Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Tutorial: Windows installation #647

Closed
car1ot opened this issue Feb 2, 2021 · 16 comments
Closed

Tutorial: Windows installation #647

car1ot opened this issue Feb 2, 2021 · 16 comments

Comments

@car1ot
Copy link

car1ot commented Feb 2, 2021

Hello,
I have an issues with installing to windows (10). Please, could you give a guide?

@ghost
Copy link

ghost commented Feb 2, 2021

Get files for Real-Time-Voice-Cloning

  1. Download this file to get the latest version of Real-Time-Voice-Cloning: https://github.com/CorentinJ/Real-Time-Voice-Cloning/archive/master.zip
  2. Extract it to C:\ The files will be in a folder called Real-Time-Voice-Cloning-master

Install Python

  1. Download Python from here: https://www.python.org/ftp/python/3.7.9/python-3.7.9-amd64.exe
    • Warning: Other versions of Python may not work.
    • Warning: Do not install Python from the Windows Store.
  2. Run the installer. Click the "Install Now" button. Do not change the defaults.
  3. Exit the installer when complete. Do not disable the path length limit.
  4. Edit the environment variables for your account. Add these locations to PATH. First is for python.exe, second is for pip.exe. Move these to the top to take precedence over any pre-existing Python installations.
%USERPROFILE%\AppData\Local\Programs\Python\Python37
%USERPROFILE%\AppData\Roaming\Programs\Python\Python37\Scripts

Set up virtual environment and install required packages

  1. Open Windows command prompt and run these commands to set up and activate a virtual environment. Real-Time-Voice-Cloning needs this dedicated environment as it uses obsolete packages that could cause conflicts with other Python programs.
cd C:\Real-Time-Voice-Cloning-master
python -m venv venv
venv\Scripts\activate.bat
  1. Next, run these commands to install Python packages.
pip install --upgrade pip
pip install torch -f https://download.pytorch.org/whl/torch_stable.html
pip install -r requirements.txt

Get FFmpeg

  1. Download FFmpeg from here: https://github.com/BtbN/FFmpeg-Builds/releases/download/autobuild-2021-10-31-12-23/ffmpeg-N-104454-gd92fdc7144-win64-lgpl.zip
  2. Extract the zip file. Move ffmpeg.exe to C:\Real-Time-Voice-Cloning-master (the same folder as demo_toolbox.py)

Pretrained models

  1. Download pretrained models from: https://github.com/blue-fish/Real-Time-Voice-Cloning/releases/download/v1.0/pretrained.zip
  2. Extract pretrained.zip. Move the included files to the below locations. See here if you need more help on this step.
C:\Real-Time-Voice-Cloning-master\encoder\saved_models\pretrained.pt
C:\Real-Time-Voice-Cloning-master\synthesizer\saved_models\pretrained\pretrained.pt
C:\Real-Time-Voice-Cloning-master\vocoder\saved_models\pretrained\pretrained.pt

Launching the toolbox

  1. Open Windows command prompt and type:
cd C:\Real-Time-Voice-Cloning-master
venv\Scripts\activate.bat
python demo_toolbox.py

@bigdog3604
Copy link

I have been trying to install this program for over two days without any luck. I've followed the "Poorly Documented" guide and believe I have followed all the steps. I try to launch the CLI toolbox and get an error: ModuleNotFound: No module name 'webrtcvad'. I go to the Python 3.7 lib folder and the module is there. I would love to work with @tomodachi94 to test out his new process for Windows 10 users. The existing process sucks.

@ghost
Copy link

ghost commented Feb 3, 2021

@bigdog3604
Since #376, webrtcvad is optional and the toolbox will run without it.
Make sure you have downloaded the latest version of Real-Time-Voice-Cloning from this site.
Here's a download link: master.zip

@ghost
Copy link

ghost commented Feb 3, 2021

The "Poorly Documented" guide is obsolete given the recent usability improvements. It is hurting users more than it is helping now. We should stop linking to it in the README.

@wyattbiggs
Copy link

wyattbiggs commented Feb 3, 2021

If you'd like to use
webrtcvad, I think it requires the 64 bit version of Python 3.7, but I could be confusing that with something else...

@CarvellScott
Copy link

I'm attempting to install this on Windows 10 as well. I used Anaconda to install Pytorch and the rest of the requirements, but as far as downloading ffmpeg goes, I'm not sure where to place it for the executable to be found...

Nonetheless, the rest of the installation seemed to proceed without any issues. Upon running the demo however, I see this message:

C:\Users\Carvell\gitrepos\Real-Time-Voice-Cloning\encoder\audio.py:13: UserWarning: Unable to import 'webrtcvad'. This package enables noise removal and is recommended.
  warn("Unable to import 'webrtcvad'. This package enables noise removal and is recommended.")
Arguments:
    enc_model_fpath:   encoder\saved_models\pretrained.pt
    syn_model_fpath:   synthesizer\saved_models\pretrained\pretrained.pt
    voc_model_fpath:   vocoder\saved_models\pretrained\pretrained.pt
    cpu:               False
    no_sound:          False
    seed:              None
    no_mp3_support:    False

A s s e r t i o n   f a i l e d !

 P r o g r a m :   C : \ U s e r s \ C a r v e l l \ a n a c o n d a 3 \ e n v s \ v o i c e _ c l o n i n g \ p y t h o n . e x e
 F i l e :   s r c / h o s t a p i / w d m k s / p a _ w i n _ w d m k s . c ,   L i n e   1 0 6 1

 E x p r e s s i o n :   F A L S E

The s p a c e d o u t text appears there as it does in my terminal. The python version in the conda env I'm using is 3.8.8.

@AyseEe501
Copy link

this can help #758 (comment)

@ghost ghost pinned this issue Sep 8, 2021
@Tomcattwo
Copy link
Contributor

Tomcattwo commented Sep 26, 2021

@CarvellScott You can install ffmpeg via sourceforge in Anaconda using this command:
conda install -c conda-forge ffmpeg-python
See here for details: https://anaconda.org/conda-forge/ffmpeg-python
Regards, Tomcattwo

@Tornakin
Copy link

I've been able to install every thing else so far but can't figure out what I'm supposed to do with (pip install -r requirements.txt)

@LukeCronquist
Copy link

LukeCronquist commented Oct 20, 2021

Error: Model files not found. even though they are there D;

Also UserWarning: PySoundFile failed. also No such file or directory: 'samples/1320_00000.mp3' when run the test config

EDIT: my bad, skimmed over the 3.6/.7 only

@ghost
Copy link

ghost commented Oct 20, 2021

Error: Model files not found. even though they are there D;

Most of the time, can be solved with #758 (comment) or #642 (comment)

@MuhammadWaqar621
Copy link

i want to train for other language, can you able to help me?

@Tomcattwo
Copy link
Contributor

Tomcattwo commented Apr 10, 2023 via email

@PixelcrafterEXE
Copy link

Get FFmpeg

  1. Download FFmpeg from here: https://github.com/BtbN/FFmpeg-Builds/releases/download/autobuild-2021-10-31-12-23/ffmpeg-N-104454-gd92fdc7144-win64-lgpl.zip
  2. Extract the zip file. Move ffmpeg.exe to C:\Real-Time-Voice-Cloning-master (the same folder as demo_toolbox.py)

Alternatively winget install FFmpeg would be more elegant.

@MaraScott
Copy link

Damn ! Thx

@goldentechie
Copy link

image
Installed Python3.7 by following the tutorial. Can't install venv

This issue was closed.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests