1 paper
Javier Latorre, Jakub Lachowicz, Jaime Lorenzo-Trueba +4
Recent speech synthesis systems based on sampling from autoregressive neural networks models can generate speech almost undistinguishable from human recordings. However, these mode…