uvr

📦 Check out my new library called Vocal for Vocal/Speech Separation built with simplicity in mind.

---

Ultimate Vocal Remover CLI

[Colab]

⚠️ Before running this project, make sure you have installed torch, torchaudio. Please check out the PyTorch documentation.

⚠️ Also make sure you have libsndfile and ffmpeg installed.

⚠️ This project currently works on CUDA.

Install dependencies

pip install -r requirements.txt

Download Model Weights

./download.sh

Separation

python separate.py

Modify these lines to fit your needs

if __name__ == '__main__':
    device = 'cuda'
    is_half=True
    model_path='uvr5_weights/2_HP-UVR.pth'
    pre_fun = _audio_pre_(model_path=model_path,device=device,is_half=True)
    audio_path = 'audio.aac'
    save_path = 'opt'
    pre_fun._path_audio_(audio_path , save_path,save_path)