benwtrent commented on issue #13699: URL: https://github.com/apache/lucene/issues/13699#issuecomment-2374932998
> So the workaround for users now is to just ask for a larger top N and discard all but the top K in the end? Yes, that is the current solution. > +1 it'd be nice to find some simple way to allow "expert" tunables to be set rather than one at a time poking them through the APIs ad-hoc. Though, lacking such a general mechanism also shouldn't block progress......But let's design for today? I agree, just wanting to clarify that the expert settings would be tied to a particular index format. Users would need to know the format they are using when utilizing the parameters. Additionally, there are some other more advanced things you can do with HNSW (early exit based on other criteria) and quantization (dynamically gather more values based on some error bias). As for how to set these tunables, its possible now to create a sub-class of the knn queries and override `getKnnCollectorManager`. Maybe `MultiLeafKnnCollector` just needs to be updated to allow more inputs (greediness is static and cannot be set via the ctor). That way folks can pretty easily sub-class the knn queries and return a customized multi-leaf collector. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
