AdaVITS: Tiny VITS for Low Computing Resource Speaker Adaptation

TL;DR

AdaVITS is a lightweight VITS-based TTS model designed for low-resource speaker adaptation, achieving stable, natural speech with significantly reduced parameters and computational complexity.

Contribution

It introduces an efficient AdaVITS model with novel components like iSTFT decoder, NanoFlow, and linear attention, enabling low-resource speaker adaptation.

Findings

Achieves stable speech synthesis with 8.97M parameters.

Reduces computational complexity to 0.72GFlops.

Maintains natural speech quality in low-resource scenarios.

Abstract

Speaker adaptation in text-to-speech synthesis (TTS) is to finetune 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 TTS model, named AdaVITS, for low computing resource speaker adaptation is proposed. To effectively reduce parameters and computational complexity of VITS, an 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, instead of using the original text encoder, phonetic posteriorgram (PPG) is utilized as linguistic feature via a text-to-PPG module, which is then used as input for the encoder. Experiment shows that AdaVITS can generate stable and natural speech in speaker adaptation with 8.97M model parameters and 0.72GFlops computational complexity.