Dataset · IlyaGusev
Russian news with brief summaries
IlyaGusev/gazeta
A corpus of articles from the Gazeta.Ru website and their annotations for training and evaluating summarization models.
What's inside
The dataset contains 10 to 100 thousand news texts in Russian from the Gazeta.Ru website and their brief summaries. The annotations were extracted from the HTML description tag and underwent additional quality selection. There are two versions of this dataset.
What you can do with it
- Training and SFT of automatic text summarization models in Russian
- Evaluating summarization quality on a standard Russian benchmark
- Generating headlines and leads for news texts
- Tasks
- Summarization
- Languages
- Russian
- Size
- 10K–100K rows
- Downloads this month
- 4.6K
- Likes
- 31
Models trained on it
- deepvk/USER2-baseSentence similarity 1.1M
- deepvk/USER-bge-m3Sentence similarity 403.5K
- sergeyzh/rubert-tiny-turboSentence similarity 46.2K
- d0rj/rut5-base-summSummarization 29.9K
- deepvk/USER-baseSentence similarity 29.4K
- sergeyzh/rubert-tiny-liteSentence similarity 24.2K
- deepvk/USER2-smallSentence similarity 12.1K
- sergeyzh/rubert-mini-fridaSentence similarity 9.5K
How to load it
from datasets import load_dataset
ds = load_dataset("IlyaGusev/gazeta")