bc7ec356/synthetic-speech-indic
收藏资源简介：
这是一个多语言语音数据集，包含31种语言（如阿萨姆语、阿瓦德语、孟加拉语、博杰普尔语、博多语、多格拉语、英语、加罗语、贡德语、古吉拉特语、印地语、霍语、卡纳达语、克什米尔语、卡西语、孔德语、科达瓦语、孔卡尼语、马加希语、迈蒂利语、马拉雅拉姆语、曼尼普尔语、马拉地语、米佐语、蒙达里语、旁遮普语、拉贾斯坦语、梵语、桑塔利语、信德语、索拉语和泰米尔语）的音频和文本数据。每个样本包括字符、单词、句子的文本转录，以及语言信息、说话者姓名和性别。音频数据以16kHz采样率提供。数据集分为训练集和验证集，适用于语音识别、自然语言处理或多语言研究任务。
This is a multilingual speech dataset comprising 31 languages (e.g., Assamese, Awadhi, Bengali, Bhojpuri, Bodo, Dogri, English, Garo, Gondi, Gujarati, Hindi, Ho, Kannada, Kashmiri, Khasi, Khond, Kodava, Konkani, Magahi, Maithili, Malayalam, Manipuri, Marathi, Mizo, Mundari, Punjabi, Rajasthani, Sanskrit, Santali, Sindhi, Sora, and Tamil). Each sample includes text transcriptions at character, word, and sentence levels, along with language information, speaker name, and gender. Audio data is provided at a 16kHz sampling rate. The dataset is split into training and validation sets, suitable for speech recognition, natural language processing, or multilingual research tasks.
