InvalidConfigException: Can't load class for name 'ConveRTTokenizer'. Please make sure to provide a valid name or module path and to register it using the '@DefaultV1Recipe.register' decorator

Rasa Open Source
1

Hi all,

I’m using the configuration from the RASA doc and get this error:

InvalidConfigException: Can't load class for name 'ConveRTTokenizer'. Please make sure to provide a valid name or module path and to register it using the '@DefaultV1Recipe.register' decorator.

My config.yml looks like this:

pipeline:
  - name: ConveRTTokenizer
  - name: ConveRTFeaturizer
    alias: "convert"
  - name: RegexFeaturizer
    alias: "regex"
  - name: LexicalSyntacticFeaturizer
    alias: "lexical-syntactic"
  - name: CountVectorsFeaturizer
    alias: "cvf-word"
  - name: CountVectorsFeaturizer
    alias: "cvf-char"
    analyzer: "char_wb"
    min_ngram: 1
    max_ngram: 4
  - name: DIETClassifier
    epochs: 300
  - name: EntitySynonymMapper
  - name: ResponseSelector
    retrieval_intent: faq
    featurizers: ["convert", "cvf-word"]
    epochs: 100
    constrain_similarities: true
  - name: ResponseSelector
    retrieval_intent: chitchat
    featurizers: ["convert", "cvf-word"]
    epochs: 100
    constrain_similarities: true
  - name: ResponseSelector
    retrieval_intent: booking
    featurizers: ["convert", "cvf-word"]
    epochs: 100
    constrain_similarities: true
  - name: ResponseSelector
    retrieval_intent: refund
    featurizers: ["convert", "cvf-word"]
    epochs: 100
    constrain_similarities: true
  - name: ResponseSelector
    retrieval_intent: planning
    featurizers: ["convert", "cvf-word"]
    epochs: 100
    constrain_similarities: true
  - name: FallbackClassifier
    threshold: 0.3

Any idea why this is happening? Many thanks!

2

Hi @MelT, the documentation says that the ConveRTTokenizer tokenizer is deprecated.

1 Like
3

Oh okay, so should one use just a different tokenizer? And is the ConveRTFeaturizer still the recommended Featurizer for English based bots?

4

SpacyFeaturizer and CountVectorsFeaturizer would suit well for the English language.

5

Okay, thanks! So you would say that RASA themselves would not recommend the ConveRTFeaturizer anymore as the “best” option for English?

6

I am not sure if it was recommended as a “best” option for English.

7

I got this reference that they recommend the ConveRTFeaturizer from here

If your training data is in English, we recommend using the [ConveRTFeaturizer]

Maybe this is not up-to-date.

And just a follow up question, out of curiosity, have you tried SpacyFeaturizer and ContVectorsFeaturizer with English and would you suggest one over the other?