This repository provides lecture materials for the INFO8010 Deep Learning course at ULiège, covering fundamental and advanced topics in deep learning. It is intended for students and researchers seeking a structured curriculum with accompanying code examples and assignments. The materials offer a comprehensive overview of deep learning concepts, from basic neural networks to modern architectures like transformers and diffusion models.

How It Works

The course material is delivered through PDF lectures and accompanying code examples, primarily using the PyTorch library. The curriculum progresses logically, starting with machine learning fundamentals and multi-layer perceptrons, then delving into automatic differentiation, training techniques, and specialized architectures like CNNs, GNNs, and LLMs. The provided code snippets illustrate key concepts and serve as a basis for practical assignments.

Quick Start & Requirements

• Installation: PyTorch installation instructions are provided within the "Homeworks" folder.
• Prerequisites: PyTorch library, Python. GPU usage is mentioned for Homework 2.
• Resources: No specific resource estimates are provided, but typical PyTorch development environments are assumed.
• Links:
  • Discord: https://discord.gg/5yZqTZhXFW
  • Previous Editions: https://github.com/glouppe/info8010-deep-learning/tree/main/archives

Highlighted Details

• Comprehensive lecture schedule covering Spring 2025 topics.
• Includes optional homework assignments with PyTorch focus.
• Features code examples for various deep learning concepts.
• Offers archived lectures and videos on older topics like RNNs and GANs.

Maintenance & Community

The course is led by Gilles Louppe, with teaching assistants François Rozet, Yann Claes, and Victor Dachet. A Discord server is available for community interaction and support.

Licensing & Compatibility

The repository content is not explicitly licensed in the README. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The repository primarily serves as educational material and does not appear to be a production-ready library. Specific version requirements for dependencies are not detailed, and the "optional" nature of homeworks means practical application may vary among students.