ICASSP2020-demo

"Teacher-Student Training for Robust Tacotron-based TTS",

Authors: Rui Liu, Berrak Sisman, Jingdong Li, Feilong Bao, Guanglai Gao, Haizhou Li

accepted and to be presented at ICASSP2020

paper

Speech samples

Examples of synthesized speech by "Tacotron2-SS" , "Tacotron2-FR" and "Tacotron2-KD" systems.

Speech samples are available at demo page.

Tacotron2-FR: Tacotron2-based model with free running (FR) mode

Tacotron2-SS: Tacotron2-based model with scheduled sampling (SS) mode

Tacotron2-KD: Tacotron2-based model with proposed teacher-student training method by using knowledge distillation (KD)

Reference

Please kindly cite this paper in your publications if it helps your research:

@inproceedings{Rui2020tacotron-KD,
    title={Teacher-Student Training for Robust Tacotron-based TTS},
    author={Rui Liu, Berrak Sisman, Jingdong Li, Feilong Bao, Guanglai Gao and Haizhou Li},
    booktitle={Proceedings of the 2020 IEEE International Conference on Acoustics,  Speech and Signal Processing (ICASSP)},
    year={2020}
}

ttslr/ICASSP2020

ICASSP2020-demo

Speech samples

Reference