search

thomasp85 / lime

Local Interpretable Model-Agnostic Explanations (R port of original Python package)
https://lime.data-imaginist.com/
Other
483 stars 110 forks source link

Flow ... through to the interactive_text_explanations #180

Open alanault opened 4 years ago

alanault commented 4 years ago

Hi there,

Love the package and have been using the interactive explainer to dive into some of our NLP models.

We have quite complex models, and the latest version of the LIME 0.5.1 broke our wrapper around the interactive explainer.

We typically pass additional arguments to the text_explainer, but they're not pushed through to the underlying explainer. e.g. we call it like this

lime::interactive_text_explanations(explainer,
                                        max_feature_to_select = 10,
                                        problem = problem)

I think all we need to adjust to make this work is line 90 of the shiny.r function where the call to explain needs to have ... = ... added and the function itself just needs ... added as an argument.

Thanks! Alan