Skip to content

Conversation

@zaoliu-aws
Copy link
Contributor

import horovod.torch does not raise exception even if it is not fully successful. Add init() function to catch it.

# This redundant import is necessary because horovod does not raise an ImportError if the library is not present
import torch # noqa
#make sure the library is correctly imported
hvd.init()
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we know if this could interfer with other distribution strategies? I'm worried running init on horovod might cause a problem if the user is running SMDDP or DDP.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants