File size: 540 Bytes
5ba1f52 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 | name: SwednClusteringP2P
description: The SWE-DN corpus is based on 1,963,576 news articles from the Swedish
newspaper Dagens Nyheter (DN) during the years 2000--2020. The articles are filtered
to resemble the CNN/DailyMail dataset both regarding textual structure. This dataset
uses the category labels as clusters.
evaluation_framework: mteb
tasks:
- id: SwednClusteringP2P
- id: SwednClusteringP2P_default_summaries
config: default
split: summaries
- id: SwednClusteringP2P_default_articles
config: default
split: articles
|