Skip to content
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

Jean/logging #77

Open
wants to merge 50 commits into
base: main
Choose a base branch
from
Open

Jean/logging #77

wants to merge 50 commits into from

Conversation

jeandut
Copy link
Contributor

@jeandut jeandut commented Jan 16, 2025

Use fed-pydeseq2 logging utils written by @umarteauowkin to log all training quantities that are exchanged in nice graphs and tables.

@jeandut jeandut marked this pull request as draft January 16, 2025 10:29
@jeandut jeandut marked this pull request as ready for review January 19, 2025 16:09
@jeandut
Copy link
Contributor Author

jeandut commented Jan 19, 2025

Hey @umarteauowkin,
Thanks again a billion times for providing those utils. Do you mind checking that everything is okay for you before merge ?
Notably regarding credit assignment.
Cheers.

Copy link
Collaborator

@honghaoli42 honghaoli42 left a comment

Choose a reason for hiding this comment

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

Thanks for the great work! Just a minor comment, is it possible to add a piece of test for the added components?

@jeandut
Copy link
Contributor Author

jeandut commented Jan 21, 2025

Thanks for the great work! Just a minor comment, is it possible to add a piece of test for the added components?

Unfortunately we won't be able to do explicit testings. Implicit testing works to the extents that I tested all FL algorithms which produce correct graphs and tables. As the goal of this util is really for power users I don't mind that it can break if some other FL algos are developed. Monitoring is orthogonal to algorithmic functionalities which themselves should be properly tested.

However I cannot merge it as it is right now because of the immense overhead it adds (CI goes up to like 4 hours). I am trying to find a way to cancel every monitoring by default so that most users don't have to face the consequences.

Copy link

@umarteauowkin umarteauowkin left a comment

Choose a reason for hiding this comment

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

From my perspective (credit assignent), LGTM ! Thanks @jeandut :)

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.

3 participants