| 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 |
|
|
