Eliminate requirement to return loss and thus require label key #819
Description
conditionally compute loss so the interpretability method don't need to worry about faking a label tensor
{{ message }}
conditionally compute loss so the interpretability method don't need to worry about faking a label tensor