Deep Learning by Code: Build Neural Networks from Scratch

By Thompson Carter

Build Neural Networks from the Ground Up-No Black Boxes, Just Code

If you've ever wanted to truly understand how deep learning works-not just use high-level libraries-Deep Learning by Code is your roadmap.

This hands-on guide teaches you how to implement the core building blocks of neural networks from scratch using pure Python and NumPy. You'll code every layer, activation function, and optimization algorithm yourself-gaining a deep, intuitive understanding of how modern AI really works under the hood.

Whether you're a developer, data science student, or aspiring machine learning engineer, this book gives you the confidence and skills to create and experiment with your own deep learning models-line by line.

Inside You'll Learn:

How neural networks process information, learn patterns, and make predictions

Step-by-step construction of forward and backward propagation

Implementing activation functions, loss functions, and optimizers by hand

Building feedforward, convolutional, and recurrent neural networks

Understanding gradient descent, backpropagation, and weight updates

Creating training loops without relying on frameworks

Visualizing training behavior and model accuracy

Transitioning from raw code to real-world applications

Bonus: Compare your custom models to PyTorch and TensorFlow

This isn't just another tutorial-it's a deep dive into the mechanics of deep learning. You'll leave not just knowing how, but why.

• Format: Paperback
• Language: English
• ISBN: B0FG1DF1H9
• ISBN13: 9798290086538
• Release Date: June 2025
• Publisher: Independently Published
• Length: 240 Pages
• Weight: 0.72 lbs.
• Dimensions: 0.5" x 6.0" x 9.0"