sdialog/200-dialogues-voices-libritts

这是一个包含200个对话的数据集，这些对话是通过SDialog生成的。在对话中，医生和患者两个角色都使用了gemma3:27b-it-qat语言模型，但医生的话会被截断问号。数据集被划分为训练集、开发集和测试集，分别对应不同的对话范围。同时，数据集中的音频是通过LibriTTS声音库、IndexTTS模型、dScaper和PyRoomAcoustics生成的。

This dataset consists of 200 dialogues generated using SDialog. Both roles in the dialogues use the gemma3:27b-it-qat language model, but the doctors speech is truncated with a question mark. The dataset is split into training, development, and test sets, each covering different ranges of dialogues. The audio in the dataset is generated using the LibriTTS voice bank, IndexTTS model, dScaper, and PyRoomAcoustics.