Shypd.aiThe largest open index of AI toolsReal-Time-Voice-Cloning is an open-source voice cloning tool that generates arbitrary speech in real-time from a few seconds of audio. It implements Transfer Learning from Speaker Verification to Multispeaker Text-To-Speech Synthesis (SV2TTS).
Trending
About
Real-Time-Voice-Cloning is an open-source implementation of Transfer Learning from Speaker Verification to Multispeaker Text-To-Speech Synthesis (SV2TTS), designed to clone a voice in just 5 seconds and generate arbitrary speech in real-time. This tool leverages a three-stage deep learning framework, starting with creating a digital voice representation from a short audio sample. This representation then serves as a reference for generating speech from any given text. The repository includes implementations of key papers like WaveRNN for vocoding, Tacotron for synthesis, and GE2E for the encoder. It supports both Windows and Linux, requiring ffmpeg and uv for installation, and offers both GUI and command-line interfaces. Pretrained models are automatically downloaded, and users can optionally download datasets or use their own audio files.
Capabilities
Pricing & Plans
Open Source
Free
FAQs
Trending
