Joint Dynamic Data and Model Parallelism for Distributed Training of DNNs over Heterogeneous Infrastructure
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS（2025）
Key words
Training,Data models,Pipelines,Optimization,Synchronization,Parallel processing,Scalability,Partitioning algorithms,Costs,Computational modeling,Distributed deep learning system,data parallel,model parallel,data assignment,model partitioning
AI Read Science
Must-Reading Tree
Example
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined