Alexander März

> I will check if I can set up the deployment using Poetry Ok nice, thanks. > Yes, I will look into poetry packaging and change up the files accordingly....

@Cattes I believe that if we want to also allow `pip install git+https://github.com/StatMixedML/XGBoostLSS.git` we still need the previous setup.py. So we should still keep it with the updated package version...

@Cattes It might take some time until I find the time to re-visit and review the PR. One thing I noticed is that for the unit-tests, we might want to...

@Cattes > I added the one example test because I received a deprecation error from xgboost about the use of the feval parameter. Because of that I had to switch...

@Cattes I am planning to fully migrate to pytorch-distributions, so that analytical gradients and hessians get replaced with autograd ones. To what extend would that affect the PR?

@m1sta Thanks for the question. Choosing between the two depends on the problem at hand I'd say: - the quantile regression feature of catboost allows you to model different parts...

You are so right, fully agree. I am currently also working on combining XGBoostLSS and CatBoostLSS into one single python-library so that you can chose which boosting algorithm to use...

> What if their are updates to the underlying libraries in the future, do you plan to track changes in them? Yes, would update and keep track of the major...

@varunbalupuri Kindly referring to https://github.com/StatMixedML/CatBoostLSS/issues/8#issuecomment-782819290

I don't find time to work on this problem https://github.com/StatMixedML/XGBoostLSS/issues/8 which is also true for catboostlss. In case you have a solution let me know.