LJ Speech Dataset

利用 LJ Speech Dataset 单人语音训练英文 TTS，大概有3G的文件大小，里面有15000条语音信息，以WAV为主

This dataset is the LJ Speech Dataset, intended for training English Text-to-Speech (TTS) models using single-speaker speech data. It has an approximate total file size of 3 GB, containing roughly 15,000 audio utterances, most of which are in WAV format.

搜集汇总

背景与挑战

背景概述

LJ Speech Dataset是一个包含15000条WAV格式英文语音的公开数据集，总大小约3GB，主要用于训练英文文本到语音(TTS)系统。数据集采用分卷压缩形式存储，共7个压缩文件，遵循GPL 2.0许可协议。