Text-to-Speech (TTS) System / Machine Learning Library
8.7k 2026-04-18
fishaudio/Bert-VITS2
An open-source Text-to-Speech system built on the VITS2 backbone, enhanced with multilingual BERT for improved speech synthesis.
Core Features
High-quality speech synthesis via VITS2 backbone.
Enhanced linguistic understanding with multilingual BERT.
Multilingual support for diverse speech generation.
Open-source and community-driven development.
Detailed Introduction
Bert-VITS2 is an advanced open-source Text-to-Speech (TTS) project that leverages the robust VITS2 architecture and integrates multilingual BERT. This combination aims to produce high-fidelity, natural-sounding speech across multiple languages by improving the model's understanding of linguistic nuances. While the project's direct maintenance is currently paused in favor of Fish-Speech, it represents a significant contribution to the open-source TTS community, offering a powerful foundation for researchers and developers to explore and build upon for diverse speech synthesis applications.