Migration from 1.x to 2.3.4 config.yaml pipeline

Rasa Open Source
1

Hello there. I am trying to migrate a small application from 1.7.4 to 2.3.4. I am following these resources:

My config.yml looks like this:

language: en
pipeline: supervised_embeddings

Now it is my understanding that in rasa 2 I have to specify the different stages of the pipeline, the featurizers and whatnot.

My question is, what’s the most direct replacement for the line pipeline: supervised_embeddings ?

Thank you all for your help.

2

Hi there!

The pipeline supervised_embeddings was composed of the following components:

- name: "WhitespaceTokenizer"
- name: "RegexFeaturizer"
- name: "CRFEntityExtractor"
- name: "EntitySynonymMapper"
- name: "CountVectorsFeaturizer"
- name: "CountVectorsFeaturizer"
  analyzer: "char_wb"
  min_ngram: 1
  max_ngram: 4
- name: "EmbeddingIntentClassifier"

EmbeddingIntentClassifier was deprecated, if you want to replicate the behaviour you can use DIETClassifier with the following parameters:

- name: DIETClassifier
  hidden_layers_sizes:
    text: [256, 128]
  number_of_transformer_layers: 0
  weight_sparsity: 0
  intent_classification: True
  entity_recognition: False
  use_masked_language_model: False
  BILOU_flag: False
3

Thank you very much for your help!

1 Like
4

You’re welcome! Happy bot-building :slight_smile: