enhancementhelp wanted
描述
Cannot find the way to ignore UNK words when numericalising, i.e. instead by substituting them by a 0, it just ignore that word.
Is that implemented?
This is useful in classification problems, when you just want to remove 'UNK' words.