Variational inference with implicit probabilistic models

Hi, everyone. I'm not sure if I am at the right place to ask, but I'll try it anyways :-)

I've been trying to implement the algorithm proposed in Tran et al. (2017) - "Hierarchical Implicit Models and Likelihood-Free Variational Inference" ([arXiv](https://arxiv.org/abs/1702.08896)) and I admit I'm having quite some trouble to get it done.

I have found a piece of code that actually implements the implicit KL training proposed in the paper [`implicit_klqp`](https://github.com/blei-lab/edward/blob/master/edward/inferences/implicit_klqp.py), but it is done on Edward1 and AFAIK there's no accompanying script showing how to use the function in practice.

Since the project for Edward1 seems quite inactive nowadays, I was wondering whether people over here could help me out to find a way of using the algorithm proposed in the paper that I mentioned :-)

Cheers,
Pedro