[FEA] Expose fp16 and uint8 through FAISS for all index types
Open
Nobody has claimed this yet.
feature request
- Dominant language
- Cuda
- Stars
- 854
- Forks
- 236
- Avg merge
- 3d 3h
- Merged PRs (30d)
- 62
Description
This is a placeholder / tracker issue. It's becoming more common that vectors are heavily quantized using scalar or binary quantization before they are handed off to the indexes for training. As a result, it's important that we support uint8 and fp16 through our integrations, especially FAISS.
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
This is a placeholder tracker with no files, tests, or entry points named. Start by mapping the FAISS integration and the supported index types, then determine how fp16 and uint8 inputs are represented across training and search. Done means the agreed index types accept both formats with coverage for their integration paths.
Written by the indexing model from the issue text.
Assessment
- Domain
- search
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100