GPT-SoVITS

Implements a few-shot voice cloning and text-to-speech web application that trains personalized, high-fidelity synthetic voices from as little as one minute of recorded audio.