PATS

该数据集名为PATS，包含了多种不同说话者的视频资料，它被广泛应用于零样本风格转移的手势动画领域。此外，该数据集还包括了在模型训练期间既可见也未见过的说话者的录音。该数据集的任务是利用多模态输入（文本和语音）来驱动手势动画。

This dataset, named PATS, contains video recordings of multiple distinct speakers, and it is widely used in the field of gesture animation for zero-shot style transfer. Additionally, this dataset also includes audio recordings of speakers that are either seen or unseen during model training. The core task of this dataset is to drive gesture animation using multimodal inputs including text and speech.