F-score is ill-defined and being set to 0.0 & won't converge about sdgym

[oregonpillow]

Hi @csala ,
that's really helpful. Thanks for writing such a detailed explanation. I will try with alternative synthesizers and be more patient with my execution ;)

[csala]

Hello @oregonpillow

This is completely normal and expected. All these messages come from the scikit-learn estimators and metrics, and are caused because either the synthesized data is not good enough for the estimator to work in it, or the estimator is not performing well enough on that particular problem. Since we are running a bunch of different estimators on a bunch of different datasets, having those warning messages appear randomly is both expected and unavoidable (unless you explicitly silence them :-P). Also, note that the Quickstart is using the IndependentSynthesizer because it is the simplest and fastest one, but also the one that performs the worse.

I also notice that on your notebook you interrupted the execution, but if you do not, and wait enough, at the end you'll get the results table.