VOICE · CLONING · FEW-SHOT
GPT-SoVITS
@RVC-BOSS
Use a few seconds to a minute of reference speech to train your partner a set of their own voices.
GPT-SoVITS turns a small amount of reference audio into a continuously recallable synthesized voice, and is a basic tool for many self-hosted companions to obtain exclusive sounds.
Project address (can be copied to AI):https://github.com/RVC-Boss/GPT-SoVITS
PROJECT INTRO
Project introduction
GPT-SoVITS supports zero-sample TTS with short reference audio, and can also be fine-tuned with about one minute of data to improve timbre similarity, and covers cross-language synthesis such as Chinese, Cantonese, English, Japanese, and Korean.
The project provides data slicing, ASR annotation, training, inference WebUI and API, and can run in Windows, Linux, macOS or Docker.
- Suitable for:People who wish to create their own synthesized vocals for their self-hosted companion.
- Platform:Windows, Linux, macOS and Docker; GPU experience is better.
- Usage:Prepare reference audio that you have permission to use, and directly infer or fine-tune the model after segmentation and annotation.
- License:MIT; sound material and training data still require legal authorization.
COMMENTS & FEEDBACK
Comments and Feedback
People who have used this project can come back and tell newcomers: which platform it ran on, whether there were any pitfalls during the installation, and what the actual experience was like. The current version is in guest mode, and you do not need to register an account to leave a message.
The message is waiting for review and will be publicly displayed here after approval.
SOURCE & CREDIT
Atlas is responsible for the introduction and navigation, and the use still returns to the original author.
We do not mirror project files, nor do we intercept author traffic. For installation, download, version updates and the latest instructions, please refer to the project page provided by the original author.
Reading public comments...