Flow matching distillation · 13 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 Flow Matching loss data (75% ratio) (ablation)SoFlow · ImageNet 256 DiT-B/4 1-NFEFID-50K: 66.36 -> 58.57Improved
use loss weighting p=1.0 instead of p=0 (ablation)SoFlow · ImageNet 256 DiT-B/4 1-NFEFID-50K: 68.25 -> 58.57Improved
use exponential instead of cosine l-to-t schedule (ablation)SoFlow · ImageNet 256 DiT-B/4 1-NFEFID-50K: 60.17 -> 58.57No clear effect
use constant instead of exponential schedule (ablation)SoFlow · ImageNet 256 DiT-B/4 1-NFEFID-50K: 58.57 vs 59.52No clear effect
use loss weighting p=1.5 instead of 1.0 (ablation)SoFlow · ImageNet 256 DiT-B/4 1-NFEFID-50K: 58.57 -> 63.72Hurt
use linear schedule with Euler parameterization instead of trigonometric/trigonometric (ablation)SoFlow · ImageNet 256 DiT-B/4 1-NFE with CFGFID-50K: 19.01 -> 11.59Improved
increase CFG strength from 1.5 to 3.0 (ablation)SoFlow · ImageNet 256 DiT-B/4 1-NFE with CFGFID-50K: 34.47 -> 11.59Improved
lower velocity mix ratio m from 1.0 to 0.25 (ablation)SoFlow · ImageNet 256 DiT-B/4 1-NFE with CFGFID-50K: 17.22 -> 11.59Improved
replace MeanFlow loss with solution consistency loss (efficient attention)SoFlow · ImageNet 256 training costGPU memory: 51.45 GB -> 35.44 GBImproved
replace MeanFlow loss with solution consistency loss (efficient attention)SoFlow · ImageNet 256 training costtraining speed: 2.39 -> 2.94 iters/secImproved
use SoFlow instead of MeanFlowXL/2 · ImageNet 256 1-NFEFID-50K: 3.43 -> 2.96Improved
use SoFlow instead of MeanFlowXL/2 · ImageNet 256 2-NFEFID-50K: 2.93 -> 2.66Improved

Later papers that compared with this method

1 later accepted paper reports SoFlow-XL 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