Music-Speech-Separation

ARCHITECTURE USED FOR TRAINING

Usage

1. Download the pre-trained model here

Pretrained Python Code

Python Code is used. Change the paths as required.

python MSClassifierR1.py

Using colab or jupyter

Notebook can be used to split the uploaded file into music and speech components. Either upload a folder or a single file.

Training Purposes

1. STFT based
2. Chroma STFT based

Dataset

Download the used dataset here.

License

This project is licensed under the MIT License - see the LICENSE file for details.