argmaxinc/librispeech-openbench

这个数据集包含了两种配置的音频数据，分别为200和full。每种配置都包括音频文件、音频持续时间、音频的文本转录和音频的唯一标识符。音频采样率为16000Hz。数据集分为测试集，其中200配置的测试集包含200个示例，而full配置的测试集包含2620个示例。数据集的总大小和下载大小也有所不同。

The dataset includes audio data in two configurations: 200 and full. Each configuration consists of audio files, audio duration, text transcription of the audio, and a unique identifier for the audio. The audio sampling rate is 16000Hz. The dataset is split into a test set, with the 200 configurations test set containing 200 examples, and the full configurations test set containing 2620 examples. The total size and download size of the dataset also vary.