ReMoE: Boosting Expert Reuse through Router Fine-Tuning in Memory-Constrained MoE LLM Inference 文章

ArXiv CS.AI2026-05-27NEWSen作者: Xiongwei Zhu, Xiaojian Liao, Tianyang Jiang, Yusen Zhang, Liang Wang, Limin Xiao

查看原文 →

ReMoE: Boosting Expert Reuse through Router Fine-Tuning in Memory-Constrained MoE LLM Inference · 相关技术

相关技术

ORM TTE ODE divide-and-conquer partitioning Token Stan SPA Referring expression comprehension (REC)REUSE Parts-of-Speech (POS) tags MoE Mixture-of-Experts LMM ICT For ENA ANN Table iOS