license: gpl-3.0
task_categories:
- text-generation
language: - zh
tags: - not-for-all-audiences
- hentai
- nsfw
- literature
- novel
- roleplay
- uncensored
- mutt
- zhtw
- taiwan
size_categories: - 1K<n<10K
What this dataset does
A whole dataset of Chinese h-novels translated into Traditional Chinese(Taiwan) forked from
Sex-novel-filtered and sex-novel
Motivation
There aren't a lot of Traditional Chinese(Taiwan) datasets on huggingface and it pretty much
ruins the mood when the AI spits out Simplified Chinese to people from Taiwan(especially those
who uses qwen3 as the base training model)
How this works
It's pretty much well known that Taiwan uses a different variant of Chinese, different from what mainland
China uses. So, I've written a shellscript called replace.sh. It uses opencc to first translate
it into Traditional Chinese and then replaces words that are used in China with those equivilant to
what people in Taiwan says. For example, 「早上好」or good morning in China has the same meaning
as「早安」in Taiwan. To preserve the original efforts of the author, I've kept the original Simplified Chinese as is and a suffix 'TW' in the filename for the Taiwanese version.
Developer Note:
The count.sh file counts through how many Chinese characters there are in total so that
you can estimate how long it'll take for you to fully fine-tune your model!
Feel free to give me pull requests for this project to produce more high-quality datasets!
Special Thanks:
Thanks to Ruri_Helesta for working this project with me! He's the one
who fine-tunes the model with these datasets for me to test!