M3-JEPA: Multimodal Alignment via Multi-gate MoE based on the Joint-Embedding Predictive Architecture

Hongyang Lei, Xiaolong Cheng, Qi Qin, Dan Wang, Huazhen Huang +3 more
2/14/2026

No official code found. Best implementation: HongyangLL/M3-JEPA (28 stars, Apache-2.0, last commit 2026-03-13). 1 more below.

Semantic Scholar

Code Implementations(2)

HongyangLL/M3-JEPACommunity72%
Apache-2.0

[ICML 2025] Repository for M3-JEPA: Multimodal Alignment via Multi-gate MoE based on the Joint-Predictive Embedding Architecture

283Shell, PythonAug 7, 20247 months agoApache-2.0
JinyuCong/M3-JEPACommunity70%

M3-JEPA: Multimodal Alignment via Multi-gate MoE based on the Joint-Embedding Predictive Architecture by Hongyang Lei and al, complition of the code make it run in local.

10PythonDec 16, 20256 months ago

Cite this paper

@article{lei2026mjepa,
  title  = {M3-JEPA: Multimodal Alignment via Multi-gate MoE based on the Joint-Embedding Predictive Architecture},
  author = {Hongyang Lei and Xiaolong Cheng and Qi Qin and Dan Wang and Huazhen Huang and Qingqing Gu and Yetao Wu and Luo Ji},
  year   = {2026},
  url    = {https://api.semanticscholar.org/CorpusID:bd03c5cf724eebb4cfbb9d67dcc78afcd018773c},
  journal = {ICML 2025 2025}
}

Discussion