Momentum-driven adaptive synchronization model for distributed DNN training on HPC clusters

Journal of Parallel and Distributed Computing（2022）

引用 1|浏览1

暂无评分

摘要

•Big Data application for Deep Learning.•DNN training on High-Performance Computing Platform.•DNN training slows down due to the stale gradients.•Stochastic Error Constraint for DNN training.•Adaptive Learning Rate for DNN training.

查看译文

关键词

Distributed deep learning system,Momentum-driven,Adaptive,Synchronization model,High-performance computing

AI 理解论文

溯源树

样例

生成溯源树，研究论文发展脉络

Chat Paper

正在生成论文摘要