whybe-choi/OmniWikiRetrieval
OmniWikiV2IRetrieval An MTEB dataset Massive Text Embedding Benchmark Retrieve the Wikipedia page screenshot on which a given video is embedded. Each query is a video embedded on an English Wikipedia page (drawn from Tevatron Wiki-SS-NQ train positives, kept only if 10-180s and decodable by torchcodec), and the gold documents are the page-screenshot images of every corpus page the video appears on. The corpus is those gold pages plus the BM25 hard negatives of the questions whose… See the full description on the dataset page: https://huggingface.co/datasets/whybe-choi/OmniWikiRetrieval.
| Author / organization | whybe-choi |
|---|---|
| License (as declared) | cc-by-sa-4.0 |
| Downloads | 212 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 8 September 2026 |
| Last updated at source | 8 September 2026 |
Popularity, last 3 days
212 now (-60 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
More from whybe-choi
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.