Gabriel Stechschulte

> I don't know all the details of the optimization methods. I guess we would get a single number for every parameter in the posterior, right? If that is the...

> LGTM in terms of working with `bayeux`. It looks like CI needs to be fixed and @tomicapretto give final approval for it being in the `bambi` style. > >...

> @GStechschulte I think this is a great addition and the PR is in great shape. Just want to know your opinion on my suggestion about handing, for now, the...

> @tomicapretto @GStechschulte will the samplers work with Bambi's mixture modes like the zero-inflated and hurdle models? > > Z. Yup! Though, some samplers are better suited for different problems,...

Ugh. pylint is making the CI fail. It says it cannot import bayeux. However, when I check the logs of the step "Install Bambi and all its dependencies", I can...

> Thanks! We also need to figure out if the licenses are compatible and how to do proper attribution if you took inspiration from someone else's implementation. It looks like...

Hey @rlouf and others, I will give this PR a go. I am following #1335 and #1284 for more explanations on the implementations. Likewise, I will comment here on progress...

After using NumPyro, I remembered that they have a JAX implementation of the [Multinomial distribution](https://num.pyro.ai/en/v0.2.4/_modules/numpyro/distributions/util.html), albeit following the design of the PyTorch distributions module. Therefore, I adapted the code to...

@danieltomasz, yes, we are preparing a new release. Not 100% sure on the time frame; any thoughts @tomicapretto

Hey @NathanielF thanks for raising the issue. This should work. What version of Bambi are you using? A couple months ago we added enhancements of how the interpret package parses...