minato-ryan/starrail-voice-xcodec2-example

该数据集包含不同长度的音频片段，每个片段都有相应的游戏内文件名、转录文本、语言、说话者、声音类型和持续时间信息。数据集分为三种配置：长音频（long）、短音频（short）和非常长音频（very_long），每种配置都有对应的训练集。

The dataset contains audio clips of varying lengths, each with corresponding in-game filenames, transcriptions, languages, speakers, voice types, and duration information. The dataset is divided into three configurations: long audio (long), short audio (short), and very long audio (very_long), with each configuration having its respective training set.