BatchFlow: Benefit-Aware Data Pipeline Allocation and Batch Reuse for Multi-Job Training
A data pipeline service that improves aggregate training throughput by dynamically allocating pipeline workers across jobs by using online profiling to identify jobs bottlenecked by the data pipeline and to estimate the throughput gain they would achieve from additional workers.
Patrick Watters, Xiao-Long Ma, Syed Zawad et al.
· 0 citations