Summary
In this chapter, we discussed how Rasa works. Rasa can be divided into two parts: NLU and policies. The key part of NLU is a pipeline composed of components. The core work of the strategy is to convert the tracker into input data that can be used in the model and then train the model. We also introduced how to write Rasa extensions for various functions. Finally, we showed you how to create and use a custom English tokenizer through a practical project.
In the next chapter, we will discuss testing and production deployment in Rasa.
