Datasets:

Modalities:
Text
Formats:
parquet
Languages:
Swedish
ArXiv:
License:
Samoed commited on
Commit
5ba1f52
·
verified ·
1 Parent(s): f8dbf10

Add eval config

Browse files
Files changed (1) hide show
  1. eval.yaml +14 -0
eval.yaml ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ name: SwednClusteringP2P
2
+ description: The SWE-DN corpus is based on 1,963,576 news articles from the Swedish
3
+ newspaper Dagens Nyheter (DN) during the years 2000--2020. The articles are filtered
4
+ to resemble the CNN/DailyMail dataset both regarding textual structure. This dataset
5
+ uses the category labels as clusters.
6
+ evaluation_framework: mteb
7
+ tasks:
8
+ - id: SwednClusteringP2P
9
+ - id: SwednClusteringP2P_default_summaries
10
+ config: default
11
+ split: summaries
12
+ - id: SwednClusteringP2P_default_articles
13
+ config: default
14
+ split: articles