jtibshirani commented on pull request #267: URL: https://github.com/apache/lucene/pull/267#issuecomment-910812274
> I took an inspiration from the `nmslib` library where this parameter [is configurable](https://github.com/nmslib/nmslib/blob/master/similarity_search/src/method/hnsw.cc#L203) Thanks, I didn't poke through the C++ and had missed that. It still might not be common to set it, it looks like the HNSW implementations in ann-benchmarks don't tune over `ml`. I don't feel strongly, but we usually try to minimize configuration options, only exposing ones that are really useful and we expect users to tune. Maybe we could at least prevent `ml` from being `0` so we could remove the flat graph code path? This would be a nice simplification and I can't think of a use case for keeping a flat graph implementation as part of the HNSW codec. > This test fails because it marks many expected results docs (from 0 to 9) as deleted which don't end up being retrieved. That makes sense. This feels like an independent fix, maybe we could pull it into its own PR against `main` so it doesn't just sit on the feature branch? -- 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
