Skip to content

Dataset · NandemoGHS

Japanese voice data from visual novels with transcriptions

NandemoGHS/Japanese-Eroge-Voice

A large corpus of Japanese audio recordings from games and automatically generated subtitles for TTS and ASR.

What's inside

The dataset contains from 100 thousand to 1 million pairs of Japanese audio and text. The audio tracks were extracted from licensed Japanese adult visual novels (eroge) and volume-normalized via ffmpeg. The text transcriptions for the recordings were automatically generated using the litagin/anime-whisper model.

What you can do with it

  • Training automatic speech recognition (ASR) systems for Japanese language and anime styling
  • Training text-to-speech (TTS) models with emotional game voice acting
Tasks
Text → speech, Automatic speech recognition
Languages
Japanese
Size
100K–1M rows
Downloads this month
317
Likes
38

How to load it

from datasets import load_dataset

ds = load_dataset("NandemoGHS/Japanese-Eroge-Voice")