nguyenvulebinh/libris_clean_100
收藏资源简介:
LibriSpeech是一个包含约1000小时16kHz英语朗读语音的语料库,数据来源于LibriVox项目的有声读物,并经过仔细的分段和对齐处理。数据集支持自动语音识别(ASR)和音频说话人识别任务,并提供了两个配置:clean和other。数据集的结构包括音频文件路径、音频数据、文本转录、说话人ID、章节ID和唯一ID等信息。数据集分为训练集、验证集和测试集,训练集进一步分为train.100、train.360和train.500。数据集的创建者包括Vassil Panayotov、Guoguo Chen、Daniel Povey和Sanjeev Khudanpur,使用CC BY 4.0许可证。
LibriSpeech is a corpus of approximately 1000 hours of 16 kHz English read speech, derived from audiobooks in the LibriVox project, and meticulously segmented and aligned. This dataset supports automatic speech recognition (ASR) and audio speaker recognition tasks, and provides two configurations: clean and other. The dataset structure includes information such as audio file path, audio data, text transcription, speaker ID, chapter ID and unique ID. The dataset is divided into training, validation and test sets, where the training set is further subdivided into train.100, train.360 and train.500. The dataset's creators include Vassil Panayotov, Guoguo Chen, Daniel Povey and Sanjeev Khudanpur, and it is licensed under CC BY 4.0.
数据集概述
数据集名称
- 名称: LibriSpeech
数据集属性
- 语言: 英语 (en)
- 许可证: CC-BY-4.0
- 多语言性: 单语种
- 任务类别: 自动语音识别, 音频分类
- 任务ID: 说话人识别
- 大小类别: 100K<n<1M
- 源数据集: 原始数据
- Paperswithcode ID: librispeech-1
数据集结构
- 配置名称: clean, other, all
- 特征:
- file: 字符串类型
- audio: 音频类型,采样率为16000
- text: 字符串类型
- speaker_id: 整数类型
- chapter_id: 整数类型
- id: 字符串类型
- 数据分割:
- clean配置:
- train.100: 28539个样本,6619683041字节
- train.360: 104014个样本,23898214592字节
- validation: 2703个样本,359572231字节
- test: 2620个样本,367705423字节
- other配置:
- train.500: 148688个样本,31810256902字节
- validation: 2864个样本,337283304字节
- test: 2939个样本,352396474字节
- all配置:
- train.clean.100: 28539个样本,6627791685字节
- train.clean.360: 104014个样本,23927767570字节
- train.other.500: 148688个样本,31852502880字节
- validation.clean: 2703个样本,359505691字节
- validation.other: 2864个样本,337213112字节
- test.clean: 2620个样本,368449831字节
- test.other: 2939个样本,353231518字节
- clean配置:
- 下载大小:
- clean配置: 30121377654字节
- other配置: 31236565377字节
- all配置: 61357943031字节
- 数据集大小:
- clean配置: 31245175287字节
- other配置: 32499936680字节
- all配置: 63826462287字节
数据集创建
- 注释创建者: 专家生成
- 语言创建者: 众包, 专家生成




