zilliztech / zilliztech/VectorDBBench

Questions regarding custom datasets and recall calculation

Open
#528 4 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
1.2k
Forks
438
Avg merge
1d 4h
Merged PRs (30d)
1

Description

Thank you for your work on this project — it's a valuable tool for evaluating vector databases.

I have a couple of technical questions:

  • I’m interested in benchmarking with a custom scale/dimensions dataset (e.g., 100M vectors with 1024 dimensions). In this case, is it recommended that I generate a custom dataset manually, or is there an existing mechanism within VectorDBBench to support such configurations?

  • Regarding the 15 built-in datasets, could you please provide more details on how the ground truth is constructed? Specifically, I’d like to understand the methodology used to compute recall — for example, how the nearest neighbors are determined and what metrics or tools are used to establish the ground truth.

I would greatly appreciate any clarification or documentation you could provide.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No files, tests, or entry points are named. Review the project’s dataset configuration and recall or ground-truth documentation, then document how custom scale and dimensions are supported and how the 15 built-in datasets establish nearest neighbors and recall.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
databases, search
Issue type
Documentation
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.