CleanS2S

该数据集为类似于人类对话的语音到语音交互提供了一个框架，它将自动语音识别、大型语言模型和文本到语音合成技术集成到一个统一的流程中。该框架支持多种模型，确保了研究想法的高效原型设计和快速迭代，同时保持了模块化和可复现性。其任务目标是实现语音到语音的交互。

This dataset provides a framework for human-like conversational speech-to-speech interaction, which integrates Automatic Speech Recognition (ASR), Large Language Models (LLMs), and Text-to-Speech (TTS) synthesis technologies into a unified workflow. This framework supports multiple models, enabling efficient prototyping and rapid iteration of research ideas while maintaining modularity and reproducibility. Its task objective is to enable speech-to-speech interaction.