AtrCheema
hyperopt.utils.Real: ValueError: prior should be 'uniform' or 'log-uniform' got log
For the record; I will submit a PR for these kind of things, if appropriate.
Running compare_ml.ipynb at commit a8dae5065fbb236272c7e2797bed9f2f35139cb1, comparisons.fit(run_type="dry_run") leads to:
"prior should be 'uniform' or 'log-uniform' got log"
visibly in skopt, using version 0.9.0
This is likely related to the change made for scikit-optimize #1017
@jmp75 Thanks for pointing it out. But How is commit a8dae50 related to this error or related to running of 'compare_ml.ipynb'?
Moreover, as you might have noticed, we do development in dev branch so any PR should be directed to it with tests whenever possible.
Did you get this error even when skopt version is between 0.22 and 0.24?
The commit hash is superfluous here; just a habit for repro.
I have not tested with scikit-optimize version 0.8.1 (0.9.0 is less than a month old), but it very probably worked. Just a change of behavior to know about skopt. Not sure what you meant with versions 0.22 to 0.24 which seem not related skopt.
I started from the dev branch indeed. I won't submit PRs without first asking by the way; for now just capturing on a fork in case they are of value in my "discovery process"..
oh my mistake. You are right. Just ignore my comment about skopt's version.