咨询与建议

限定检索结果

文献类型

  • 2 篇 会议

馆藏范围

  • 2 篇 电子文献
  • 0 种 纸本馆藏

日期分布

学科分类号

  • 2 篇 工学
    • 2 篇 计算机科学与技术...

主题

  • 2 篇 distributed data...
  • 1 篇 recommendation m...
  • 1 篇 knowledge graph ...
  • 1 篇 graph neural net...
  • 1 篇 prefetching
  • 1 篇 caching
  • 1 篇 efficient traini...

机构

  • 1 篇 univ southern ca...
  • 1 篇 instacart ca usa
  • 1 篇 ibm res almaden ...

作者

  • 1 篇 reinwald berthol...
  • 1 篇 lei chuan
  • 1 篇 annavaram murali
  • 1 篇 sheikh nasrullah
  • 1 篇 balasubramanian ...
  • 1 篇 qin xiao
  • 1 篇 choe joshua
  • 1 篇 alshabanah abdul...

语言

  • 2 篇 英文
检索条件"主题词=Distributed Data Parallel Training"
2 条 记 录,以下是1-10 订阅
排序:
Scaling Knowledge Graph Embedding Models for Link Prediction  2
Scaling Knowledge Graph Embedding Models for Link Prediction
收藏 引用
2nd European Workshop on Machine Learning and Systems (EuroMLSys)
作者: Sheikh, Nasrullah Qin, Xiao Reinwald, Berthold Lei, Chuan IBM Res Almaden San Jose CA 95120 USA Instacart San Francisco CA USA
Developing scalable solutions for training Graph Neural Networks (GNNs) for link prediction tasks is challenging due to the inherent data dependencies which entail high computational costs and a huge memory footprint.... 详细信息
来源: 评论
cDLRM: Look Ahead Caching for Scalable training of Recommendation Models  21
cDLRM: Look Ahead Caching for Scalable Training of Recommend...
收藏 引用
15th ACM Conference on Recommender Systems (RECSYS)
作者: Balasubramanian, Keshav Alshabanah, Abdulla Choe, Joshua Annavaram, Murali Univ Southern Calif Los Angeles CA 90007 USA
Deep learning recommendation models (DLRMs) are typically composed of two sets of parameters: large embedding tables to handle sparse categorical inputs, and neural networks such as multi-layer perceptrons (MLPs) to h... 详细信息
来源: 评论