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
Tell us about your request. Provide a summary of the request and all versions that are affected.
As of today, k-NN plugin only supports vectors of type float for each dimension which is 4 bytes. This is getting expensive in terms of storage especially for those use cases that requires ingestion on a large scale as we need to construct, load, save and search graphs which gets more and more costlier. There are few use cases where customers prefer to reduce memory footprints by trading off recall with a minimal loss.
Using the Lucene ByteVector feature, we can add support for Byte Sized Vector where each dimension of the vector is a byte integer with range [-128 to 127].
As of now, we are planning to support this feature from 2.9 and future versions of OpenSearch.
What other resources are available? Provide links to related issues, POCs, steps for testing, etc. opensearch-project/k-NN#812
The text was updated successfully, but these errors were encountered:
What do you want to do?
Tell us about your request. Provide a summary of the request and all versions that are affected.
As of today, k-NN plugin only supports vectors of type float for each dimension which is 4 bytes. This is getting expensive in terms of storage especially for those use cases that requires ingestion on a large scale as we need to construct, load, save and search graphs which gets more and more costlier. There are few use cases where customers prefer to reduce memory footprints by trading off recall with a minimal loss.
Using the Lucene ByteVector feature, we can add support for Byte Sized Vector where each dimension of the vector is a byte integer with range [-128 to 127].
As of now, we are planning to support this feature from 2.9 and future versions of OpenSearch.
What other resources are available? Provide links to related issues, POCs, steps for testing, etc.
opensearch-project/k-NN#812
The text was updated successfully, but these errors were encountered: