malaysia-ai/common_voice_17_0

Common Voice Corpus 17.0是一个公共领域的音频数据集，包含多个特征字段，如用户ID、路径、句子ID、句子内容、领域、投票数、年龄、性别、口音、变体和地区等。数据集分为训练集、测试集和验证集，分别包含668,9550、533,642和507,874个示例。数据集总大小约为3.31GB，下载大小约为1.03GB。此数据集是mozilla-foundation/common_voice_17_0数据集的镜像，方便下载和解压音频文件。

Common Voice Corpus 17.0 is a public domain audio dataset that includes various feature fields such as user ID, path, sentence ID, sentence content, domain, votes, age, gender, accents, variants, and locale. The dataset is split into training, testing, and validation sets containing 6,689,550, 533,642, and 507,874 examples respectively. The total size of the dataset is approximately 3.31GB, with a download size of about 1.03GB. This dataset is a mirror of the mozilla-foundation/common_voice_17_0 dataset, making it easy to download and extract audio files in parquet format.