Lukas Heumos comments

Results 522 comments of


                                            Lukas Heumos

(scCODA) Final Parameter and log2-fold change have different signs & run_hmc error

Hi, could you please ensure that your issue is formatted correctly and could you please provide a minimal working example to reproduce this?

(scCODA) Final Parameter and log2-fold change have different signs & run_hmc error

That would work. Could you text me on https://scverse.zulipchat.com/ so that we can coordinate this? Cheers

Issues with MLP convergence in dl.multilevel_modeling

Dear @Mari123i I'm sorry that you're running into this. We already have an overhauled version of DIALOGUE in the works but it'll take some time for it to land on...

Add a checklist when creating a PR

I would not use an automatically created comment for this. Instead, we can add a pull request template: https://github.com/theislab/ehrapy/blob/main/.github/pull_request_template.md They're pretty ugly but it's still the best Github has.

`assign_mixture_model()` is really slow, even using JAX with CUDA support

Thanks! We haven't really performance tested it on GPUs but something seems to be off. It might take a week or two for us to get to it as the...

`assign_mixture_model()` is really slow, even using JAX with CUDA support

Got it, thanks! We'll try to look into it soon. Please remain patient for a bit.

`assign_mixture_model()` is really slow, even using JAX with CUDA support

I'll look into this in about 14 days.

Inconsistent covariate formatting for credible scCODA effects

Thank you for digging into this already! I'll get back to you in a few days.

Inconsistent covariate formatting for credible scCODA effects

@johannesostner do you have any reasoning? Otherwise, I'll simplify and harmonize this.

Inconsistent covariate formatting for credible scCODA effects

@mschilli87 I am in favor of making them pretty and more importantly consistent. However, if we do so, we'd need to ensure that we don't break existing user code. It...