A comprehensive, structured deep learning course using Keras and TensorFlow for developers who want to master AI and deep learning.

This repository contains a complete, progressive deep learning curriculum with 70+ Jupyter notebooks (35+ each for Keras and PyTorch) organized into 10 progressive modules, from beginner fundamentals to advanced research topics.

Choose your preferred deep learning framework:

• Keras/TensorFlow - High-level API, rapid prototyping, production deployment
• PyTorch - Flexible, research-friendly, explicit control over training

Both implementations teach identical concepts and have the same learning objectives!

• 01_Basics - CNN fundamentals and regression tasks
• 02_Image_Classification - Image classification from scratch
• 03_Advanced_CNN - Advanced architectures (DenseNet, etc.)
• 04_Regularization - Preventing overfitting and improving generalization
• 05_Transfer_Learning - Leveraging pretrained models
• 06_Image_Segmentation - Pixel-level predictions and medical imaging
• 07_Time_Series - Sequence modeling and forecasting
• 08_NLP_Text - Natural language processing and text classification
• 09_Advanced_Topics - Physics-informed neural networks, PDEs, real-time video AI

→ Framework Documentation:

• Keras Version: See Keras/README.md
• PyTorch Version: See PyTorch/README.md

• Python 3.7 or higher
• pip or conda

# Clone the repository
git clone https://github.com/yourusername/goodplacedeeplearning.git
cd goodplacedeeplearning

# Install dependencies
pip install -e .

Or with conda:

conda create -n keras-course python=3.9
conda activate keras-course
pip install -e .

For Keras/TensorFlow:

# Navigate to the Keras folder
cd Keras

# Launch Jupyter
jupyter notebook

# Start with 01_Basics

For PyTorch:

# Navigate to the PyTorch folder
cd PyTorch

# Launch Jupyter
jupyter notebook

# Start with 01_Basics

• RAM: 8GB minimum (16GB recommended for transfer learning)
• GPU: Optional but recommended (NVIDIA GPU with CUDA support)
• Storage: 10GB free space for datasets

• NumPy - Numerical computing
• Pandas - Data manipulation
• Matplotlib & Seaborn - Visualization
• Scikit-learn - Machine learning utilities
• Jupyter - Interactive notebooks
• Pillow - Image processing
• h5py - HDF5 support

For Keras:

• TensorFlow/Keras 2.10+

For PyTorch:

• PyTorch >= 1.9.0
• TorchVision - Pretrained models and image transforms
• TorchAudio (optional)

All dependencies are specified in setup.cfg and will be installed automatically with pip install -e .

1. 01_Basics - Learn CNN fundamentals
2. 02_Image_Classification - Classic image tasks
3. 03_Advanced_CNN - Deeper architectures
4. 04_Regularization - Improve your models

1. Complete beginner path first
2. 05_Transfer_Learning - Efficient learning from pretrained models
3. 06_Image_Segmentation - Advanced vision tasks
4. 07_Time_Series - Sequential data

1. Complete beginner and intermediate paths
2. 08_NLP_Text - Natural language understanding
3. 09_Advanced_Topics - Research-level topics

• Deep Learning Fundamentals

  • Convolutional Neural Networks (CNNs)
  • Recurrent Neural Networks (RNNs)
  • Attention mechanisms

• Computer Vision

  • Image classification
  • Semantic segmentation
  • Object detection (implicit)
  • Medical image analysis

• Natural Language Processing

  • Text classification
  • Sentiment analysis
  • Sequence models

• Advanced Techniques

  • Transfer learning and fine-tuning
  • Regularization and dropout
  • Batch normalization
  • Physics-informed neural networks
  • Fourier Neural Operators

• Practical Applications

  • Real-time video processing
  • Time series forecasting
  • Model deployment

• Keras Notebooks: 35+
• PyTorch Notebooks: 18 (Keras → PyTorch conversions)
• Total Notebooks: 53+
• Total Modules: 10 (identical across frameworks)

1. Run every cell - Don't just read, execute and experiment
2. Modify code - Change hyperparameters and observe results
3. Visualize - Pay attention to plots and interpretations
4. Take notes - Jot down key concepts and insights
5. Challenge yourself - Try to solve problems before looking at solutions
6. GPU acceleration - Use GPU for faster training if available

For development and contributions:

# Clone and install in editable mode
git clone https://github.com/yourusername/goodplacedeeplearning.git
cd goodplacedeeplearning
pip install -e ".[dev]"

# Run linters and formatters
black Keras/
flake8 Keras/

Each notebook follows the pattern: NN_notebook_name.ipynb

• NN = Lesson number within the module
• notebook_name = Descriptive name of the content

Example: 00_fashion_mnist_basic_cnn.ipynb

• Contributing Guide - How to contribute to this project
• Notebook Index - Complete index of all notebooks
• PyTorch README - PyTorch-specific documentation
• PyTorch Conversion Summary - Keras to PyTorch conversion details
• Conversion Complete - Full conversion documentation

Contributions are welcome! Please see docs/CONTRIBUTING.md for details.

1. Fork the repository
2. Create a new branch for your changes
3. Commit with clear messages
4. Push and create a pull request

This course material is provided for educational purposes.

This course was designed to provide a comprehensive, hands-on introduction to deep learning using Keras and TensorFlow. It progresses from fundamental concepts to cutting-edge research applications, making it suitable for:

• Students learning deep learning for the first time
• Practitioners looking to expand their skills
• Researchers exploring advanced techniques
• Developers building production ML systems

• Issues: Report bugs and request features on GitHub
• Discussions: Ask questions in GitHub Discussions
• Documentation: See docs/ folder and framework READMEs (Keras, PyTorch)

• TensorFlow Official Documentation
• Keras Documentation
• NVIDIA Modulus (for PDE notebooks)
• DeepStream SDK

Happy Learning! 🚀

Start with Keras/01_Basics and work your way through the course at your own pace.