Skip to content

Dataset · daekeun-ml

Korean news with summaries

daekeun-ml/naver-news-summarization-ko

27,400 news text and summary pairs in Korean for summarization model training and testing.

What's inside

The dataset consists of document-summary pairs in Korean. 27,400 pairs were collected from Naver News over ten days in July 2022. Originally created for a Korean NLP workshop; includes a technical report with collection protocols and baseline results.

What you can do with it

  • Fine-tuning Korean news summarization models
  • Testing seq2seq models in Korean
  • LoRA for news summarization
Tasks
Summarization
Languages
Korean
Size
10K–100K rows
Downloads this month
909
Likes
70

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("daekeun-ml/naver-news-summarization-ko")