Hi, has anybody tried making a deep markov model in PyMC3?
You mean something like below?
I dont think anybody is working on this, but since the repository is in Theano it might not be too difficult to get the loss function and play with it in PyMC3.