Batuka0901/mn-tts-omnivoice-data
收藏Hugging Face2026-04-29 更新2026-05-03 收录
下载链接：
https://hf-mirror.com/datasets/Batuka0901/mn-tts-omnivoice-data
下载链接
链接失效反馈官方服务：
资源简介：
这是一个蒙古语单说话者文本到语音(TTS)数据集。数据集使用说话者25秒的原始参考声音，通过OmniVoice语音克隆模型增强和克隆该声音，并从Google FLEURS (mn)数据集中筛选出包含特定字符的文本。每个文本仅由1个声音朗读，形成音频+文本数据集。包含895个句子/wav文件，使用蒙古语(西里尔字母)，采样率为24 kHz。音频生成方法为OmniVoice克隆+参考音频文件。
A Mongolian single-speaker text-to-speech (TTS) dataset. The dataset uses the speakers original 25-second reference voice, enhances and clones the voice using the OmniVoice voice-cloning model, and filters texts containing specific characters from Google FLEURS (mn). Each text is read by only 1 voice, forming an audio + text dataset. Contains 895 sentences/wav files, uses Mongolian (Cyrillic), with a sampling rate of 24 kHz. Audio generation method is OmniVoice clone + reference audio file.
提供机构：
Batuka0901
