aghoshpub

Swish activation saves beta as weight even if it is not trainable

2

**- What I did** Allows to save beta as weights when a model is saved even if beta is not trainable. **- How I did it** Adding "trainable" argument to...

How is the error calculated for Permutation Importance?

16

If we do: ``` perm = PermutationImportance(D, random_state=1, n_iter=2, scoring=significance_scorer ).fit(X_test,y_test)``` ```eli5.show_weights(perm, feature_names = data.columns.tolist())``` Then we get some scores with ```+-``` errors. But these errors don't correspond to ```std/sqrt(n_iterations)```....