automl / neps

Neural Pipeline Search (NePS): Helps deep learning experts find the best neural pipeline.
https://automl.github.io/neps/
Apache License 2.0
44 stars 11 forks source link

fix(metahyper): Print useful information in logs for exception occuring #53

Closed eddiebergman closed 5 months ago

eddiebergman commented 6 months ago

This is more of a bandaid fix such that the error/traceback shows up at least somewhere. For debugging pipelines and initial setup, there should be some option where the error just bubbles all the way up and crashes the program.

eddiebergman commented 5 months ago

I learned something new today, I did not know logger.exception does that by default! Good to know :)

In that case, I think the main feature that should be added is to give an option to raise immediatly on failure. This makes developing easier, i.e. no one will get their evaluation function right first time.

As an addendum, I also think that the error logic in the parsing of the result should always mention what it got as the result that it failed to parse. Simply saying something didn't work isn't very informative

I will close this PR as that seems like two things: