dianna
dianna copied to clipboard
Deep Insight And Neural Network Analysis
This commit makes dianna swallow more kinds of data items without users having to mangle them manually themselves. As long as your data is numpy ArrayLike it will go. The...
Due to the lower limit on the Tensorflow version that dianna uses, I cannot use dianna on my particular GPU setup with its particular CUDA version installed (DAS6 @ VU)....
Removed from the repo the models and data that can be downloaded from Zenodo.
still need to add: - [ ] link from documentation to notebook - [ ] references from readmes
Elements to include: interview with Joeri about scientific case interview with Leon about DIANNA show images of frbs, telescopes, etc dashboards? audience would be other astronomers goal messages: Dianna is...
in lime_tabular, `None` is set as a default value for the input argument `feature_names`, see [here](https://github.com/dianna-ai/dianna/blob/48a9994ac4cb652b4c6e54217d6bf0cf9df1e62d/dianna/methods/lime_tabular.py#L17). Later, the variable `self.num_features` is calculated as `len(feature_names)` which does not work on NoneType...
The method kernelshap generates a tqdm progress bar, see [here](https://github.com/shap/shap/blob/8c8ee841a6d9d095ae23ac3f86246cde746e6540/shap/explainers/_kernel.py#L267). It is useful to be able to hide the progress bar if explainer should be run several times. The original...
Preferably remove the folder completely See #171
2 Teaching video Teaching astronomers how to use dianna. Notebook tutorial Similar to Giulia’s video. Can be longer if needed. Based on the FRB tutorial. See also issue #785
A speech recording can be seen as a time series. Usually sound analysis is done with another view of the data than the raw signal that we show now. Maybe...