Has anyone tried working with Rasa NLU with the fast.ai library? Based on my preliminary investigation it has native support only for the spacy_sklearn and tensorflow_embedding pipelines.
I’m not completely following the Q. Rasa is an NLU engine that you can use with other chatbot and conversation management frameworks, if you don’t want to use the NLU engines they often provide. I once used it with botkit. If you had a use case where rasa or wit.ai or Luis or diagflow didn’t meet your needs but home-rolled NLU with fastai did, I think I’d try integrating fastai with a bot framework, not with rasa. A fastai powered rasa competitor would be a potential project. IMHO. Happy to understand if you see it differently.