Deep Learning Fundamentals

An introduction to deep learning for programmers, focused on practical implementation with PyTorch.

Learning Path

Main Track

Work through these notebooks in order:

#	Notebook	Description
1	Introduction to Neural Networks	History, perceptron, MLPs, activation functions, training concepts, architectures overview
2	PyTorch Basics	Tensors, GPU usage, nn.Module, nn.Sequential, building models
3	Your First Neural Network	End-to-end FashionMNIST classification: data loading, model, training loop, evaluation
4	Training Dynamics	Loss functions, optimizers, regularization (L2, Dropout), learning rate scheduling

Complementary Material

Optional deep-dives referenced from the main track:

Resource	Description
XOR & Linear Separability	Why linear models fail on XOR, feature engineering vs. representation learning, universal approximation theorem
Gradient Descent from Scratch	ADALINE in NumPy, 3D cost landscape, comparison with PyTorch autograd
TensorFlow Playground Guide	Hands-on experiments with TF Playground + equivalent PyTorch code

Prerequisites

• Python programming experience
• Basic linear algebra (vectors, matrices)
• PyTorch installed (pip install torch torchvision)