sleeping-ai/11Labs

11labs数据集是由11Labs API编译而成的，旨在促进音频深度学习和扩散模型领域的研究和创新。该数据集包含大量独特的人声和情感，每个语音片段都有不同的发声特点，与其他类似类型的数据集相比具有很高的独特性。数据集提供了增强版和超级分辨率版的wav文件、csv格式的元数据和说话人嵌入。共有25K个语音片段，每个片段时长为8-11秒，总时长为55.6小时。数据集的版权属于Sleeping AI和友好实验室，目前仅提供学术和研究用途，未提供商业使用权利。

Introducing 11labs, a dataset compiled from the 11Labs API, aimed at fostering research and innovation in the field of audio deep learning and diffusion models. The dataset includes a large number of unique human voices and emotions, with each voice snippet having a distinct set of vocal bursts, setting it apart from other similar datasets. It provides enhanced and superRES wav files, metadata in csv format, and speaker embeddings. There are 25K voice snippets, each 8-11 seconds long, totaling 55.6 hours of audio data. The dataset is copyrighted by Sleeping AI and friendly labs, and is currently available for academic and research purposes only, with no commercial usage rights provided.