SST2-audio
收藏arXiv2025-09-30 收录
下载链接：
https://docs.coqui.ai/en/latest/models/xtts.html
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是对SST2在语音领域的扩展，它保留了情感极性，同时尽量减少了说话人和环境变化的影响。这些数据是使用可控的文本转语音模型（XTTSv2）生成的，旨在用于分析语音模型中的语义概念。
This dataset is an extension of the SST2 to the speech domain, which preserves sentiment polarity while minimizing the impact of variations among speakers and environmental changes. It was generated using a controllable text-to-speech model (XTTSv2), and is designed for the analysis of semantic concepts in speech models.
提供机构：
Coqui AI
