TeoGchx/BEAT_HML3D_whisper_wavtokenizer_motokenizer_emage_all_spk

该数据集包含了运动的浮点数序列、文本序列、元数据（包括时长、名称、帧数、词汇数、说话者等）、音频数据以及其他与运动相关的结构化数据（包括贝塔值、表情、性别、动作捕捉帧率、模型、姿态、平移等）。此外，还有与文本对齐的信息和语言相关的数据段。数据集划分为训练集、验证集和测试集，大小分别为8.46GB、6.77GB和1.64GB。

The dataset includes sequences of motion as float32, text sequences, metadata (including duration, name, number of frames, number of words, speaker, etc.), audio data, and other structured data related to motion (including betas, expressions, gender, mocap frame rate, model, poses, translation, etc.). Additionally, there are aligned text information and language-related data segments. The dataset is split into training, validation, and test sets, with sizes of 8.46GB, 6.77GB, and 1.64GB respectively.