Document Retriever and TF-IDF score

facebookresearch / DrQA

Reading Wikipedia to Answer Open-Domain Questions

Other

4.48k stars 898 forks source link

Document Retriever and TF-IDF score #251

Open JosephPai opened 4 years ago

JosephPai commented 4 years ago

Hi, thanks for the amazing system! I'd like to know how you build the sparse representation for the docs in db. Specifically, have you considered the text within the document? Or just build it according to the doc id? It's important to me because I only want to use the Document Retriever part. Thanks!