Hi All,
I have a latent discrete variable in my model (so cool that you guys got that working btw) but I don’t see any posterior samples in the samples object returned from NUTS. This makes sense, since you guys are integrating out the discrete parameters, but it would still be nice to see some estimate (whether that is map or even just monte carlo estimate of mean or w/e). In my case, the parameter is interpretable and I would like to make some claims about its values in the associated paper.
Thanks!