A 1,000‑hour Japanese singing voice dataset (~73 GB) in WebDataset format for training SVC models, containing 15,311 MP3 files with rich YouTube metadata and quality scores.
Description in other languages
Statistics
Statistics as of Nov 16, 2026
Hugging Face downloads (last 30 days)
458
Hugging Face likes
2
Listed
Nov 17, 2026
Popular in the same category
- voicevox-voice-corpusayousanz885VOICEVOX‑generated synthetic voice dataset comprising 445,793 .wav files (totaling 577 h 51 m 23 s) built from the ITA, つくよみちゃん, and ROHAN corpora.
- makise-kurisu-vn-voicelineszhonglongbao847Transcribed Makise Kurisu VN dialogue from a video using Whisper Large‑V2, split into clips with pydub; uncleaned text is intended for TTS model training and does not belong to the author.
- spoken-magpie-jaAtotti648Japanese instruction‑tuning dataset for the Spoken‑magpie LLM, built with CosyVoice2 TTS on 10 NVIDIA H200 GPUs, featuring diverse speaker voices and audio responses for prompts under 500 characters, with 50 sample entries publicly available.
- nagatoro-soundezfiez424A 140‑clip Japanese voice dataset of Nagatoro Hayase from “Nagatoro‑san” for AI training in RVC, So‑VITS‑SVC, TTS fine‑tuning and audio autoencoders.
- Hadou-Voice-Datasethadou1225419A Japanese-language voice dataset of 966 “Calm Voice” clips (~114 min) and 424 “ITA Corpus” clips (~39 min) by Hadou, usable for AI training, TTS, voice conversion, LoRA creation, or direct inclusion in any project with optional credit.
- Lux-Japanese-Speech-CorpusLami332Lux Japanese Speech Corpus is a CC‑BY‑4.0 dataset of 96 kHz/16‑bit WAV recordings by the character Lux, providing raw and cleaned audio files with transcriptions in metadata.csv for use in TTS research and training.
Resources in similar research fields
aica-corpusreinehonoka23AI character, filler, laughter, and emotion-focused Japanese TTS corpus (CC0)GitHubShared: Speech synthesis・Corpora & datasets- joyo-kanji-yomi-benchmark-parakeetParakeet-Inc272A benchmark dataset and evaluation toolkit for measuring how accurately Japanese G2P, morphological analyzers, and TTS systems read kanji from the Ministry of Culture’s Jōyō‑kanji list, featuring 4,512 kanji‑reading pairs with three annotated example sentences each and detailed annotations for multiple readings.Shared: Speech synthesis・Corpora & datasets
- joyo-kanji-yomi-benchmarksbintuitions224A Japanese TTS benchmark that evaluates kanji pronunciation across all 2,136 Joyo kanji with 4,378 readings using 13,095 native‑speaker‑verified sentences, each targeting a single reading and annotated for automatic CER calculation under an MIT license.Shared: Speech synthesis・Corpora & datasets
- KakologArchivesKakologArchives74kAggregates over 150 GB of NicoNico Live’s comment logs from 2009‑2024—including pre‑transition, post‑transition, and real‑time NX‑Jikkyo captures—providing an API for easy retrieval of historical TV‑broadcast discussions.Shared: Corpora & datasets
GPT-SoVITSRVC-Boss62k1 min voice data can also be used to train a good TTS model! (few shot voice cloning)MultilingualShared: Speech synthesis
barksuno-ai39k🔊 Text-Prompted Generative Audio Model.MultilingualShared: Speech synthesis