DenisSud/ai-lecture-materials
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
No description
2 commits 1 branch 0 tags 209 KiB
  • Python 63.6%
  • HTML 35.5%
  • Nix 0.9%
Find a file
Exact
DenisSud 3c6662682e Add slides: The Foundation of Modern AI 
HTML presentation deck with embedded assets covering:
- What is AI (everyday examples, common systems)
- Classical programming vs ML (explicit engineering vs learned patterns)
- Model architecture + parameters
- Loss functions, gradient descent, training loop
- Interactive demo section (links to notebooks)
2026-07-21 03:54:23 +03:00
layouts Initial commit: organize AI lecture materials 2026-07-21 03:43:12 +03:00
notebooks Initial commit: organize AI lecture materials 2026-07-21 03:43:12 +03:00
slides Add slides: The Foundation of Modern AI 2026-07-21 03:54:23 +03:00
.gitignore Initial commit: organize AI lecture materials 2026-07-21 03:43:12 +03:00
.python-version Initial commit: organize AI lecture materials 2026-07-21 03:43:12 +03:00
devenv.lock Initial commit: organize AI lecture materials 2026-07-21 03:43:12 +03:00
devenv.nix Initial commit: organize AI lecture materials 2026-07-21 03:43:12 +03:00
devenv.yaml Initial commit: organize AI lecture materials 2026-07-21 03:43:12 +03:00
pyproject.toml Initial commit: organize AI lecture materials 2026-07-21 03:43:12 +03:00
README.md Initial commit: organize AI lecture materials 2026-07-21 03:43:12 +03:00
uv.lock Initial commit: organize AI lecture materials 2026-07-21 03:43:12 +03:00

README.md

AI Lecture Materials

Interactive AI/ML lecture notebooks built with Marimo — a reactive Python notebook that doubles as a deployable app.

Notebooks

Notebook Description
notebooks/gradient-descent.py Interactive gradient descent demo on linear regression with a 3D loss landscape, ChartPuck starting-point picker, and playable step/run controls.
notebooks/titanic-survival.py Hands-on ML experiment: train a small neural network on the Titanic dataset with adjustable data size, features, learning rate, and model complexity.
notebooks/greedy-search.py Population-based greedy search visualisation on 2D loss landscapes (Rastrigin, Ackley, Rosenbrock, etc.) using ChartPuck.
notebooks/chartpuck-demo.py Quick reference / example of the ChartPuck draggable overlay widget API.

Quick start

# Install marimo and dependencies
uv sync

# Run a notebook
uv run marimo run notebooks/gradient-descent.py

# Or edit it
uv run marimo edit notebooks/gradient-descent.py

Development

This project uses devenv + uv for a reproducible environment with CUDA support (for optional JAX acceleration).

devenv shell    # activates the environment
marimo edit     # starts the marimo editor

Project structure

├── notebooks/           # Marimo notebooks
├── layouts/             # Notebook layout configurations
├── pyproject.toml       # Project metadata and dependencies
├── devenv.nix           # Declarative dev environment (Nix)
├── devenv.yaml          # Devenv inputs
└── uv.lock              # Locked dependency versions