Spleeter, a Python+tensorflow voice-from-everything-else separation algorithm (deezer.io)

I'd be very interested in seeing if it can extract the bass line in any of the tracks of Ride the lighting album by Metallica