[FEA] Quantized Fastener index merge
Nobody has claimed this yet.
- Dominant language
- Cuda
- Stars
- 854
- Forks
- 236
- Avg merge
- 3d 3h
- Merged PRs (30d)
- 62
Description
Is your feature request related to a problem? Please describe.
In line with #2285, it would be good to support index merge on quantized vectors.
Describe the solution you'd like
The only non-plumbing thing that would need to change in the Fastener implementation is replacing the batched GEMM distance comparisons used to choose leaders and compute intra-partition NNs with quantized distance comparisons. This may be annoying to implement or slow if batched quantized comparisons in the shapes used by these two operations are not already available.
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
Start by reading the Fastener implementation and the related work in #2285. Identify the batched GEMM distance comparisons used for leader selection and intra-partition nearest neighbors, then check whether quantized comparisons support those shapes. Done means both operations use quantized distance comparisons without losing index-merge behavior.
Written by the indexing model from the issue text.
Assessment
- Domain
- machine-learning, performance, search
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 42/100