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鈥檒l occasionally send you account related emails.

Already on GitHub? Sign in to your account

鈿楋笍馃幆 Fix norm limit regularizers hpo defaults #1274

Merged
merged 3 commits into from
May 19, 2023

Conversation

mberr
Copy link
Member

@mberr mberr commented May 19, 2023

Fix #1273

@mberr mberr mentioned this pull request May 19, 2023
3 tasks
@mberr mberr changed the title Fix norm limit regularizers hpo defaults 馃鈴帮笍 Fix norm limit regularizers hpo defaults May 19, 2023
@mberr mberr marked this pull request as ready for review May 19, 2023 18:06
@mberr mberr changed the title 馃鈴帮笍 Fix norm limit regularizers hpo defaults 鈿楋笍馃幆 Fix norm limit regularizers hpo defaults May 19, 2023
@mberr mberr requested a review from cthoyt May 19, 2023 19:02
Copy link
Member

@cthoyt cthoyt left a comment

Choose a reason for hiding this comment

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

Maybe there鈥檚 a more automated way to check all of these components have HPO defaults? Do we have a class resolved for it?

@mberr mberr merged commit cb226fc into master May 19, 2023
14 checks passed
@mberr mberr deleted the fix-norm-limit-regularizers-hpo-defaults branch May 19, 2023 20:34
@mberr
Copy link
Member Author

mberr commented May 19, 2023

Maybe there鈥檚 a more automated way to check all of these components have HPO defaults? Do we have a class resolved for it?

I guess we have, at least for some of them 馃

These ones also look like we may need to test some things:

default_kwargs_ranges=lr_schedulers_hpo_defaults[self.lr_scheduler],

default_kwargs_ranges=optimizers_hpo_defaults[self.optimizer],

@mberr
Copy link
Member Author

mberr commented May 19, 2023

It should also suffice to only have one of kwargs_ranges and default_kwargs_ranges be non-empty

def _get_kwargs(
trial: Trial,
prefix: str,
*,
default_kwargs_ranges: Mapping[str, Any],
kwargs: Optional[Mapping[str, Any]] = None,
kwargs_ranges: Optional[Mapping[str, Any]] = None,
) -> Mapping[str, Any]:
_kwargs_ranges = dict(default_kwargs_ranges)
if kwargs_ranges is not None:
_kwargs_ranges.update(kwargs_ranges)
return suggest_kwargs(
trial=trial,
prefix=prefix,
kwargs_ranges=_kwargs_ranges,
kwargs=kwargs,
)

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.

HPO with TransH not working
2 participants