I made a GUI for the spleeter audio processing library for a client. It allows you to easily use the spleeter library installed on an Anaconda virtual environment.
You can install it from my github here.
Spleeter is a library made by developers at deezer. It allows you to split a song file into separate tracks using machine learning. You can split instrumentals and vocals just from an audio file for pretty much any song, and it does a very good job. There's also an option to split the track into drums, vocals, and synths as well. This is useful for those of you who want to do karaoke to a song that doesn't have an instrumental track released yet, or for people who want to do a vocal cover of a song.
Unfortunately, it doesn't really work on unclean vocals in death metal or metalcore, but that's kind of expected.