Datasets:

Modalities:
Text
Formats:
webdataset
Size:
< 1K
Libraries:
Datasets
WebDataset
License:
MAC / README.md
shuailiu6626's picture
Update README.md
08e4220 verified
|
raw
history blame
462 Bytes
metadata
license: cc-by-sa-4.0

We introduce the Million Authors Corpus (MAC), a novel dataset encompassing contributions from dozens of languages on Wikipedia. It includes only long and contiguous textual chunks taken from Wikipedia edits and links those texts to their authors.

This repository includes the data used for training and testing only. The full MAC dataset is shared in Zenodo repository: https://zenodo.org/records/15538126

Here are few examples: