shap icon indicating copy to clipboard operation
shap copied to clipboard

[BUGFIX] ignore_inds should not be tuple

Open peterdhansen opened this issue 3 years ago • 2 comments

ignore_inds doesn't work unless there is only one index to ignore

np.where returns a tuple the same length as the number of dimensions of the array. since the array has 1 dimension, we need to take the first index

Example Error

ipdb> i in ignore_inds
*** ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()
ipdb> ignore_inds
(array([ 1,  5,  7, 12, 20, 27, 30, 33, 34, 36, 40, 41, 42, 43, 44, 45, 47,
       48, 49], dtype=int64),)

peterdhansen avatar Sep 10 '21 15:09 peterdhansen