I am running the below code to fit my model:
advi_fit = pm.fit(method=pm.ADVI(), callbacks=[CheckParametersConvergence()])
When I run below commands to get mean and standard deviation:
means_dict = advi_fit.bij.rmap(advi_fit.params[0].eval()) and means_dict = advi_fit.bij.rmap(advi_fit.params[1].eval())
I see some variables renamed by adding _lowerbound__ , for example- alpha_lowerbound__ and some have kept their naming structure.
I am not sure what the lowerbound means here and also most of the standard deviation values are negative. I used truncated normal with lower bound = 0, for those values mean value is less than zero. I think I lack an understanding of the output of the code. Ideally, I am looking to get the mean and standard deviation values of the variables.