Diff-MoE: Diffusion Transformer with Time-Aware and Space-Adaptive Experts

Kun Cheng, Xiao He, Lei Yu, Zhijun Tu, Mingrui Zhu +3 more
2/14/2026

No official code found. Best implementation: kunncheng/Diff-MoE (33 stars, last commit 2025-11-10).

Semantic Scholar

Code Implementations(1)

kunncheng/Diff-MoECommunity71%

[ICML 2025] Diff-MoE: Diffusion Transformer with Time-Aware and Space-Adaptive Experts

333PythonMay 10, 202511 months ago

Cite this paper

@article{cheng2026diffmoe,
  title  = {Diff-MoE: Diffusion Transformer with Time-Aware and Space-Adaptive Experts},
  author = {Kun Cheng and Xiao He and Lei Yu and Zhijun Tu and Mingrui Zhu and Nannan Wang and Xinbo Gao and Jie Hu},
  year   = {2026},
  url    = {https://api.semanticscholar.org/CorpusID:70552806f2906801081f1137af262bd8d32ffc8e},
  journal = {ICML 2025 2025}
}

Discussion