-
Notifications
You must be signed in to change notification settings - Fork 178
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Update Version Syntax for PyTorch and PyTorch Lightning Examples #205
Conversation
hello @c-bata @toshihikoyanase. my PR is ready for review. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you for your PR. I have a question about the reason why we need to drop the support for PyTorch lighitning v1.6 and Pytorch v1.11.
pytorch/requirements.txt
Outdated
@@ -1,9 +1,9 @@ | |||
mpi4py | |||
plotly | |||
pytorch-ignite | |||
pytorch-lightning>=1.6.0 | |||
pytorch-lightning>=2.0 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In my understanding, the pytorch lightning examples works with the latest version without changes, so I guess we can keep the support for v1.6 as well. Do we have difficulties to keep the the current lower bound?
pytorch-lightning>=2.0 | |
pytorch-lightning>=1.6 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
As you said, there are no difficulties in keeping >=1.6 since that syntax logic will install whichever is the latest version.
Regarding explicitly setting a minimum at 2.0 in the requirements file, I would say this is more of a request as we move PL and Lightning Fabric into 2.0.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks for your comment.
Taking your comments into account, I've reconsidered our approach. Up until now, we've only changed version constraints when code modifications required it.
On the other hand, the current CI only tests against the latest version as you mentioned, so we can't be certain that the lower version bounds are being accurately checked.
Therefore, rather than setting a lower limit, it might be more realistic to simply remove the version constraints and ensure functionality solely with the latest version.
What do you think of it?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That sounds good to me!
Co-authored-by: Toshihiko Yanase <[email protected]>
This pull request has not seen any recent activity. |
Sorry for the delayed response. Let me check the update today. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you for your update. Looks good to me.
Motivation
Fixes #204
Description of the changes
After installing PyTorch 2.0 and PyTorch Lightning 2.0, I ran
pytorch_simple
andpytorch_lightning_simple
to confirm there were no issues with either example. Each ran without issue.Additionally, I bumped the minimum Python version to 3.8 in
pytorch.yml
.