Seth Axen
Seth Axen
@MansMeg I have created a repo [PosteriorDB.jl](https://github.com/sethaxen/PosteriorDB.jl) that, similarly to `posteriordb-r` and `posteriordb-python`, provides functionality for working with posteriordb. If you haven't started on a Julia repo yet, would you...
Great! IIRC, the process would be: 1. I transfer ownership of the repo to a user who has repository creation rights in the stan-dev org 2. That user then transfers...
+1 this would also be great to have!
+1 for this. Currently needing polar axes and using TikZ. Would be nice to use Makie.
I think with #458, a single-line change to loop over `["deps", "weakdeps", "extras"]` instead of just `["deps", "weakdeps"]` would resolve this issue.
I agree. Also, for `[extras]`, people may want the ability to opt out of updating that particular section's compats (similar to how they currently need to opt in to `test/Project.toml`...
Yeah, makes sense.
> I'm down with this, but it's worth pointing out that just calling `rand(rng, condition(model, x))` is probably not the greatest idea as it defaults to `NamedTuple` which can blow...
Hrm. Maybe then `predict` should use a `NamedTuple` if `x` is a `NamedTuple` (imperfect because you can have few parameters but many data points). Or provide an API for specifying...
How does Enzyme treat complex numbers? e.g. if I wanted to also support `BLAS.dotc`, would the `Duplicated`s all store complex derivative vectors, and would the `Active` store a complex derivative?