MnTTS2: An Open-Source Multi-Speaker Mongolian Text-to-Speech Synthesis Dataset

Introduction

This is the experimental description of MnTTS2.

Kailin Liang, Bin Liu, Yifan Hu, Rui Liu *, Feilong Bao, Guanglai Gao.

Links

20 audio tracks synthesised for three speakers. Griffin-Lim vocoder, HiFi-GAN vocoder, VITS and Ground-Tuth are included, with corresponding texts given for reference： Inference Audios

Please contact the Author to access the Pre-trained Models. liangkailin98@foxmail.com

Acknowledgements:

Tensorflow-TTS: https://github.com/TensorSpeech/TensorFlowTTS

VITS: https://github.com/jaywalnut310/vits

MnTTS: https://github.com/walker-hyf/MnTTS