You would think getting the notes and converting them to chords, and tablature, would be one of those "exactly suited for neural nets" type of problem. If you want the chords to "Here comes the sun," you can find dozens of hits, but try something slightly obscure and they are hard to come by. (People with great ears have no idea what I am talking about.)
Thus propagating the NN "If all you have is a hammer" trend. I don't quite know the nuance of music theory, but could you not get away with traditional Fourier analysis? You just need to decompose the song into its constituent frequency "bins" right?
No, at the very least you also need to do the instrument separation, e.g. bass might match guitar chord root note at one moment, and do something entirely different half a second later, and of course in a coherent song all the different instruments will merge in the same frequency bins, either directly or with their overtones. Also, for certain instruments and effects on e.g. guitar amps the strongest peaks of Fourier transform for the instrument output may not necessarily match the notes that are actually being played.