Fun with Apache Lucene and Bert Embeddings

dmitry-kan.medium.com

4 pointsdmitrykan3 comments

Second post in the series about building Neural Search with BERT and Solr -- with the focus on practical aspects, speed, memory consumption and semantics. Will appreciate your feedback!

Really great Dmitry! So happy to see the workaround after deprecation of CustomScoreQuery.

I just managed to get Hello-NLP to work with elastic for KNN using SBERT, but don't have it working for Solr yet. I can vectorize queries and text for Solr, but cant yet store/score. Interested in a collaboration?

It will be great to collaborate, Max! So far I have the storage part figured out. And scoring is currently applied to the entire set of indexed documents. My goal is to work on some sort of nearest neighbours approach. So it'll be great to see if we could port KNN over to Solr.

Reply on news.ycombinator.com