Practical Deep Learning for Coders

A hands-on exploration of deep learning concepts following the Fast.ai course and accompanying book. This repository demonstrates practical image classification using fastai and PyTorch with different ResNet architectures.

What This Project Does

• Image Classification: Train models to recognize different categories (pets, landscapes, objects)
• Model Comparison: Compare ResNet18, ResNet50, and ResNet152 architectures
• Interactive Testing: Upload images to test trained models in real-time
• Hardware Optimization: Leverages Apple Silicon (MPS), CUDA, or CPU depending on availability

Notebooks

• 01_intro.ipynb - Main introduction with custom image classification and model comparison
• 01_intro-resnet18.ipynb - Pet breed classification using ResNet18
• 01_intro-resnet50.ipynb - Pet breed classification using ResNet50
• 01_intro-resnet152.ipynb - Pet breed classification using ResNet152

Features

• Multi-category image classification (cats, dogs, forests, mountains, cities)
• Pet breed recognition with 37 different breeds
• Interactive image upload for real-time testing
• Model architecture comparison across ResNet variants
• Hardware optimization for Apple Silicon (MPS), CUDA, or CPU

Prerequisites

• Python: 3.13.5 (managed via pyenv)
• Operating System: macOS (optimized for Apple Silicon) or Linux (for CUDA support)
• Hardware:
  • Apple Silicon Mac (for MPS acceleration)
  • NVIDIA GPU (for CUDA acceleration)
  • CPU-only systems supported

Installation

1. Clone the repository:

   git clone <repository-url>
   cd deep-learning-for-coders

2. Set up Python environment (using pyenv):

   pyenv install

3. Create and activate virtual environment:

   python -m venv .venv
   source .venv/bin/activate  # On macOS/Linux

4. Install dependencies:

   pip install -r requirements.txt

Usage

Quick Start

1. Launch Jupyter Notebook:

   jupyter notebook

2. Start with the main notebook: Open 01_intro.ipynb for a comprehensive introduction to the project

3. For specific model experiments:

   • 01_intro-resnet18.ipynb - Fast training with ResNet18
   • 01_intro-resnet50.ipynb - Balanced performance with ResNet50
   • 01_intro-resnet152.ipynb - Maximum accuracy with ResNet152

Resources

• Fast.ai Course - Free deep learning course
• Fast.ai Book - Comprehensive deep learning book

License

This project is for educational purposes. Please respect the licenses of the datasets and libraries used.