How to evaluate slot-filling?


I’m trying to run some evaluations on the model that I’ve trained.

I use the command rasa train nlu for training and rasa test nlu for evaluating.

However, the results that are returned (the ones inside the folder “results”) refer only to the performances of intent prediction, but don’t say anything about how well the entities were predicted.

How do I evaluate this?

Thank you, Tiziano