---
task_categories:
- text-generation
language:
- av
pretty_name: avarCorpora
size_categories:
- 10K<n<100K
---
This is a dataset of pure (unlabeled) texts in the Avar language (av: магIарул (avar) мацI; ru: аварский язык). The volume of its data is 4,455,123 tokens (in 63,973 texts).

The texts were collected from open sources, the list of sources is listed in [this document](https://drive.google.com/file/d/1LCT3O6pNkx1RRic1carlJNiu7mqhindl/view?usp=sharing) (in Russian). Mostly the dataset contains data from the press, internet mass media and other public resources (such as Wikipedia), personal blogs and comments. There are no poetry or religious texts in the dataset (or at least I tried to exclude them).

You can use this dataset for any purpose (it would be better if you provide a link to this repository). If you cannot open [the document with the sources](https://drive.google.com/file/d/1LCT3O6pNkx1RRic1carlJNiu7mqhindl/view?usp=sharing) or have any questions, contact me by e-mail: volina092@ya.ru