Is it possible to run in parallel many chains for variational Bayes (vb from RStan)?
Basically for some models (of which all parameters have normal distribution checked with NUTS) I have to explore the tails of the posterior with many draws, and although the inference is somewhat quick, the sampling stage takes hours.
If I could split those draws among parallel chains would be amazing. I could do it manually but I don’t know how to print the progresses for all chains to the user. Usually with parallelism in R I loose the standard output.
Thanks