How to resolve urllib.error.URLError for ConveRT pipeline

Rasa Open Source
1

I am getting urllib.error.URLError: <urlopen error [Errno 8] nodename nor servname provided, or not known> when I am trying to run rasa train nlu using ConveRT in config.yml file. Please help me resolve this error.

2

Hi, can you share your config file? Note that ConveRT took their model off of the public url, so you will now either need to use a community-hosted model, or local model. See the note under “Configuration” here

3

Hey, thanks for replying. my config file looks like this-

  • 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: 100
  • name: EntitySynonymMapper
  • name: FallbackClassifier threshold: 0.75
4

Sure! It looks like you haven’t specified a model_url for ConveRTFeaturizer. You’ll need to find a community hosted model and specify it as a parameter

5

Okay, Thanks a lot.

1 Like
6

Alternatively, we’ve had good success using the pre-trained embeddings in LMFeaturizer. You could try to remove the ConveRT components, and just use LMFeaturizer + some tokenizer (WhitespaceTokenizer could work if you’re writing a bot for a whitespace-tokenized language)