Scholarlink

In this chapter, we present CoEdge (Cooperative Edge), a runtime system that orchestrates cooperative DNN inference over multiple heterogeneous edge devices. CoEdge does not apply any structural modifications or tuning requirements to the given DNN model, and does not sacrifice model accuracy as it reserves input data and model parameters of the given DNN model. CoEdge employs parallel workflow in which the input is split initially and the execution is parallelized on multiple devices at runtime.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Cooperative Edge Model Inference over Heterogeneous Edge Devices

  • 章节

  • 发表日期：

    2025年

查看全文
  • Sen Lin,
  • Zhi Zhou,
  • Zhaofeng Zhang,
  • Xu Chen,
  • Junshan Zhang

摘要

In this chapter, we present CoEdge (Cooperative Edge), a runtime system that orchestrates cooperative DNN inference over multiple heterogeneous edge devices. CoEdge does not apply any structural modifications or tuning requirements to the given DNN model, and does not sacrifice model accuracy as it reserves input data and model parameters of the given DNN model. CoEdge employs parallel workflow in which the input is split initially and the execution is parallelized on multiple devices at runtime.