Vladimir Starostin

Here is the basic implementation of the convolutional layer with KAN, which seems to work just fine on the mnist dataset. https://github.com/StarostinV/convkan

I am pretty sure. Please share the link so that we can compare the implementation. Everybody would benefit from that! EDIT: I [found ](https://github.com/paulestano/LeKan) it, looks good! My implementation supports...

> Be my guest mate https://github.com/paulestano/LeKan I was not aware that one could use unfold as a module. However, your implementation lacks support for padding_mode and groups, and it has...

> > The main reason I did it was my surprise that nobody else had done so, especially given that others have argued it's impossibe in another open issue on...

> > @hesamsheikh Basically, you can rephrase a convolution operator as a simple matrix multiplication as in Linear layer, that's why if we could replace the normal Linear layer with...

> Hi, here I implement ConvKAN with different activation formulations with their corresponding inference time. https://github.com/XiangboGaoBarry/ConvKAN-Zoo We evaluate the result on CIFAR10 dataset. You can also make a pull request...

Do I understand correctly that it's about changing `sbi.inference.posteriors.score_posterior.ScorePosterior` to `sbi.inference.posteriors.vector_field_posterior.VectorFieldPosterior` in `mkdocs/docs/reference/posteriors.md` ? I thought it is updated automatically. :/ I can fix that if you haven't already @janfb...

This might be a possible solution for sharing the protocol among different estimators (if I understood the protocol logic correctly): ```python ConditionalEstimatorType = TypeVar( 'ConditionalEstimatorType', bound=ConditionalEstimator, covariant=True, ) class ConditionalEstimatorBuilder(Protocol[ConditionalEstimatorType]):...

It's great to have it fixed! This PR will have merge conflicts with #1497 , so I decided to take a look. I have a question - as far as...

> We can absolutely support this, but we could also just point people to the [training interface](https://sbi.readthedocs.io/en/latest/advanced_tutorials/18_training_interface.html) for achieving this. I'm sure power users can find ways to work around...