TeoGchx/BEAT_HML3D_whisper_wavtokenizer_smpl_motokenizer

该数据集包含了运动数据（包括贝塔值、表情、性别、动作捕捉帧率、模型、姿态和变换）、文本、元数据（包括持续时间、名称、帧数、单词数和发言人）、耳语（包括语言、片段的起始和结束时间以及文本）、对齐文本（包括片段和单词的起始和结束时间、得分和单词本身）、音频数据（采样率为24000Hz）和音频标记、运动标记。数据集分为验证集、训练集和测试集，分别包含不同数量的示例和大小。

The dataset includes motion data (including betas, expressions, gender, mocap frame rate, model, poses, and transformations), text, metadata (including duration, name, number of frames, number of words, and speaker), whisper (including language, start and end times of segments, and text), aligned text (including start and end times, scores, and words of segments and word segments), audio data (with a sampling rate of 24000 Hz), and audio and motion tokens. The dataset is split into validation, training, and test sets, each containing a different number of examples and size.