Heemin Kim
Heemin Kim
I can still see the same problem with the 'develop' branch.
@mdouze , would you mind taking a look?
I search HNSW through IndexIDMap. Do you have a plan to expand the ResultHandler to IndexIDMap? I need a way to access `id_map` data to dedupe result properly.
@mdouze Any further thought on this? As @navneet1v mentioned, we can add something like IDDeduper(or IDGroup) and pass it through search parameter. During the result collection, we can dedupe the...
I think we can support the feature incrementally starting hnsw index. For all other index, we can throw an error message like what we do for search parameter. `FAISS_THROW_IF_NOT_MSG(!params, "search...
@mdouze I raised a draft PR for vector deduplication in HNSW. https://github.com/facebookresearch/faiss/pull/3140 I am going to continue to work to make the PR to meet the bar to be merged...
@mdouze I added a new PR which introduce a result collector so that caller can implement deduplication logic outside of faiss repo. Could you take a look and provide a...
After the refactoring around ResultHandler, https://github.com/facebookresearch/faiss/pull/3190, I came up with new plan to support grouping of result. 1. Create IDGrouper similar to IDSelector. Add two grouper to start with: Bitmap,...
Same for both Faiss and Lucene. Wasn't able to find the why.
@ryanbogan?