Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.
And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.
Bad monitoring metrics with integration callbacks are difficult to notice.
See original GitHub issueMotivation
Bad/wrong monitoring metrics represented by key strings with the following integration callbacks are difficult to notice. For instance, a simple typo in the metric would make the pruner a NOOP pruner. One might expect an error.
- https://github.com/optuna/optuna/blob/master/optuna/integration/keras.py (resolved by #1759)
- https://github.com/optuna/optuna/blob/master/optuna/integration/tfkeras.py (resolved by #2156)
- https://github.com/optuna/optuna/blob/master/optuna/integration/pytorch_lightning.py (resolved by #2157, ongoing)
Description
Above callbacks are not doing dictionary lookups that raise on failure but uses dict.get. I’m wondering if we cannot raise instead as that’d clearly communicate to the user that the pruner isn’t doing anything. However, I’m not too familiar with all of the above libraries and it might not be possible.
Issue Analytics
- State:
- Created 3 years ago
- Reactions:2
- Comments:5 (3 by maintainers)
Top Results From Across the Web
Top Related Medium Post
Top Related StackOverflow Question
Troubleshoot Live Code
Top Related Reddit Thread
Top Related Hackernoon Post
Top Related Tweet
Top Related Dev.to Post
Top Related Hashnode Post
You’re right, let me close it. Thanks @nzw0301 for the heads up and again, thanks @bigbird555 !
@bigbird555 Kudos to your all contributions to this issue!
Do we have something to do related to this issue? Otherwise, I suppose that we can close this isseu.