Binary Quantization Support for pgvector HNSW Algorithm

zilliztech / VectorDBBench

A Benchmark Tool for VectorDB

MIT License

558 stars 151 forks source link

Binary Quantization Support for pgvector HNSW Algorithm #389

Closed Sheharyar570 closed 3 weeks ago

Sheharyar570 commented 3 weeks ago

This update introduces binary vector quantization support to the pgvector HNSW algorithm within our benchmarking tool. The implementation includes binary vector quantization queries with an option for reranking.

New CLI and UI Options

reranking (boolean): Enables or disables reranking for quantized queries.
quantized_fetch_limit (integer): Specifies the number of rows to fetch for distance calculations between binary vectors before reranking.
reranking_metric (string): Chooses the distance metric for reranking, with options:
- COSINE
- L2
- IP

Sheharyar570 commented 3 weeks ago

/assign @xuanyang-cn

greenhal commented 3 weeks ago

This is a great addition to VectorDBBench. My only suggestion is would be to try to reduce the large section of duplicate code in pgvector with the filtered/unfiltered sql query definition. Lines 132-194 & 195-245, this code is identical except for the where clause.

Sheharyar570 commented 3 weeks ago

@greenhal I've refactored the code as you suggested and removed code duplication.

sre-ci-robot commented 3 weeks ago

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: alwayslove2013, Sheharyar570 To complete the pull request process, please assign xuanyang-cn after the PR has been reviewed. You can assign the PR to them by writing /assign @xuanyang-cn in a comment when ready.

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these files: - **[OWNERS](https://github.com/zilliztech/VectorDBBench/blob/main/OWNERS)** Approvers can indicate their approval by writing `/approve` in a comment Approvers can cancel approval by writing `/approve cancel` in a comment