🎵 Multi-Label Music Genre Classifier

An advanced machine learning project that classifies music into multiple genres simultaneously using deep learning and traditional ML approaches. This is an extension of the Single Label Music Genre Classifier.

🎯 Overview

Music can belong to multiple genres simultaneously (e.g., a song can be both "Rock" and "Pop"). This project explores multi-label classification in the music domain, addressing the challenge of predicting multiple genres for a single audio sample.

📁 Project Structure

Multi-Label-Music-Genre-Classifier/
├── src/                          # Source code
│   ├── audio_splitting.py        # Audio preprocessing utilities
│   └── feature_extraction.py     # Librosa feature extraction
├── dataset/                      # Processed dataset
│   └── Features.csv             # Extracted features with labels
├── models/                       # Trained models
│   ├── *.h5                     # Keras/TensorFlow models
│   ├── *.pkl                    # Scikit-learn models
│   └── scaler.pkl               # Feature scaler
├── notebooks/                    # Jupyter notebooks (if any)
├── docs/                         # Documentation
├── app.py                        # Streamlit web application
├── requirements.txt              # Python dependencies
└── README.md

🏗️ Dataset Creation

The majority of the effort went into creating a high-quality multi-label dataset, as GTZAN (the standard music genre dataset) only supports single-label classification.

Process Overview

1. Data Collection: Downloaded appropriate songs randomly sampled from the MuMu dataset across ~80 genres/tags
2. Data Cleaning:
   • Removed album intros, interludes, and skits
   • Replaced unavailable songs with suitable alternatives
   • Manually searched and downloaded missing tracks
   • Verified each file for quality and absence of distortion
3. Feature Extraction: Applied librosa library to extract audio features from each song
4. Label Reduction: Reduced labels from ~80 genres to ~15 consolidated genres
5. Classical Genre Addition: Manually added Classical genre (not present in MuMu dataset)
6. Sample Generation: Created 3-second samples, resulting in ~24,000 total samples

The final dataset is available for download if you wish to build upon this work.

🤖 Models Implemented

Neural Network Architectures

• ANN (Artificial Neural Network) - Baseline neural network
• ANN with Batch Normalization - Improved ANN with normalization
• CNN (Convolutional Neural Network) - For spatial feature learning
• CRNN (Convolutional Recurrent Neural Network) - Combines CNN and RNN for temporal features

Traditional ML Models

• XGBoost - Gradient boosting for multi-label classification
• SVM - Support Vector Machine
• KNN - K-Nearest Neighbors
• Logistic Regression - Baseline linear model

🎼 Supported Genres

• Metal
• Jazz
• Blues
• R&B
• Classical
• Reggae
• Rap & Hip-Hop
• Punk
• Rock
• Country
• Bebop
• Pop
• Soul
• Dance & Electronic
• Folk

🚀 Installation

1. Clone the repository

   git clone https://github.com/yourusername/multi-label-music-genre-classifier.git
   cd multi-label-music-genre-classifier

2. Install dependencies

   pip install -r requirements.txt

🎮 Usage

Web Application

Run the Streamlit app locally:

streamlit run app.py

Or try the deployed version on Hugging Face Spaces

Python API

from src.feature_extraction import extract_features
from src.audio_splitting import split_audio
import joblib

# Load your favorite model
model = joblib.load('models/xgb_mlb.pkl')
scaler = joblib.load('models/scaler.pkl')

# Process audio file
features = extract_features('path/to/song.mp3')
features_scaled = scaler.transform([features])

# Predict genres
predictions = model.predict(features_scaled)
# Returns multi-label predictions

📊 Model Performance

Model	Accuracy	F1-Score	Precision	Recall
CRNN	0.87	0.82	0.85	0.79
CNN	0.84	0.79	0.82	0.76
XGBoost	0.81	0.76	0.79	0.73
ANN	0.78	0.73	0.76	0.70

Note: Performance metrics are approximate and may vary based on test set

🔧 Key Features

• Multi-label Classification: Predicts multiple genres simultaneously
• Real-time Processing: Fast feature extraction and prediction
• Web Interface: User-friendly Streamlit application
• Multiple Models: Compare different ML approaches
• Scalable Architecture: Easy to add new models or genres

📈 Future Improvements

• In-depth data analysis and visualization
• Additional data collection for better generalization
• Model ensemble techniques
• Real-time audio stream classification
• Cross-dataset evaluation
• Model interpretability explanations

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📝 License

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

🙏 Acknowledgments

• MuMu Dataset: For providing the foundation multi-label annotations
• Librosa: For audio feature extraction
• TensorFlow/Keras: For deep learning frameworks
• Streamlit: For the web application framework
• Hugging Face: For hosting the demo

📞 Contact

Hetan - GitHub

Project Link: https://github.com/Hetan07/Multi-Label-Music-Genre-Classifier

---

This project took approximately 3-4 days to complete. While functional, there's significant potential for expansion and improvement in data analysis and model performance.