DIETClassifier not working properly

Rasa Open Source
1

Hi Team, recently I install rasa 1.8.4. And as per the documents, I have started using DietClassifier istead of EmbeddingsIntentClassifier. When I start running the rasa as a http server and parse a message. I only get one enitity in the output, that too with no confidence value. But when I run rasa shell nlu in the terminal, I get multiple entities for the same message with confidence values as well. Any idea why Rasa is acting like this or I am doing any mistake from my end ?

2

Hi @anirbankar983,

can you please post the content of your config.yml such that we can check on that one? If your data isn’t confidential, then a brief part of your nlu.md would also be good, aswell as the corresponding story.

Kind regards
Julian

3

Configuration for Rasa NLU.

Components

language: “en”

pipeline:

  • name: “ConveRTTokenizer”
  • name: “RegexFeaturizer”
  • name: “ConveRTFeaturizer”
  • name: “EntitySynonymMapper”
  • name: LexicalSyntacticFeaturizer
  • name: CountVectorsFeaturizer
  • name: CountVectorsFeaturizer analyzer: “char_wb” min_ngram: 1 max_ngram: 4
  • name: “DIETClassifier”

Configuration for Rasa Core.

Policies

policies:

  • name: MemoizationPolicy
  • name: TEDPolicy
  • name: MappingPolicy

Sorry, I can’t post the nlu contents as it is confidential.