Hidden Markov model in pyro and numpyro I took this great unsupervised example written in pyro and translated it into numypro (jax). Also check this article how to do this in pymc