SeanMcCarren
SeanMcCarren
Hi @adrinjalali @romanlutz, I think scikit-learn's `LabelEncoder` can be used? I could give it a try if this issue is still relevant :)
I've delved a bit deeper into the issue now. Starting with a list of algorithms to update (as @adrinjalali mentioned): - [ ] threshold optimizer - [ ] exponentiated gradient...
> * using TF or PyTorch in tests is a MUCH easier task than in the docs because we'd have to include these packages in the dev requirements to make...
> yes, but the issue there is that since it's not compiled code and the API changes quite frequently, it either becomes dead code or needs to be maintained, and...
Parts of adversarial fairness are hard to test without PyTorch or TensorFlow loaded, so the important test cases for adversarial fairness are in test_othermlpackages.
> In that case, it seems they shouldn't be a cron job, rather run for all PRs? I'm not sure what a cron job is, but this could run whenever...
If it is too much of a hassle, I could run my test cases locally? I don't expect to be needing to change much anyways. Of course without test cases...
At first, I found it slightly confusing that every subsection of Metrics (except for the first) is about grouped metrics, while the header of the second subsection is called 'Metrics...
Thanks for all the great feedback! Sorry, I took a while :) > > > ``` > > * What stylistic features should the plots have by default? I already...
With regards to the auto labelling, I like it but I also agree that I don't like an extra dependency. I think sort of mimicking their techniques would be feasible,...