eb8680

> enumerating a d dimensional Bernoulli vector would inevitably cost O(2^d) right? Yes, that's right. In our implementation of enumeration in Pyro (and NumPyro) we've chosen to have programs fail...

@fritzo can you take a look at this again? I was helping someone today with issues building the example HTML and this seemed like it would be a useful check...

@r3v1 no, sorry, it's unlikely I/we will have the bandwidth to fix this ourselves any time soon. You're welcome to play around with the branch linked above and see if...

@qinqian thanks for looking at this. Can you turn the example above into a unit test in `tests/infer/test_discrete.py` that fails without this fix and passes with it?

@hjnnjh can you provide a minimal example that reproduces the error? To work around this you could try passing a custom `create_plates` function to your guide, as in e.g. [this...

@hjnnjh great, thanks!

We don't have any plans for this, but open source contributions are welcome!

@null-a what do you think about the feasibility of this? Are there any major stumbling blocks you're aware of?

> Would be want to support models in which response variables have different response families? I think a first version supporting only `Normal` responses with uncorrelated residuals would be fine,...

@neerajprad Is this something we could support upstream with more generic transformations in Pyro, as in e.g. [this paper](https://arxiv.org/abs/1906.03028)?