davanstrien/test_audio
收藏Hugging Face2024-12-17 更新2024-12-21 收录
下载链接：
https://hf-mirror.com/datasets/davanstrien/test_audio
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含多个字段，涵盖了音频文件的起始时间、结束时间、持续时间、文本内容等信息。此外，还包含了音频转录的多种结果（如Whisper和Wav2Vec的转录结果）以及相关的评估指标（如BLEU和WER分数）。数据集还包含音频文件的基本信息（如音频文件路径、采样率）以及说话者的相关信息（如姓名、性别、角色等）。数据集的训练集包含100个样本，总大小为78075901字节。
This is a dataset containing speech transcriptions and related metadata. The dataset includes various features such as start and end times of speech, duration, text transcriptions, transcription results from different models (e.g., Whisper and wav2vec), BLEU and WER scores, speech identifiers, protocol ID, sub IDs, language probabilities, audio file paths, speaker information (e.g., gender, role, district), dates, year, shard information, and the audio data itself. The dataset is divided into a training set with 100 samples.
提供机构：
davanstrien
