MockingBird real-time voice cloning system v1.0

MockingBird is an AI onomatopoeia system that can clone your voice and generate arbitrary speech content.

characteristic

1. The text supports Mandarin and is tested using a variety of Chinese data sets: aidatatang_200zh, magicdata, aishell3, biaobei, MozillaCommonVoice, data_aishell, etc.

2. PyTorch is suitable for pytorch, has been tested in 1.9.0 (latest in August 2021), GPU Tesla T4 and GTX 2060

3. Windows + Linux can run in Windows operating system and Linux operating system (the Apple system M1 version also has successful community operation cases)

4. Easy & Awesome Just download or newly train the synthesizer (synthesizer) to achieve good results, reuse the pre-trained encoder/vocoder, or real-time HiFi-GAN as the vocoder

5. Webserver Ready can serve your training results for remote calling

start

Installation requirements

Follow the original repository to test whether you have all your environments ready. Python 3.7 or higher is required to run the toolbox.

Install PyTorch.

If ERROR: Could not find a version that satisfies the requirement torch==1.9.0+cu102 (from versions: 0.1.2, 0.1.2.post1, 0.1.2.post2) occurs when installing with pip It may be that python is too low, 3.9 can be installed successfully

Install ffmpeg.

Run pip install -r requirements.txt to install the remaining necessary packages.

Install webrtcvad pip install webrtcvad-wheels.