GPT-SoVITS-WebUI

GPT-SoVITS-WebUI is a few-shot voice conversion and text-to-speech WebUI. It targets voice cloning workflows where only a short sample or a small training set is available, turning reference speech into a voice for synthesized text rather than requiring a large recorded corpus.

Zero-shot TTS accepts a 5-second vocal sample for immediate text-to-speech conversion. Few-shot TTS can fine-tune with 1 minute of training data to improve voice similarity and realism. Cross-lingual inference supports English, Japanese, Korean, Cantonese, and Chinese, including inference in a language different from the training dataset.

The WebUI includes tools for voice accompaniment separation, automatic training set segmentation, Chinese ASR, and text labeling to help build training datasets and GPT/SoVITS models. It can run from a Windows integrated package, a Python environment with webui.py, or Docker Compose, with an AutoDL Cloud Docker option for users in China.