End-to-End Text-to-Speech Using Latent Duration Based on VQ-VAE

Publication
ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)