Hello, good day to you!
I’m new to RASA and I’ve been reading on Pipeline components, custom config.yml in order to figure out if I can use my own model for predicting question intent and slot-filling then using RASA Core for question answering. I have no problem with using Rasa NLU other than the fact that we’ve already collected and prepared our data according to this implementation and trained our model so we are afraid that preparing our data to be suitable for training Rasa NLU now would be harsh on time. We have also read these answer regarding the usage of only Rasa Core:
But, still we are not getting how to use our model with Rasa Core. Now, as for our model. Given sentence “Flight from Denver to Philadelphia” . Our model outputs following:
- Intent: Flight
- Slots: “O O FromLoc O ToLoc”
We have knowledge base answers in json format. And now wish to use Rasa Core for question answering with our trained model. How would we go about it?
Thank you very much!