2021-10-05 17:50:16 +01:00
|
|
|
# Welcome to the Voice Cloning App
|
|
|
|
|
|
|
|
|
|
A Python/Pytorch app for easily synthesising human voices.
|
|
|
|
|
|
|
|
|
|
## Getting started
|
|
|
|
|
|
|
|
|
|
After installing & running the app it should open in your browser at [localhost:5000](http://localhost:5000/)
|
|
|
|
|
|
2021-10-05 18:12:05 +01:00
|
|
|
|
2021-10-05 17:50:16 +01:00
|
|
|
|
|
|
|
|
### Windows
|
|
|
|
|
|
|
|
|
|
Download the latest release from [releases](https://github.com/BenAAndrew/Voice-Cloning-App/releases)
|
|
|
|
|
|
|
|
|
|
- If you have an NVIDIA GPU make sure to download the executable with GPU support
|
|
|
|
|
- Otherwise, download the 'cpuonly' version
|
|
|
|
|
|
|
|
|
|
### Linux
|
|
|
|
|
|
|
|
|
|
1. Clone this repository
|
|
|
|
|
2. Run `./install.sh` from the root of the repository
|
|
|
|
|
3. Run `python3.6 main.py`
|
|
|
|
|
|
|
|
|
|
### Manual Install (Linux/ Windows)
|
|
|
|
|
|
|
|
|
|
Users familiar with Python may prefer building the app themselves:
|
|
|
|
|
|
|
|
|
|
1. Clone this repository
|
|
|
|
|
2. Install [Python](https://www.python.org/) (version 3.6)
|
|
|
|
|
3. Run `pip install -r requirements.txt`
|
|
|
|
|
- (or `pip install -r requirements-cpu.txt` for the 'cpuonly' build)
|
|
|
|
|
4. Run `python main.py`
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
## Adding languages
|
|
|
|
|
|
|
|
|
|
If you are using a language other than English you can add it to the app.
|
|
|
|
|
|
|
|
|
|
Firstly, you'll need to find a deepspeech model for your language by going to [coqui](https://coqui.ai/models).
|
|
|
|
|
You'll then need to download the `model.pbmm` and `alphabet.txt` files for your language.
|
|
|
|
|
|
|
|
|
|
To add this to the app go to "Settings" -> "Add a language", enter your language name and upload your language files
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Please note that language model quality may vary and so successful generation of voices in other languages is not guarenteed.
|
