Reubencf/Adaption-multilingual-speech

该数据集是一个多语言语音数据集，主要用于文本到语音（TTS）和自动语音识别（ASR）任务。数据集是从原始68,677条多语言合成语音语料库中选取的10,274条音频和文本行，经过Adaption的Adaptive Data平台处理，增加了`enhanced_prompt`和`enhanced_completion`列。数据集包含9种语言（日语、俄语、德语、韩语、西班牙语、葡萄牙语、中文、英语、法语），每种语言的行数和比例在README中有详细说明。数据集还包含了音频、文本、语言、风格、声音等元数据。

This dataset is a multilingual speech dataset primarily intended for text-to-speech (TTS) and automatic speech recognition (ASR) tasks. It consists of 10,274 audio and text rows selected from the original 68,677-clip multilingual synthetic speech corpus, processed through Adaptions Adaptive Data platform, with added `enhanced_prompt` and `enhanced_completion` columns. The dataset covers 9 languages (Japanese, Russian, German, Korean, Spanish, Portuguese, Mandarin Chinese, English, French), with detailed breakdowns of row counts and proportions for each language provided in the README. The dataset also includes metadata such as audio, text, language, style, and voice.