Skip to content

Dataset · fancyzhx

AG News: a million news articles for classification

fancyzhx/ag_news

Over a million news articles from 2,000+ sources with topic labeling for text classification.

What's inside

A collection of over a million news articles gathered by the ComeToMyHead academic search engine since 2004 from more than 2,000 sources. Written in English, the scale ranges from 100 thousand to 1 million examples for topic classification. The data is provided for research in data mining and information retrieval.

What you can do with it

  • Train and test a news classifier
  • Build a quick baseline for encoder comparison
  • Fine-tune LoRA for topic categorization
Tasks
Text classification
Languages
English
Size
100K–1M rows
Downloads this month
114.9K
Likes
218

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("fancyzhx/ag_news")