kotoba-speech/simul_eval_aws_ja2en_1b_audio_labels

该数据集包含音频文件及其对应的开始和结束时间戳，这些时间戳以日语和英语两种语言给出。每个音频文件还有一个与之相关的文本内容，同样提供日语和英语两种版本。此外，还包括一些可能用于标注或记录音频质量的字段，如错误数、缺失数据和不自然标记。数据集分为训练集，共包含75个样本。但没有提供关于数据集具体内容和用途的描述。

The dataset includes audio files along with their start and end timestamps in both Japanese and English. Each audio file is associated with a text content that is provided in both Japanese and English. Additionally, there are fields that may be used for annotation or recording audio quality, such as the number of errors, missing data, and unnatural markers. The dataset is split into a training set, containing 75 samples in total. No specific description or purpose of the dataset is provided.