Runa Eschenhagen

No, this looks exactly like what we need for an improved CL interface. I'll look into the details when the PR is ready.

> default 100 linewidth? I think that's fine. If not, I'd rather have something like 90 than something > 100. > strings with ' instead of " (this wasn't even...

@ArturPrzybysz One other thing I randomly noticed when briefly looking at your code was that the tuning of the prior precision is probably not working as you intended: you [pass](https://github.com/ArturPrzybysz/LaplaceConvDemo/blob/main/src/la_conv.py#L45)...

When this PR is ready and our backend also supports regression with ASDL, we can also change the default backend to `AsdlGGN`. This avoids issues with quite a few architectures,...

Hi Nikola, Thanks for your interest in our library. At first glance the reply you got on Stack Overflow looks correct to me and the issue is not related to...

`BCEWithLogitsLoss` support was just added to `curvlinops` in https://github.com/f-dangel/curvlinops/pull/99.

Sounds good, can you create issues for all comments that you want to address in future PRs?

Included in #44.

I tried reproducing the issue by running the target setting run on the current `dev` branch with `pytorch_eval_num_workers=4`, but I don't see the drop in eval metrics compared to an...