The goal here is to have Imputation layer for both float and categorical values and I DO NOT want to change/preprocess the type of array that is used as imput. The problem i
