Skip to content

Dataset · fancyzhx

DBpedia 14-category text classification dataset

fancyzhx/dbpedia_14

Large balanced dataset of 630K English texts for multi-class classification.

What's inside

Data is sourced from the DBpedia 2014 ontology and divided into 14 mutually exclusive classes. The training split contains 560K examples (40K per class), and the test split contains 70K (5K per class). Both parts are in English and share a three-column structure.

What you can do with it

  • Training and benchmarking text classifiers (BERT, RoBERTa, and similar encoders)
  • Evaluating text embeddings on thematic clustering tasks
Tasks
Text classification
Languages
English
Size
100K–1M rows
Downloads this month
21.9K
Likes
37

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("fancyzhx/dbpedia_14")