Why CountVectorsFeaturizer is used twice in config.yml?

Rasa Open Source
1

Below is the part of config.yml generated by rasa init. It shows CountVectorsFeaturizer is used twice. Why is it so?

- name: WhitespaceTokenizer
- name: RegexFeaturizer
- name: LexicalSyntacticFeaturizer
- name: CountVectorsFeaturizer
- name: CountVectorsFeaturizer
  analyzer: char_wb
  min_ngram: 1
  max_ngram: 4
- name: DIETClassifier
2

In the first use, a bag of words representation is generated for the individual words. The second, creates features for n-grams from 1 to 4 characters within word boundaries. Both could be useful for intent classification in the DIETClassifier.

There’s a good blog post on the NLU pipeline here.