Maximilian Pichler

scale ALE/PDP plots to the same y range

threads / parallelism

1

- torch::torch_set_num_threads(n) - torch::torch_set_num_interop_threads(n) missing

It would be nice to have some of the SOTA local diagnostic methods - [Anchor](https://ojs.aaai.org/index.php/AAAI/article/view/11491); [Book](https://christophm.github.io/interpretable-ml-book/anchors.html#fn61) - [SHAP](https://christophm.github.io/interpretable-ml-book/shap.html) (can be also used to get feature interactions) - Lime (but doesn't...

Or is it too much work?

ALE and PDP variable/behaviour

1

- if variable = NULL, PDP should be calculated for all variables - it should be also possible to pass a vector of variables names

- add details and description section to dnn (see f407521997893fa37216bde62793835a02735f60) - improve argument documention (e.g. see optimizer argument, which optimizers are available?) - improve GPU documentation (explain the details section...

progress bar tokens

3

Hi @mihaiconstantin just to clarify, there is currently no way to pass tokens from the workers to the progress, right? with progress: ``` steps = 10 pb = progress::progress_bar$new(total =...

no progress bar with backend = NULL

1

Hi @mihaiconstantin thanks for the great package. Is it intended that there's no progress bar for backend = NULL?: ``` parabar::configure_bar(type = "modern", format = "[:bar] :percent :eta", width =...

first and second order derivative of gamma samples

1

Hi @dfalbel, Why can I calculate the first but not the second order derivative of gamma samples? It would be really great if I could . ``` alpha = torch_tensor(0.5,...