Fritz Obermeyer

OK, so if I understand correctly, the idea is to use a user-provided covariance/precision/scale_tril estimate. Since we eventually need a cholesky decomposition, let's assume wlog that the user can provide...

> `scale_tril_fn` needs access to the model I guess you could use either `functools.partial(fn, model)` or define a model as a class with a `scale_tril_fn` method and then pass `model.scale_tril_fn`...

Hi @dirmeier, we'd welcome contributions of any new automatic guides. For compatibility with Pyro, we'd like to reserve the name `AutoStructured` to the design in Pyro that @fehiepsi pointed to....

Thanks, we're behind on torch versions at the moment...

@gioelelm could you check to see if #3069 fixes your issue?

Thanks for checking @gioelelm. I might have time in the next few weeks to dive deeper. If you have time I can recommend some strategies (what I'd try): - get...

> Don't you think that the bug could have gone unnoticed? It could have, but `TraceEnum_ELBO` is pretty heavily used, and we've done a lot of memory profiling in the...

Hmm maybe we should relax pytorch requirements and release to allow Pyro 1.8.2 to work with PyTorch 1.10. We'd need to do the same with Funsor. I think I was...

cc @stefanwebb

@OlaRonning sounds good, I'm in no hurry. I guess the first task is to find a dataset, then I could start a tutorial in the next couple weeks.