After this solution LIME (plot_features) does not give factors for outcome variable (for example: labels = 'Yes', 'No') as 'No' and 'Yes', instead of them I get '1' and '2'. This problem is actual for Keras.
Today I have used Lime for h2o automl and when I specify ' bin_continuous' as TRUE, I get the explanation of continuous variables. What about the rest of data...
Hi @thomasp85 I have several problems with Lime.