1 paper · 1 filter
Konrad Szewczyk, Daniel Gallo Fernández, James Townsend
Directly learning to generate audio waveforms in an autoregressive manner is a challenging task, due to the length of the raw sequences and the existence of important structure on…