You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've been following Orama for quite some time now. That's why I want to first say a great thank you to the whole team to provide this amazing universal and local search solution! The hybrid search is also a very unique feature.
Now, I was wondering if you already considered providing the option to enable a binary index for vector search?
This could boost performance 10x or more, which is especially interesting on local/edge devices or servers with millions of records.
In this regard, this article describes an interesting technique that basically uses a binary index for a wide retrieval followed by a KNN re-ranking to boost accuracy ( up to 95% ).
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
I've been following Orama for quite some time now. That's why I want to first say a great thank you to the whole team to provide this amazing universal and local search solution! The hybrid search is also a very unique feature.
Now, I was wondering if you already considered providing the option to enable a binary index for vector search?
This could boost performance 10x or more, which is especially interesting on local/edge devices or servers with millions of records.
In this regard, this article describes an interesting technique that basically uses a binary index for a wide retrieval followed by a KNN re-ranking to boost accuracy ( up to 95% ).
Binary Vector Search: The 30x Memory Reduction Revolution with Preserved Accuracy
Also, the HN comments on the article provide some more interesting methods.
Let me know what your thoughts on this are.
Beta Was this translation helpful? Give feedback.
All reactions