In this guide you will learn:
- How to write and train a story 📖
- How to create and train a simple NLU model 🤓
- How to add a bot response
Train and try your stories
Select stories in the sidebar, create a new group called Greetings, paste the content of 2 stories above (starting with
chitchat.bye) in two dictinct stories, and train.
Then open the chat window. Since we haven't trained our bot to understand natural language, we'll talk with intents. It can be done by prepending a
/ to the intent name. So you can start the conversation with
/chitchat.greet. You should be able to reproduce the conversation above: if you say
/chitchat.greet three times, you should get 3 different answers.
Add natural language
Select NLU from the side bar and go the Chitchat tab. Select
chitchat.bye, import and train.
Add a bot response
Select Bot responses in the sidebar, then create 4 bot responses. The names of the bot responses are those used in stories:
Congrats, now you can finally have a real (!) conversation with our bot!
You can easily add a new language to your project. Go to the project settings, and add a language. Once you save, your new language will be available in the top left selector.
Let's add our training data and retrain (since we added French we have ready to use data here as well).
And voilà, our bot speaks French!
You can dive in the official Rasa docs or simply follow the rest of the tutorial.