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