most citedRelax: Composable Abstractions for End-to-End Dynamic Machine Learning

12 citations · 12 across the 2 of their papers we have counts for

collaborators
Showing cs.DCShow all

1 paper · 1 filter

cs.DC2024

GraphPipe: Improving Performance and Scalability of DNN Training with Graph Pipeline Parallelism

Byungsoo Jeon, Mengdi Wu, Shiyi Cao +11

Deep neural networks (DNNs) continue to grow rapidly in size, making them infeasible to train on a single device. Pipeline parallelism is commonly used in existing DNN systems to s…