This code is an adaption of ws-choi Conditioned-Source-Separation-LaSAFT to run over CPU in a Docker container.
docker build -f Dockerfile -t lasaft .
docker run --rm -it -d -v `pwd`/src:/app -v /root:/root --name lasaft-test lasaft
docker exec -it lasaft-test bash
python Separation_wav.py -mix_scp /root/test.wav -yaml options/train/train.yml -model /root/best.pt -save_path ./checkpoint
python inference.py --url "youtube-url" --s 0 --e 10
python inference.py --a /path/to/audio.wav --s 0 --e 10