Masked Audio Generation using a Single Non-Autoregressive Transformer Paper โข 2401.04577 โข Published Jan 9 โข 42