Steps

Install anaconda

mkdir -p ~/Downloads
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/Downloads/miniconda.sh
bash ~/Downloads/miniconda.sh -b -p $HOME/miniconda

# Any new terminal will start the base enviroment
echo 'eval "$($HOME/miniconda/bin/conda shell.bash hook)"' >> ~/.bashrc

# This will activate the enviroemnt for you on this instance
eval "$($HOME/miniconda/bin/conda shell.bash hook)"

Activate conda enviroment

conda create -n dles python=3.7 pip -y
conda activate dles

# After the repo has been cloned, and inside the project root folder, with the
# (dles) enviroment, execute:
pip install -r rtvc/requirements.txt

Play with the project

Commmandline:

For the command line interface, put the pre-trained models in the correct locations

rtvc/vocoder/saved_models/pretrained/pretrained.pt
rtvc/synthesizer/saved_models/logs-pretrained/taco_pretrained/<tf checkpoints>
rtvc/encoder/saved_models/pretrained/pretrained.pt

And run the code with:

cd rtvc
python demo_cli.py

Graphical interface:

cd rtvc
python demo_toolbox.py

Quantization experiments

Files associatied with quantization can be found in:

quantization-encoder
quant_demo_enc.ipynb
quant_demo_synth.ipynb
quantization-vocoder