Download
CosyVoice is a multilingual large voice generation model that offers a full-stack solution for training, inference, and deployment of high-quality TTS systems. The model supports multiple languages, including Chinese, English, Japanese, Korean, and a range of Chinese dialects such as Cantonese, Sichuanese, Shanghainese, Tianjinese, and Wuhanese. It is designed for zero-shot voice cloning and cross-lingual or mix-lingual scenarios, so a single reference voice can be used to synthesize speech across languages and in code-switching contexts. CosyVoice 2.0 significantly improves on version 1.0 by boosting accuracy, stability, speed, and overall speech quality, making it more suitable for production environments. The repository contains training recipes, inference pipelines, deployment scripts, and integration examples, positioning it as a comprehensive toolkit rather than just a set of model weights.
Features
- Multilingual TTS with support for major languages and many Chinese dialects
- Zero-shot voice cloning, including cross-lingual and code-switching speech synthesis
- CosyVoice 2.0 architecture offering higher accuracy, stability, and faster generation than 1.0
- End-to-end recipes for training, inference, and deployment in real applications
- Integration examples with other FunAudioLLM components for emotional voice chat and complex audio agents
- Open-source code and models that can run on standard GPU hardware with documented requirements
Project Samples
