OFFICIAL PAGE: Open-Unmix
According to their web page: "Open-Unmix, is a deep neural network reference implementation for music source separation, applicable for researchers, audio engineers and artists. Open-Unmix provides ready-to-use models that allow users to separate pop music into four stems: vocals, drums, bass and the remaining other instruments." "The model is available for three different frameworks. However, the pytorch implementation serves as the reference version that includes pre-trained networks trained on the MUSDB18 dataset." Open-Unmix is a command line application with no GUI (graphical user interface).
Image courtesy of Inria