setup-all.sh

#!/bin/bash
#The intent here is to set up a server that can do the things the scripts require. I'm using runpod.io servers that already have pytorch installed; they use /workspace as a root working directory. Change BASEDIR as appropriate or your setup. 

#apt update
#apt --yes install libsndfile1-dev ffmpeg emacs python3.9-dev nq less sox

export TORTOISE_MODELS_DIR=$BASEDIR/models
export NQDIR=$BASEDIR/outputs
cd $BASEDIR
mkdir inputs
mkdir logs
mkdir outputs
mkdir models
git clone https://github.com/neonbjb/tortoise-tts
git clone https://github.com/NVIDIA/NeMo-text-processing/
cp -r  wikipedia-tts/voices/* tortoise-tts/tortoise/voices/
python3.9 -m pip install -r tortoise-tts/requirements.txt
python3.9 -m pip install wikipedia Cython pynini
python3.9 -m pip install git+https://github.com/NVIDIA/NeMo.git@main#egg=nemo_toolkit[nlp]

cd tortoise-tts

python -m pip install .

cd models
# wget `../wikipedia-tts/models.txt`