Dataset assetOpen Source CommunityVideo Retrieval

VideoRetrieval

The dataset includes three configurations: corpus, default, and queries. The corpus configuration contains document IDs, text, and titles, split into a dev partition with 100,930 samples and a total size of 8,580,491 bytes. The default configuration contains query IDs, document IDs, and scores, also in a dev split with 1,000 samples and 27,968 bytes. The queries configuration contains query IDs and text, in a dev split with 1,000 samples and 34,156 bytes.

Source

huggingface

Created

Nov 28, 2024

Updated

Dec 1, 2024

Signals

131 views

Availability

Linked source ready

Overview

Dataset description and usage context

Dataset Overview

Dataset Configurations

Configuration Name: corpus

Features:
- _id: string
- text: string
- title: string
Split:
- dev:
  - Bytes: 8,580,491
  - Samples: 100,930
Download Size: 7,277,662 bytes
Dataset Size: 8,580,491 bytes
Data Files:
- dev: corpus/dev-*

Configuration Name: default

Features:
- query-id: string
- corpus-id: string
- score: int64
Split:
- dev:
  - Bytes: 27,968
  - Samples: 1,000
Download Size: 17,445 bytes
Dataset Size: 27,968 bytes
Data Files:
- dev: data/dev-*

Configuration Name: queries

Features:
- _id: string
- text: string
Split:
- dev:
  - Bytes: 34,156
  - Samples: 1,000
Download Size: 29,116 bytes
Dataset Size: 34,156 bytes
Data Files:
- dev: queries/dev-*

Need downstream help?

Pair the dataset with AI analysis and content workflows.

Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.

Explore AI studio