Uberduck is a tool for fun and creativity with audio machine learning, currently focused on voice cloning and neural text-to-speech. This repository includes development tools to get started with creating your own speech synthesis model. For more information on the state of this repo, please see the Wiki.
An overview of the subpackages in this library:
models
: TTS model implementations. All models descend from models.base.TTSModel
.
trainer
: A trainer has logic for training a model.
exec
: Contains entrypoint scripts for running training jobs. Executed via a command like
python -m uberduck_ml_dev.exec.train_tacotron2 --your-args here
conda create -n 'uberduck-ml-dev' python=3.8
source activate uberduck-ml-dev
pip install git+https://github.com/uberduck-ai/uberduck-ml-dev.git
To start contributing, install the required development dependencies in a virtual environment:
pip install pre-commit black
Clone the repository:
git clone [email protected]:uberduck-ai/uberduck-ml-dev.git
Install required Git hooks:
pre-commit install
Install the library:
python setup.py develop
python -m pytest
- It is important for you to spell the name of your user and repo correctly in
settings.ini
. If you change the name of the repo, you have to make the appropriate changes insettings.ini