I have already implemented my Rasa chatbot and the UI is accessible through HTML page. But I cannot see how I can do the same with JupyterLab in Amazon SageMaker studio. I will be very grateful if someone can help me. Thank you!
Technically, it’s just some commands from the command line. You may notice how, in Jupyter, when you run:
pwd as if it was ran from the command line. That means that if you’ve gotten your data in the right folders you can simply train via;
! rasa train
Serving the model might be trickier and more involved though. I remember that SageMaker works by spinning up containers, but how exactly it wants the models served is something the Amazon docs can tell you.
I gotta ask though, wouldn’t a Rasa X installation be simpler if you want to host the bot?