IKH

Pipelines of NLU model-Rasa

For customising your model and how to act on your data set, you have to choose an  NLU pipelines.

The two most important pipelines are:

  • supervised_embeddings
  • pretrained_embeddings_spacy

The biggest difference between them is that the pretrained_embeddings_spacy pipeline, as the name suggests, uses pre-trained word vectors from either GloVe or fastText, whereas the supervised_embeddings pipeline feeds specifically on your data set and doesn’t use any pre-trained word vectors.

So it is generally recommended that you use the pretrained_embeddings_spacy pipeline if you have less than 1,000 total training examples, and there is a spaCy model for your language. Or use the supervised_embeddings pipeline if you have more than 1,000 examples.

You can specify the pipeline by defining them in config.yml.

To use the supervised_embeddings pipeline, the template is:

Python
language: "en"

pipeline: "supervised_embeddings"

To use the pretrained_embedding _spacy pipelines,the template is

Python
language: "en"

pipeline: "pretrained_embeddings_spacy"
 

You can read more here on choosing the pipeline for your NLU model.