Ask a Question

Prefer a chat interface with context about you and your work?

WDMoE: Wireless Distributed Mixture of Experts for Large Language Models

WDMoE: Wireless Distributed Mixture of Experts for Large Language Models

Large Language Models (LLMs) have achieved significant success in various natural language processing tasks, but the role of wireless networks in supporting LLMs has not been thoroughly explored. In this paper, we propose a wireless distributed Mixture of Experts (WDMoE) architecture to enable collaborative deployment of LLMs across edge servers …