[RaphaelFakhri]

What does this PR do?

LearningRateMonitor(log_weight_decay=True) raises KeyError: 'weight_decay' when the optimizer has no weight_decay key in its param groups, for example torch.optim.Rprop, torch.optim.LBFGS or torch.optim.SparseAdam.

This change reads the value with param_group.get("weight_decay", 0), which matches how log_momentum=True already logs 0 for optimizers without momentum. A param group without weight_decay applies no weight decay, so 0 is the accurate value.

The new test test_log_weight_decay_no_weight_decay_optimizer trains a BoringModel with Rprop. It fails with KeyError on master and passes with this change.

Fixes #22035

No breaking changes.

Before submitting

• Was this discussed/agreed via a GitHub issue? (not for typos and docs)
• Did you read the contributor guideline, Pull Request section?
• Did you make sure your PR does only one thing, instead of bundling different changes together?
• Did you make sure to update the documentation with your changes? (if necessary)
• Did you write any new necessary tests? (not for typos and docs)
• Did you verify new and existing tests pass locally with your changes?
• Did you list all the breaking changes introduced by this pull request?
• Did you update the CHANGELOG? (not for typos, docs, test updates, or minor internal changes/refactors)

PR review

Anyone in the community is welcome to review the PR.
Before you start reviewing, make sure you have read the review guidelines. In short, see the following bullet-list:

Reviewer checklist

• Is this pull request ready for review? (if not, please submit in draft mode)
• Check that all items from Before submitting are resolved
• Make sure the title is self-explanatory and the description concisely explains the PR
• Add labels and milestones (and optionally projects) to the PR so it can be classified