[FEA] Reconcile the inlined vs separated cluster assignments in VPQ preprocessing
Open
Nobody has claimed this yet.
feature request
- Dominant language
- Cuda
- Stars
- 854
- Forks
- 236
- Avg merge
- 3d 3h
- Merged PRs (30d)
- 62
Description
The VPQ code inlines the VQ labels into the encoded output.
For compatibility reasons an option was added into the VPQ code to support it: inline_vq_labels.
Both codepaths should be reconciled to avoid that option.
As an first idea for that the solution could involve using a strided view for the pq_codes
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
No files or tests are named. Locate the VPQ preprocessing entry point and both codepaths using inline_vq_labels, then compare how cluster assignments are laid out in the encoded output. Done means the codepaths are reconciled, the compatibility option is no longer needed, and both layouts remain compatible.
Written by the indexing model from the issue text.
Assessment
- Domain
- data, search
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100