English|French
Create a French chatbot for answering questions based on a document using Google Colab. In our case, we use data extracted from France law website: French Law Website
Dataset extracted from the website can be found in "data.csv".
Two models (based on Camembert model) are used for creating answers:
-
Dense Passage Retrieval Model: This model is used to efficiently retrieve relevant passages of text from the document.
-
Question Answering Model: This model is used for answering questions based on the retrieved passages.
Model being used for question answering: Camembert Base SQuAD-FR