Hello everyone, iwish you all a happy new year. I have a problem In my training set, i have two section of intent and name’s entity.The name entity when it is in one of the section led the recognition of intent of a new message … I don’t know if it’s clear ^^ But i want to know if there is a way to separate entity name of recognition of intents ? In my case, sometime the sentence is exactly like the sentence in a section of my training set and the only difference is the name who is in the other section and the NLU fail to recognize the good intent … My pipeline is like this : language: “fr” pipeline:
- name: WhitespaceTokenizer
- name: RegexFeaturizer
- name: RegexEntityExtractor use_regexes: True
- name: “CRFEntityExtractor”
- name: LexicalSyntacticFeaturizer
- name: EntitySynonymMapper
- name: CountVectorsFeaturizer
- name: CountVectorsFeaturizer analyzer: “char_wb” min_ngram: 1 max_ngram: 4
- name: DIETClassifier epochs: 100 batch_strategy: sequence batch_size: 64
Thank you for your help …