AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Soft-Di[M]O: Improved one-step Image Discrete Model

Yuanzhi Zhu, Xi WANG, Stéphane Lathuilière, Vicky Kalogeiton

ICLR 2026 · Image generation › Fast diffusion sampling · approach: Distillation for few-step generation

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 soft embeddings plus GAN refinement to Di[M]ODi[M]O with MaskBit teacher · ImageNet-256FID: 2.89 -> 1.96Improved
train Soft-Di[M]O longer with slightly changed hyperparametersSoft-Di[M]O MaskBit · ImageNet-256FID: 1.96 -> 1.56Improved
add soft embeddings and GAN to Di[M]ODi[M]O with MaskBit teacher · ImageNet-256recall: 0.49 -> 0.55Improved
add soft embeddings and GAN to Di[M]ODi[M]O with MaskBit teacher · ImageNet-256IS: 310.1 -> 281.4Hurt
add soft embeddings (GAN) to Di[M]O with MaskGit teacherDi[M]O MaskGit · ImageNet-256FID: 6.91 -> 6.40Improved
add differentiable reward fine-tuning with Di[M]O loss as regulariserDi[M]O MaskGen-L · Text-to-image GenEval overallGenEval: 0.42 -> 0.51Improved
add differentiable reward fine-tuningDi[M]O Meissonic · Text-to-image HPS V2.1HPS averaged: 28.59 -> 32.35Improved
add test-time embedding optimization (TTEO)Soft-Di[M]O-MaskGen-L · Text-to-image GenEval overallGenEval: 0.51 -> 0.63Improved
use soft embedding instead of Gumbel straight-through estimator (ablation)Soft-Di[M]O · ImageNet-256 (5k FID)FIDImproved
use GAN loss only without Di[M]O loss (ablation)Soft-Di[M]O · ImageNet-256 and MaskGen-LFID / diversityHurt
add Di[M]O loss as regulariser to the GAN loss (ablation)Soft-Di[M]O · ImageNet-256FIDImproved
add random masking of soft embeddings for the discriminator, larger range (ablation)Soft-Di[M]O GAN · ImageNet-256FIDImproved

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