AdaVITS: Tiny VITS for Low Computing Resource Speaker Adaptation

Kun Song, Heyang Xue, Xinsheng Wang, Jian Cong, Yongmao Zhang, Lei Xie, Bing Yang, Xiong Zhang, Dan Su

科研成果: 书/报告/会议事项章节会议稿件同行评审

5 引用 (Scopus)

摘要

Speaker adaptation in text-to-speech synthesis (TTS) is to fine-tune a pre-trained TTS model to adapt to new target speakers with limited data. While much effort has been conducted towards this task, seldom work has been performed for low computational resource scenarios due to the challenges raised by the requirement of the lightweight model and less computational complexity. In this paper, a tiny VITS-based [1] TTS model, named AdaVITS, for low computing resource speaker adaptation is proposed. To effectively reduce the parameters and computational complexity of VITS, an inverse short-time Fourier transform (iSTFT)-based wave construction decoder is proposed to replace the upsampling-based decoder which is resource-consuming in the original VITS. Besides, NanoFlow is introduced to share the density estimate across flow blocks to reduce the parameters of the prior encoder. Furthermore, to reduce the computational complexity of the textual encoder, scaled-dot attention is replaced with linear attention. To deal with the instability caused by the simplified model, we use phonetic posteriorgram (PPG) as a frame-level linguistic feature for supervising the model process from phoneme to spectrum. Experiments show that AdaVITS can generate stable and natural speech in speaker adaptation with 8. 97M model parameters and 0.72 GFlops computational complexity.1

源语言英语
主期刊名2022 13th International Symposium on Chinese Spoken Language Processing, ISCSLP 2022
编辑Kong Aik Lee, Hung-yi Lee, Yanfeng Lu, Minghui Dong
出版商Institute of Electrical and Electronics Engineers Inc.
319-323
页数5
ISBN(电子版)9798350397963
DOI
出版状态已出版 - 2022
活动13th International Symposium on Chinese Spoken Language Processing, ISCSLP 2022 - Singapore, 新加坡
期限: 11 12月 202214 12月 2022

出版系列

姓名2022 13th International Symposium on Chinese Spoken Language Processing, ISCSLP 2022

会议

会议13th International Symposium on Chinese Spoken Language Processing, ISCSLP 2022
国家/地区新加坡
Singapore
时期11/12/2214/12/22

指纹

探究 'AdaVITS: Tiny VITS for Low Computing Resource Speaker Adaptation' 的科研主题。它们共同构成独一无二的指纹。

引用此