-
### Description
It would be nice to support multiple values in a Knn vector field.
This must be compatible with both the Exact and Approximate Nearest Neighbor search.
There are two sides to the …
-
Hi,
I want to perform a fixed radius search with FLANN using the autotuned index (to minimize the search time). My problem involves a large number of point (about a million) at a dimension around 5 -…
-
While performing some experiments and reading the code it seems for more than N(=4096) points the algorithm always does an approximate nearest neighbor search. Is there a way to force it to search for…
-
We're currently lacking space storage on OVH2, due to Robotoff. We store in PostgreSQL all logo embeddings (7M currently, but many are missing) + all image embeddings (7M images, but only ~1M are in D…
-
The approximate nearest neighbors algorithms (IVF-Flat #652, IVF-PQ) use a balanced k-means implementation. In #652, balanced kmeans is separate implementation from #605. These two implementations sho…
-
This is a super fast and still super effective approximate method for computing nearest neighbors and is especially useful for all-neighbors methods. We have had a plan to implement this for awhile bu…
-
**Is your feature request related to a problem? Please describe.**
The k-NN plugin adds support for a new field type, `knn_vector`, which can be thought of as an array of floating point numbers. Th…
-
Multiple approximate nearest neighbors methods will soon be made available as alternatives to bruteforce KNN into the `NearestNeighbors` model (see #2780). Let's make this issue a reminder that it mig…
-
### Description
Lucene always does an approximate nearest neighbors search when no filter is provided.
This seems like unnecessary work. Some benchmarks would have to be done, but some ideas I ha…
-