Navneet Verma

@mdouze is there any faiss mailing list on which we can send out this feature and see if there are users interested in this feature? Reason why I saying this...

@heemin32 were you able to figure out why the query with filter is not returning inner hits? Plus is this consistent with Faiss engine too?

@heemin32 its pretty interesting that inner hits is working for standard search and not for filtered vector search even when we run almost same code in both the places.

One more thing I would like to see in this reach. would be we tracking the native memory usage during indexing and training steps too. I understand that indexing is...

> @navneet1v we actually do track memory during training https://github.com/opensearch-project/k-NN/blob/main/src/main/java/org/opensearch/knn/training/TrainingJob.java#L120-L142, but yes this will include tracking during indexing too Thanks for clarification. Lets track memory during indexing.

> > High Level Comments: > > > > 1. Add examples of radial search with filters. > > 2. Add examples of radial search with neural query clause. >...

@junqiu-lei for adding examples of neural search you should use this page: https://opensearch.org/docs/latest/query-dsl/specialized/neural/

We can move to a node level CB but removing this CB can this lead to OOM killer on some nodes? because they ran out of memory?

> @navneet1v idea is that individual nodes would reject requests instead of a node rejecting the request because some other node is under duress. This can lead to partial success...

@luyuncheng not storing vector fields in _source is always an option for user. But this comes with its own limitations: 1. If you remove the fields from _soruce then you...