There is a newer version of the record available.

Published October 6, 2023 | Version v1
Software Open

Artifact for DynaPipe: Optimizing Multi-task Training through Dynamic Pipelines

Authors/Creators

  • 1. The University of Hong Kong
  • 2. Amazon Web Services, Inc.

Description

This artifact describes the steps to reproduce Fig. 12 through 17 in the paper DynaPipe: Optimizing Multi-task Training through Dynamic Pipelines. A docker image containing all software dependencies is included. The experiments are expected to be run on a single AWS EC2 p4d instance. The evaluation is expected to take around 26 hours to complete.

Files

DynaPipe_Artifact_Appendix.pdf

Files (20.0 GB)

Name Size
md5:ade79b9feafc056e9cec2de0c30ed2fc
321.1 kB Preview Download
md5:4e9c99f75cb5de3f88304d0f696666d0
19.8 GB Download
md5:23b6f69b3fb9708aacddea710a2779e5
238.0 MB Download