AIAI — Deep Generative Models Coursework Repository

This repository contains coursework, answers, descriptions, and solutions for the AIAI / Deep Generative Models university course.

Overview

The project is organized by homework number and includes:

• assignment descriptions,
• working answers (LaTeX and notebooks),
• final solutions,
• and supporting course slides.

Primary goals of this repository:

• keep coursework versions organized,
• track progress across homeworks,
• preserve reproducible experiments and writeups.

Repository Structure

AIAI/
├── HomeWorks/
│   ├── HW1/
│   │   ├── Answers/
│   │   ├── description/
│   │   └── Solutions/
│   ├── HW2/
│   │   ├── Answers/
│   │   ├── description/
│   │   └── Solutions/
│   ├── HW3/
│   │   ├── Answers/
│   │   ├── description/
│   │   └── Solutions/
│   ├── HW4/
│   │   ├── Answers/
│   │   ├── description/
│   │   └── Solutions/
│   └── HW5/
│       ├── description/
│       └── Solutions/
├── Slides/
└── README.md

Homework Layout Convention

For each homework folder:

• description/: original assignment statement and starter files
• Answers/: your in-progress and submitted answers
• Solutions/: polished/final solutions or reference implementations

Tech Stack

• LaTeX for theoretical writeups (.tex, .pdf)
• Jupyter Notebook (.ipynb) for practical/modeling tasks
• Python ecosystem for experiments (inside notebooks)

Quick Start

1) Clone

git clone <your-repo-url>
cd AIAI

2) Open in VS Code

code .

3) Work on a specific homework

cd HomeWorks/HW1/Answers

LaTeX Build

Example for HW1 answer sheet:

cd HomeWorks/HW1/Answers
pdflatex -interaction=nonstopmode -halt-on-error main.tex

Run twice if table of contents, references, or numbering are not stable.

Notebook Workflow

Recommended steps for notebook-based assignments:

1. Create/select Python virtual environment.
2. Install core packages (jupyter, numpy, pandas, matplotlib, torch, etc. as needed).
3. Open notebooks from Answers/ and save final versions in Solutions/ when complete.

Course Progress Snapshot

• HW1: LaTeX-based theoretical solutions in progress
• HW2: VAE/iTransformer notebooks and theory notebooks included
• HW3: GAN/Flow practical and theoretical notebooks included
• HW4: complete structure initialized with assignment and solution files
• HW5: LaTeX description/solution files initialized

Version Control Guidelines

• Commit in small logical chunks.
• Keep generated artifacts out of Git when possible.
• Prefer clear commit messages, e.g.:
  • hw2: complete VAE latent interpolation section
  • hw1: fix partition function derivation

Contributing

Contributions are welcome for structure, documentation quality, and reproducibility improvements.

Please read CONTRIBUTING.md before opening a pull request.

License

This project is licensed under the MIT License.

See LICENSE for details.

Maintainer

• Taha Majlesi

---

If you want, this README can be further tailored with semester-specific grading rules, exact package versions, and reproducibility commands for each notebook.