linkedin / detext

DeText: A Deep Neural Text Understanding Framework for Ranking and Classification Tasks

BSD 2-Clause "Simplified" License

1.26k stars 134 forks source link

Fix horovod support bugs #26

Closed StarWang closed 4 years ago

StarWang commented 4 years ago

Description

Running DeText with use_horovod=True is causing training crashes. This PR fixes this error.

Fixes # (issue)

Type of change

Please delete options that are not relevant.

[ ] Bug fix (non-breaking change which fixes an issue)

List all changes

Please list all changes in the commit.

Add horovod initialization and worker setup

Testing

run_detext.sh
run_detext_multitask.sh
k8s run_detext.py test

Checklist

[ ] My code follows the style guidelines of this project
[ ] I have performed a self-review of my own code
[ ] I have commented my code, particularly in hard-to-understand areas
[ ] I have made corresponding changes to the documentation
[ ] My changes generate no new warnings
[ ] I have added tests that prove my fix is effective or that my feature works
[ ] New and existing unit tests pass locally with my changes
[ ] Any dependent changes have been merged and published in downstream modules

xwli-chelsea commented 4 years ago

LGTM. Thanks!