README.md

metadata

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 and cleaned.

File contents:

• OpenSubtitles-2023-04-18.parquet: The text and the time data.

• OpenSubtitles_meta-2023-04-18.parquet: The existing metadata for each title.

  Both tables can be joined by the ID column. (The value can be NULL in the meta table).