1
Dense2MoE: Pushing the Pareto Frontier of On-Device LLMs via Unified Pruning and Upcycling
提出Dense2MoE统一剪枝与升级方法，将稠密LLM转化为MoE架构，显著提升设备端大模型的效率与性能，推动帕累托前沿。
arXiv:2605.26496v1 Announce Type: cross Abstract: The Mixture of Experts MoE architecture is highly promising for resource constrained on device deplo…