Control synthesis

Audio examples are available here

Usage

Prepare dataset

put 16k mono audio in a directory like so:

datasets/<dataset_name>/audio_16k

from root directory:

sh scripts/generate_controlsynthesis_dataset.sh <dataset_name>

Train synthesis model

sh scripts/train_synthesis_model.sh <dataset_name>

Train control model

sh scripts/train_control_model.sh <dataset_name>