This repository provides a curated collection of single-file, dependency-free Python implementations of modern AI algorithms. It aims to demystify complex machine learning concepts for ML engineers, students, and researchers by offering executable code that bridges the gap between high-level framework calls and dense academic notation, enabling a deeper understanding of AI's inner workings.

How It Works

The core approach involves creating self-contained Python scripts, each dedicated to a single algorithm. These scripts are designed to be runnable out-of-the-box, requiring only Python's standard library. They include both the training loop and inference capabilities, demonstrating the full algorithm lifecycle. This "naked algorithm" philosophy prioritizes clarity and understanding over performance optimization, making complex AI building blocks accessible and executable on standard CPU hardware within minutes.

Quick Start & Requirements

• Primary install/run command: Clone the repository (git clone https://github.com/Mathews-Tom/no-magic.git), navigate into the directory (cd no-magic), and execute scripts directly (e.g., python 01-foundations/microgpt.py).
• Non-default prerequisites: Python 3.10+, 8 GB RAM, and any modern CPU (2019-era or newer). Small datasets are downloaded automatically on first run.
• Links: CONTRIBUTING.md for contribution guidelines.

Highlighted Details

• Tier 01 (Foundations) covers core algorithms like GPT, BERT, CNN, GAN, VAE, diffusion, RAG,