Using SampleRNN to train a raw-audio model on NOFX. Definitely sounds like NOFX, but nonsense lyrics - View it on GitHub