ezerhouni/mls-eng-10k-tags_tagged_10k_generated
收藏Hugging Face2024-07-05 更新2024-07-06 收录
下载链接:
https://hf-mirror.com/datasets/ezerhouni/mls-eng-10k-tags_tagged_10k_generated
下载链接
链接失效反馈官方服务:
资源简介:
该数据集包含多个特征字段,涉及音频文件的路径、时间信息、音频时长、说话者ID、书籍ID、音高统计信息、信噪比、语音速率、音素、性别、音高、噪声、混响、语音单调性、原始文本、处理后的文本和文本描述等。数据集分为开发集(dev)、测试集(test)和训练集(train),分别包含3807、3769和2420047个样本,总大小约为2.8GB。
The dataset includes multiple feature fields related to audio file paths, timing information, audio duration, speaker ID, book ID, pitch statistics, signal-to-noise ratio, speaking rate, phonemes, gender, pitch, noise, reverberation, speech monotony, original text, processed text, and text descriptions. The dataset is divided into development (dev), test, and training (train) sets, containing 3807, 3769, and 2420047 samples respectively, with a total size of approximately 2.8GB.
提供机构:
ezerhouni
原始信息汇总
数据集概述
数据集特征
- original_path: 字符串类型,原始路径
- begin_time: 浮点数类型,开始时间
- end_time: 浮点数类型,结束时间
- audio_duration: 浮点数类型,音频时长
- speaker_id: 字符串类型,说话者ID
- book_id: 字符串类型,书籍ID
- utterance_pitch_mean: 浮点数类型,音高均值
- utterance_pitch_std: 浮点数类型,音高标准差
- snr: 浮点数类型,信噪比
- c50: 浮点数类型,C50值
- speaking_rate: 字符串类型,说话速率
- phonemes: 字符串类型,音素
- gender: 字符串类型,性别
- pitch: 字符串类型,音高
- noise: 字符串类型,噪声
- reverberation: 字符串类型,混响
- speech_monotony: 字符串类型,语音单调性
- original_text: 字符串类型,原始文本
- text: 字符串类型,文本
- text_description: 字符串类型,文本描述
数据集分割
- dev:
- 字节数: 4403292
- 样本数: 3807
- test:
- 字节数: 4386291
- 样本数: 3769
- train:
- 字节数: 2794801012
- 样本数: 2420047
数据集大小
- 下载大小: 1450361224 字节
- 数据集总大小: 2803590595 字节
配置
- config_name: default
- data_files:
- split: dev, path: data/dev-*
- split: test, path: data/test-*
- split: train, path: data/train-*
- data_files:



