Machine Learning for Sound and Music

Overview

This repository contains materials for the second half of the "Machine Learning for Sound and Music" course. The course introduces key concepts of machine learning applied to audio and music processing. The materials include lecture notebooks, practical exercises, and examples to help students learn and implement machine learning techniques effectively.

Repository Structure

• notebooks/: Contains Jupyter notebooks for each lecture.
• docker/: Includes Docker files to build the image and run the Jupyter server.

Getting Started

Prerequisites

• Docker and Docker Compose installed on your machine.
• Basic familiarity with Python and Jupyter Notebooks.
• Recommended: A laptop with a GPU for deep learning tasks.

Installation

1. Clone the repository:

   git clone <repository-url>
   cd machine_learning_for_sound_and_music

2. Build the Docker image:

   docker-compose build

3. Start the Jupyter server:

   docker-compose up

4. Access the Jupyter server: Open your browser and navigate to http://localhost:8888.

Course Content

Lecture Index

• Lecture 5: Supervised Learning with CNNs. Code for training small MLP and CNN models for a multi-class task.
• Lecture 6: Self-Supervised Learning I: Autoencoders
• Lecture 7: Self-Supervised Representation Learning II: Metric Learning
• Lecture 8: Transformers and the Attention Mechanism
• Lecture 9: Transfer Learning
• Lecture 10: Multi-modal Learning: Text-Audio Models

Contribution

Contributions are welcome! Please fork the repository and submit a pull request for any improvements or suggestions.

References

• mirdata: Dataset loader used in the course.
• PyTorch: Preferred library for deep learning tasks.
• Essentia: Library for audio processing.

License

This repository is licensed under the MIT License. See the LICENSE file for details.