Dialogue Test Set

该数据集是一个专门设计的测试集，包含1000个对话转录文本，旨在评估模型的性能。这些语音提示来源于Librispeech-test-clean，确保评估能够涵盖对话生成和独白生成两种能力。该数据集的规模为1000个对话转录，任务是对话生成模型的评估。

This is a specially designed test dataset containing 1,000 dialogue transcripts, with the goal of evaluating model performance. The speech prompts utilized in this dataset are sourced from Librispeech-test-clean, which ensures the evaluation encompasses both dialogue generation and monologue generation capabilities. With 1,000 dialogue transcripts, this dataset is specifically used for the evaluation of dialogue generation models.