Nan-Do/OpenSubtitlesJapanese

--- task_categories: - text-generation language: - ja tags: - text - tv - movies pretty_name: OpenSubtitles dataset in Japanese size_categories: - 100M<n<1B --- The dataset contains (almost) the entire OpenSubtittles database for Japanese: - Over 7000 tv shows and/or movies. - The subtittles are human generated. - The dataset has been parsed, cleaned and converted to UTF-8. File contents: - OpenSubtitles.parquet: The text and the time data. - OpenSubtitles_meta.parquet: The existing metadata for each title. - OpenSubtitles-OA.parquet: The dataset coded with two columns SOURCE(the name of the movie/tv show), and TEXT (the subtittles) following the Open Assistant rules. Both tables can be joined by the ID column. (The value can be NULL in the meta table).

原始信息汇总

数据集概述

基本信息

• 任务类别：文本生成
• 语言：日语
• 标签：文本、电视、电影
• 美观名称：日语版OpenSubtitles数据集
• 大小类别：1亿<n<10亿

数据集内容

• 包含内容：
  • 超过7000部电视节目和/或电影的日语字幕
  • 字幕为人为生成
  • 数据已解析、清洗并转换为UTF-8编码

文件详情

• OpenSubtitles.parquet：包含文本和时间数据
• OpenSubtitles_meta.parquet：包含每个标题的现有元数据
• OpenSubtitles-OA.parquet：按照Open Assistant规则编码的数据集，包含两列：SOURCE（电影/电视节目名称）和TEXT（字幕）

数据关联

• 关联方式：通过ID列将两个表连接，元数据表中的ID值可能为NULL