AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Dense2MoE: Restructuring Diffusion Transformer to MoE for Efficient Text-to-Image Generation

Youwei Zheng, Yuxi Ren, Xin Xia, Xuefeng Xiao, Xiaohua Xie

ICCV 2025 · Image generation › Fast diffusion sampling · approach: Training-free diffusion acceleration · 4 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 MoE conversion of FFNs instead of Diff-Pruning at activated expand ratio 1.5FLUX.1 [dev] · GenEvalaccuracy: 0.4113 -> 0.5728Improved
Use MoE at ratio 1.5 instead of Diff-Pruning at ratio 2.0FLUX.1 [dev] · GenEvalaccuracy: 0.4888 -> 0.5728Improved
Use MoB grouping instead of Lite depth pruning (9 double, 38 single blocks)FLUX.1 [dev] · GenEvalaccuracy: 0.4919 -> 0.5396Improved
Use MoB grouping instead of Lite depth pruning (9 double, 26 single blocks)FLUX.1 [dev] · GenEvalaccuracy: 0.0926 -> 0.4956Improved
Use MoB grouping instead of BK-SDM (9 double, 26 single blocks)FLUX.1 [dev] · GenEvalaccuracy: 0.3450 -> 0.4956Improved
Use larger shared expert (1S12E2A) instead of smaller shared expert (0.5S14E4A), full MoE (ablation)FLUX.1 [dev] single-stream blocks · GenEvalaccuracy: 0.6270 -> 0.6331Improved
Use larger shared expert (1S12E2A) instead of smaller (0.5S14E4A), shared-only stage (ablation)FLUX.1 [dev] single-stream blocks · GenEvalaccuracy: 0.6001 -> 0.6187Improved
Add MoE distillation stage on top of shared-only initialization (1S12E2A) (ablation)FLUX.1 [dev] single-stream blocks · GenEvalaccuracy: 0.6187 -> 0.6331Improved
Add Taylor metric-based expert initialization (ablation)MoE initialization · GenEvalaccuracy: -0.0013Improved
Add Taylor metric-based expert initialization (ablation)MoE initialization · ImageRewardIR: -0.159Improved
Add dense-to-MoE distillation stage (ablation)MoE · GenEvalaccuracy: -0.0057Improved
Add dense-to-MoE distillation stage (ablation)MoE · ImageRewardIR: -0.0204Improved

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