rvc

Retrieval-based-Voice-Conversion

Input

Audio file

booth.mov

(Audio from https://github.com/ohashi3399/RVC-demo)

Output

Audio file

output.mov

Requirements

This model requires additional module.

pip3 install librosa
pip3 install soundfile
pip3 install faiss-cpu==1.7.3
pip3 install pyworld==0.3.2

Usage

Automatically downloads the onnx and prototxt files on the first run. It is necessary to be connected to the Internet while downloading.

For the sample wav,

$ python3 rvc.py

If you want to specify the audio, put the file path after the --input option.

$ python3 rvc.py --input AUDIO_FILE

By adding the --model_file option, you can specify vc model file.

$ python3 rvc.py --model_file AISO-HOWATTO.onnx

Specify the f0 option to infer a model that uses f0. You can choice crepe or crepe_tiny for f0_method.

python3 rvc.py -i booth.wav -m Rinne.onnx --f0_method crepe_tiny --f0 1 --f0_up_key 11 --tgt_sr 48000

By adding the --file_index option, you can specify faiss feature file.

python3 rvc.py -i booth.wav -m Rinne.onnx --f0_method crepe --f0 1 --f0_up_key 11 --tgt_sr 48000 --file_index Rinne.index --index_rate 0.75

By adding the --version option, you can specify rvc model file version.

$ python3 rvc.py --model_file rvc_v2.onnx --version 2

Reference

• Retrieval-based-Voice-Conversion-WebUI
• RVC向け学習済みボイスモデルデータ

Framework

Pytorch

Model Format

ONNX opset=14

Netron

• hubert_base.onnx.prototxt
• AISO-HOWATTO.onnx.prototxt
• crepe.onnx.prototxt
• crepe_tiny.onnx.prototxt