You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We recently updated find_map() to return both raw and transformed random variables (#2523). However, I noticed that means = approx.gbij.rmap(approx.mean.eval()) (http://docs.pymc.io/en/latest/notebooks/GLM-hierarchical-advi-minibatch.html) only returns the means of the "raw" parameters, not of the original variables of interest. Should we perhaps add functionality to provide the transformed values as well, which are the parameters of human interest?