Datasets:
license: cc-by-4.0
dataset_info:
features:
- name: folder
dtype: string
- name: title
dtype: string
- name: text
dtype: string
- name: title_len
dtype: int64
- name: text_len
dtype: int64
splits:
- name: train
num_bytes: 1126829322
num_examples: 512750
download_size: 590025488
dataset_size: 1126829322
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
This is an Uzbek News Dataset with 512,750 articles (120 million words and in the Latin script) scraped from the web in 2023.
I combined and uploaded the dataset in this HF repo so that the community can fine-tune LLMs based on the Uzbek language.
@proceedings{kuriyozov_elmurod_2023_7677431,
title = {{Text classification dataset and analysis for Uzbek
language}},
year = 2023,
publisher = {Zenodo},
month = feb,
doi = {10.5281/zenodo.7677431},
url = {https://doi.org/10.5281/zenodo.7677431}
}
Github: github Arxiv: https://arxiv.org/ftp/arxiv/papers/2302/2302.14494.pdf
Original dataset was created by scraping the Internet.
From the authors:
The dataset consists text data from 9 Uzbek news websites and press portals that included news articles and press releases. These websites were selected to cover various categories such as politics, sports, entertainment, technology, and others. In total, we collected 512,750 articles with over 120 million words accross 15 distinct categories, which provides a large and diverse corpus for text classification. It is worth noting that all the text in the corpus is written in the Latin script.