How does Rasa match the examples in the Intent? Does it have to be a 1to1 match?

Hi, Only started playing with this today, very cool! But I have a question.

Looking at some of the training Yaml files such as this https://github.com/RasaHQ/NLU-training-data/blob/main/banking/nlu.yml does the users input have to be an exact match to the examples it trained on for it to match the Intent?

Or will it still match if the grammar or spelling is slightly off?

Thanks :slight_smile:

Happy New Year!

Welcome to the forum :slight_smile:

The user input does not need to be an exact match. Grammar and spelling mistakes are okay. Of course, the more examples you have for your intent, the better.

Don’t forget to include DIET Classifier in your pipeline in config.yml!

1 Like

Hi Chris,

Thanks for the info!!

All the best.

1 Like