Changelog | FAQ (Frequently Asked Questions)
English | 中文简体 | 日本語 | 한국어 (韓國語) | Français | Türkçe | Português
Check out our Demo Video here!
| Training and inference Webui | Real-time voice changing GUI |
 |
 |
| go-webui.bat | go-realtime_gui.bat |
| You can freely choose the action you want to perform. | We have achieved an end-to-end latency of 170ms. With the use of ASIO input and output devices, we have managed to achieve an end-to-end latency of 90ms, but it is highly dependent on hardware driver support. |
The dataset for the pre-training model uses nearly 50 hours of high quality audio from the VCTK open source dataset.
High quality licensed song datasets will be added to the training-set often for your use, without having to worry about copyright infringement.
Please look forward to the pretrained base model of RVCv3, which has larger parameters, more training data, better results, unchanged inference speed, and requires less training data for training.
- Reduce tone leakage by replacing the source feature to training-set feature using top1 retrieval;
- Easy + fast training, even on poor graphics cards;
- Training with a small amounts of data (>=10min low noise speech recommended);
- Model fusion to change timbres (using ckpt processing tab->ckpt merge);
- Easy-to-use WebUI;
- pymss/MSST model to quickly separate vocals and instruments;
- High-pitch Voice Extraction Algorithm InterSpeech2023-RMVPE to prevent a muted sound problem. Provides the best results (significantly) and is faster with lower resource consumption than Crepe_full;
- AMD/Intel systems use the CPU dependency set; Windows may use DirectML and Linux uses CPU;
This branch targets Python 3.12 x64. Run every command from the repository root. Ubuntu 24.04 x86_64 is recommended.
sudo apt update
sudo apt install -y python3.12 python3.12-venv python3.12-dev ffmpeg unzip libsndfile1 libportaudio2
python3.12 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip setuptools wheelInstall Python 3.12 x64, then create a virtual environment:
py -3.12 -m venv .venv
.venv\Scripts\activate
python -m pip install --upgrade pip setuptools wheel| Hardware | Installation |
|---|---|
| CPU, AMD, Intel | Use requirments_cpu_py312.txt; Windows may use DirectML, while Linux uses CPU |
| NVIDIA RTX 50 series | Install the CUDA 12.8 Torch pair first, then requirments_cu128_py312.txt |
| NVIDIA GPUs before the RTX 50 series | Install the CUDA 11.8 Torch pair first, then requirments_cu118_py312.txt |
python -m pip install -r requirments_cpu_py312.txtpython -m pip install torch==2.7.1+cu128 torchaudio==2.7.1+cu128 \
--index-url https://download.pytorch.org/whl/cu128 \
--extra-index-url https://pypi.org/simple
python -m pip install -r requirments_cu128_py312.txtpython -m pip install torch==2.7.1+cu118 torchaudio==2.7.1+cu118 \
--index-url https://download.pytorch.org/whl/cu118 \
--extra-index-url https://pypi.org/simple
python -m pip install -r requirments_cu118_py312.txtVerify Torch and CUDA:
python -c "import torch; print('torch:', torch.__version__); print('cuda:', torch.version.cuda); print('cuda available:', torch.cuda.is_available())"The three requirments_*.txt files define their package indexes at the top. Keep the default mirrors in mainland China. To use official indexes, replace only --index-url and --extra-index-url; keep package versions, CUDA suffixes, and the two-stage order unchanged.
| Default mirror | Official source |
|---|---|
https://mirrors.pku.edu.cn/pypi/simple |
https://pypi.org/simple |
https://mirrors.nju.edu.cn/pytorch/whl/cpu |
https://download.pytorch.org/whl/cpu |
https://mirrors.nju.edu.cn/pytorch/whl/cu118 |
https://download.pytorch.org/whl/cu118 |
https://mirrors.nju.edu.cn/pytorch/whl/cu128 |
https://download.pytorch.org/whl/cu128 |
The WebUI creates runtime directories automatically. Download models from the Hugging Face model repository and keep this layout:
assets/
├── hubert_base/
│ ├── config.json
│ ├── preprocessor_config.json
│ └── pytorch_model.bin
├── rmvpe/rmvpe.pt
├── pretrained/
├── pretrained_v2/
├── pymss_weights/
├── weights/ # user RVC .pth models
└── indices/ # user .index files
logs/
└── mute/ # training silence samples
# Exact paths used by the code
assets/hubert_base/config.json
assets/hubert_base/preprocessor_config.json
assets/hubert_base/pytorch_model.bin
assets/rmvpe/rmvpe.pt
assets/pretrained/*.pth
assets/pretrained_v2/*.pth
assets/pymss_weights/*
assets/weights/*.pth
assets/indices/*.index
logs/mute/*
python -m pip install --upgrade huggingface_hub
# Required for inference and feature extraction
hf download lj1995/VoiceConversionWebUI --revision main \
--include "hubert_base/*" --local-dir assets
hf download lj1995/VoiceConversionWebUI rmvpe.pt --revision main \
--local-dir assets/rmvpe
# Required for v1/v2 training
hf download lj1995/VoiceConversionWebUI --revision main \
--include "pretrained/*" "pretrained_v2/*" --local-dir assets
hf download lj1995/VoiceConversionWebUI mute.zip --revision main \
--local-dir .model-downloads
python -m zipfile -e .model-downloads/mute.zip logs
# Required only for pymss/MSST vocal separation
hf download lj1995/VoiceConversionWebUI --revision main \
--include "pymss_weights/*" --local-dir assetsWindows AMD/Intel DirectML environments additionally need:
hf download lj1995/VoiceConversionWebUI rmvpe.onnx --revision main \
--local-dir assets/rmvpeThe Ubuntu setup command above installs FFmpeg. On Windows, place these files in the repository root:
python webui.pyFor a headless Ubuntu server:
python webui.py --noautoopenThe default port is 7865. Put personal .pth models in assets/weights/ and .index files in assets/indices/.
- ContentVec
- VITS
- HIFIGAN
- Gradio
- FFmpeg
- Ultimate Vocal Remover
- pymss-project/pymss
- audio-slicer
- Vocal pitch extraction:RMVPE