eustlb/big-bench-audio-less-than-30s
收藏Hugging Face2025-02-11 更新2025-02-15 收录
下载链接:
https://hf-mirror.com/datasets/eustlb/big-bench-audio-less-than-30s
下载链接
链接失效反馈官方服务:
资源简介:
这是一个包含音频数据、类别标签、官方答案、唯一标识符、转录文本和原始问题的数据集。音频数据采样率为16000Hz,数据集分为验证集,共有792个示例。该数据集用于语音识别任务,其中的原始问题是从BIG-Bench-Hard项目中获取的。
This dataset includes audio data, category labels, official answers, unique identifiers, transcription texts, and original questions. The audio data has a sampling rate of 16000Hz, and the dataset is split into a validation set with a total of 792 examples. The dataset is used for speech recognition tasks, with the original questions retrieved from the BIG-Bench-Hard project.
提供机构:
eustlb



