I’m currently trying to build a multi-lingual chatbot. I’m trying to re-create the method used in this video . My question is how to create and the use the entity mapping file. Thank you.
Could you give an example of what you’re trying to achieve? It’s easier to answer when it’s included on the post than linked to in a video.
I’m trying create a file that maps entities of different languages to one entity that’s named with a code like in this picture.
If you have multilingual data, then you can use regular synonyms i.e.
## synonym:balance_statement
- balance statements
- relevés
...
If however you have multiple NLU models, but still wanted the value sent to the tracker to be the same for each language, you’d just provide the same synonym name in each language e.g.
English nlu.md
## synonym:balance_statement
- balance statements
- transfer statement
French nlu.md
## synonym:balance_statement
- relevés
- relevé de compte
....
Then you if your slot was called balance_statement, it would be auto filled for both languages.
I’d recommend reading more about synonyms here: Training Data Format