TL;DR: When using fit() with a test set, how to handle labels in test set that are unseen in training set (apart from removing these rows from the test set)?. Longer explanation. When fitting the ...