current position:Home>Python AI steals your voice in 5 seconds

Python AI steals your voice in 5 seconds

2022-01-31 12:41:05 Xiao Chen operation and maintenance

Introduce

Python Deep learning AI - Voice cloning 、 Sound imitation , It is a three-stage in-depth learning framework , Allows you to create a digital representation of speech from a few seconds of audio , And use it to adjust text to speech model , The model has been trained , It can be summarized as a new sound .

Environment preparation and installation

Original English address :

github.com/CorentinJ/R…

Chinese secondary development version ( This article uses this version ):

github.com/babysor/Moc…

pycharm Environment Download :

www.jetbrains.com/pycharm/dow…

conda A virtual environment :

www.anaconda.com/products/in…

FFmpeg :

github.com/BtbN/FFmpeg…

Model file :

pan.baidu.com/s/1PI-hM3sn… Extraction code 2021

Install... On the computer system FFmpeg Tools

download zip The compressed package connection is :github.com/BtbN/FFmpeg…

After downloading, unzip it to a directory and add the directory to the environment variable of the system

 picture

Open up new cmd Check to see if the installation is successful

ffmpeg -version

 picture

Use to open the project directory , Create with conda Of Python 3.9 A virtual environment

 picture

After creation , stay cmd View existing virtual environments in , And enter the virtual environment just created

conda env list

activate pythonProject1

 picture

Install after entering the environment pip The required depend on , And use the domestic source for installation to speed up the download

pip install -r requirements.txt -i pypi.tuna.tsinghua.edu.cn/simple

 picture

Install... In a virtual environment pytorch

pip install torch  -i pypi.tuna.tsinghua.edu.cn/simple

 picture

go back to pycharm in , Import the model into the project directory , Copy and paste the directory into the project

 picture

Modify a line of code , stay synthesizer/utils/symbols.py In file

 It is amended as follows :
_characters = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz12340!'(),-.:;? '
 Copy code 

 picture

After the terminal Start the toolbox in

 picture

Use the Audio Synthesis Toolbox

 picture

Linux Operation and maintenance exchange community

Linux Operation and maintenance exchange community , Internet news and technology exchange .

48 Original content

official account

 picture

blog.csdn.net/qq_33921750

my.oschina.net/u/3981543

www.zhihu.com/people/chen…

segmentfault.com/u/hppyvyv6/…

juejin.cn/user/331578…

space.bilibili.com/352476552/a…

cloud.tencent.com/developer/c…

You know 、CSDN、 Open source in China 、 Think no 、 Nuggets 、 Bili, Bili 、 Tencent cloud

This article USES the Article synchronization assistant Sync

copyright notice
author[Xiao Chen operation and maintenance],Please bring the original link to reprint, thank you.
https://en.pythonmana.com/2022/01/202201311241037502.html

Random recommended