NVIDIA / NVIDIA/cuvs

[FEA] Allow config.yaml used to describe dataset to also be used as input in cuvs_bench.generate_groundtruth

Open
#2,037 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

feature request
Dominant language
Cuda
Stars
854
Forks
236
Avg merge
3d 3h
Merged PRs (30d)
62

Description

Users should be allowed to provide the config.yaml used to describe their dataset as an input to cuvs_bench.generate_groundtruth when calculating ground truth. Currently, each field needs to be passed individually.

Sample config.yaml:

- name: BeIR-trec-news-generated-queries-1.5M
  dims: 768
  base_file: BeIR-trec-news-generated-queries-1.5M/base.fbin
  query_file: BeIR-trec-news-generated-queries-1.5M/query.fbin
  groundtruth_neighbors_file: BeIR-trec-news-generated-queries-1.5M/groundtruth.neighbors.ibin
  distance: euclidean

Contributor guide

Open the contributing guide

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

Start at the cuvs_bench.generate_groundtruth entry point and trace how the individual dataset fields are currently accepted. Use the sample config.yaml as the input case; the work is done when that file can provide the dataset description for ground-truth calculation without requiring every field separately.

Written by the indexing model from the issue text.

Assessment

Domain
cli
Issue type
Feature
Difficulty
3/5
Estimated time
1-2 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
63/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.