Dataset Viewer
The dataset viewer is not available for this subset.
The de config contains 329 while it should generally contain 3 splits maximum (train/validation/test). If the splits 2004_q4_12, 2004, 2004_q4, 2005_q1_01, 2005... are not used to differentiate between training and evaluation, please consider defining configs of this dataset instead. You can find how to define configs instead of splits here: https://huggingface.co/docs/hub/datasets-data-files-configuration
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Wikinews
The dataset contains news articles from Wikinews in different languages. Each article is associated with metadata like title, url, and date. The articles grouped into data splits by the article month, quarter, and year (the date is one mentioned in the article text, changes might have been after, see revision timestamp). The dataset config name defines the language.
Usage
from datasets import load_dataset
# all English news from 2008
ds = load_dataset("malteos/wikinews", config_name="en", split="2008")
# all German news from January 2017
ds = load_dataset("malteos/wikinews", config_name="de", split="2017_q1_01")
Languages
- en
- es
- fr
- it
- de
License
All text created after September 25, 2005 available under the terms of the Creative Commons Attribution 2.5 License.
- Downloads last month
- 2,029