Michael Osthege

Results 202 comments of Michael Osthege

Hi @msibaev , Sorry about the delay. I just got the chance to read your descriptions and look at the changes. I believe the history was such that in the...

Hi @msibaev , I see you're working on this branch again. Cool! Let us know when it's ready for review

Yes, we should be more strict about the CI. A red ❌ CI on main should be an immediate call to action to get it fixed. Something like the past...

@OriolAbril would you say the simplifications made by #5911 close this issue? If not, which are the most important items that we can track in separate smaller issues?

The google search now leads to this issue, and the old page didn't show up for me at all.. Also, we have abandoned `v3`, so... `wontfix`

Yeah, this is indeed very annoying. By deleting the `Distribution.__new__` docstring, VS Code shows the right docstring, but still with the uninformative signature. Maybe one can monkeypatch autogenerated `__init__` methods...

I think @ricardoV94 is right that we should refactor distributions to be functions. This is, however, a major refactoring and needs to be done right. The logic from `pm.Distribution` doesn't...

When the roadmap is clear, I'd like to help with this one.

This sounds serious! If you look at the [control flow of `Model.__init__(coords=...)`](https://github.com/pymc-devs/pymc/v4.1.3/main/pymc/model.py#L579) you can see that internally it gets routed into `Model.add_coords` and thereby `Model.add_coord`. But in the example you...