Rasa NLU only extracts a single entity even though training data contains multiple entities. This seems to be a common problem, but the solution I found on this forum is only to add more training examples. So I added more training data, around 30 per intent (maybe thats not enough) with every single example containing 2 or more entities but its still only extracting the one at the end of the sentence. I checked the configuration and couldn’t find anything. Im using the classic spacy_sklearn pipeline.
Also another Issue: I have some single word examples for Intents like hello, goodbye and stuff. Funny enough it always extracts the opposite of what im saying. If I put in “hello” it extracts Intent: Goodbye and if I put in “goodbye” it extracts “hello” even though these are the exact training examples I provided. I am adding lookup tables now so if that is the answer don’t bother answering this issue
Im running rasa nlu in python on windows10 with spacy_sklearn.
Thank you in advance!