Turing Mixture Models with Dirichlet weightings

@torfjelde Is that a similar issue as for ADVI here?