mikemccand commented on issue #14007: URL: https://github.com/apache/lucene/issues/14007#issuecomment-2494484887
Also, note that, at least for the current scalar quantization (`int7`, `int4`), those full precision `float[]` vectors remain on disk during searching. They are only used during indexing (merging) to recompute the quantiles and re-quantize the vectors more accurately (possilby) for that newly merged segment. For apps using near-real-time segment replication (the best/preferred replication approach IMO), it may be possible eventually to strip out the `float[]` from the Lucene index sent to searchers, but this has only been discussed so far -- no implementation -- I'll try to track down the issue (@benwtrent also referred to this above). -- 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
