I’ve set in config that 0.7 is the limit, so 0.71 has passed but he isn’t extracting entities, why could that be?
Can you show me ur config.yml
file configuration?
# Configuration for Rasa NLU.
# https://rasa.com/docs/rasa/nlu/components/
language: en
pipeline:
- name: WhitespaceTokenizer
- name: RegexFeaturizer
- name: LexicalSyntacticFeaturizer
- name: CountVectorsFeaturizer
- name: CountVectorsFeaturizer
analyzer: "char_wb"
min_ngram: 1
max_ngram: 4
- name: DIETClassifier
epochs: 100
- name: EntitySynonymMapper
- name: ResponseSelector
epochs: 100
# Configuration for Rasa Core.
# https://rasa.com/docs/rasa/core/policies/
policies:
- name: MemoizationPolicy
- name: TEDPolicy
max_history: 5
epochs: 100
- name: TwoStageFallbackPolicy
nlu_threshold: 0.65
- name: MappingPolicy
Configuration seems to be fine, can you just test your utterances using the below command and let me know the output:
rasa shell nlu
The problem is with your Rasa nlu model it’s not able to predict the entity.
Do you have any idea why is that?
I’ve tried training again and noticed that my entity “ingredients” isn’t being registered at all, why is that?
I made a new project with this entity “ingredients”, copied all the code and it extracts the entities, it just doesn’t work in my project? Do you have any idea why could it that be?
@rasafan @JiteshGaikwad I have a similar issue but not related to slots, just entity extraction.
When I give the utterance “i have bacon berries baguette”
I get my output as one entity
"entity": "ingredients",
"start": 7,
"end": 29,
"confidence_entity": 0.9851168036734895,
"value": "bacon berries baguette",
"extractor": "CRFEntityExtractor"
Any idea how I could get the three separate ingredients out of it ?
Hi @Tobias_Wochinger facing the similar problem @shubhamnatraj has mentioned. Can you please share your insights on the issue as how can we get them as three separate ingredients entities i.e. ‘bacon’ ‘berries’ ‘baguette’ instead of one single entity as “bacon berries baguette”. Thanks
@noman @shubhamnatraj Your bug seems to be related to your training data. Can you show your nlu.md
file? Each entity has to be annotated separately, e.g.
## my intent
- [bacon](ingredient) [berries](ingredient) [baguette](ingredient)
@Tobias_Wochinger, yes, the data has been annotated separately. I made another post on this topic a while back and @Tanja confirmed that this is expected behaviour. When multiple instances of the same entity are picked up in a row, they are concatenated by default. Doesn’t happen if I put an ‘and’ in between.
@Tobias_Wochinger i have also annotated entities separately as you have mentioned. @shubhamnatraj can you point me to that post? and also can we stop this default rasa behaviour?. Thanks
Did you solved it? I 'm facing the same mistake. You can help me. Thanks.