Audio Generation Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/facebookresearch/audiocraft
下载链接
链接失效反馈官方服务：
资源简介：
该数据集用于训练文本到音频生成模型的样本，最初由Kreuk等人（2022a）采用。该数据集包含了从完整曲目中随机抽取的30秒音频片段。所涉及的任务是文本到音频的生成。
This dataset, used as training samples for text-to-audio generation models, was originally adopted by Kreuk et al. (2022a). It contains 30-second audio clips randomly sampled from full tracks. The task involved is text-to-audio generation.
提供机构：
Kreuk et al.
