license: cc0-1.0
task_categories:
- question-answering
language:
- ru
pretty_name: otvet.mail.ru questions
size_categories:
- 100M<n<1B
Dataset Card for otvet.mail.ru questions
Table of Contents
Dataset Description
Dataset Summary
This is a dataset of questions and answers scraped from otvet.mail.ru. There are about 130 million questions with all their corresponding metadata that were posted before 03/05/2022 (the date the dataset was collected). This is a reupload of my dataset on Kaggle
Languages
The dataset is mostly in Russian, but there may be other languages present.
Dataset Structure
Please refer to the Dataset Viewer for more information on the dataset structure.
For now the Dataset Viewer doesn't work because of inconsistent data type across samples. I'll try to fix it later, but for now, the dataset could be used by downloading ZSTD compressed chunks, each consisting of 2_500_000 samples.
Dataset Creation
The data was scraped using AJAX endpoints that return full question and answers metadata by id that is auto-incremented.