OmniMoE: An Efficient MoE by Orchestrating Atomic Experts at Scale

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Shi, Jingze, Peng, Zhangyang, Zhu, Yizhang, Wu, Yifan, Liu, Guang, Luo, Yuyu
Format:	Preprint
Published:	2026
Subjects:	Computation and Language Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2602.05711