After this track you can Understand backpropagation as the chain rule, not as magic Train deep networks with the right optimizer and regularisation Build and train a convolutional network for images in PyTorch Explain attention and why transformers replaced RNNs Deep learning on proper foundations: what a single neuron computes, why stacking them bends the decision boundary, backpropagation as the chain rule, training deep networks without them blowing up, PyTorch, convolutional networks and transformers. The first lesson is free — build your own network right inside it and watch it learn.
Read a free lesson: What a Neural Network Actually Computes
What you'll learn, module by module 6 modules · 21 lessons Every module and lesson in this track. Lessons marked free are open to everyone; the rest open once you own the track.
Every lesson in this track, with summaries 01 What a neural network actually computes, why the nonlinearity is the whole trick, and how the weights learn.
1 free lesson 4 lessons 1 What a Neural Network Actually Computes Free 2 Activation Functions: ReLU, GELU and Friends 3 Backpropagation Without the Fear 4 Your First Training Loop in PyTorch 02 Optimizers, learning-rate schedules, initialisation, normalisation and regularisation — what makes training actually converge.
4 lessons 1 Optimizers: SGD, Momentum, Adam 2 Learning-Rate Schedules and Warmup 3 Initialisation and Normalisation 4 Regularisation: Dropout, Weight Decay, Early Stopping 03 Tensors, autograd, nn.Module, DataLoader and GPU training — enough PyTorch to work on your own.
4 lessons 1 Tensors and Autograd 2 Building Models with nn.Module 3 Datasets and DataLoaders 4 GPUs and Mixed Precision 04 Convolution, pooling and CNN architectures — why weight sharing makes images learnable.
3 lessons 1 How Convolutions Work 2 CNN Architectures: from LeNet to ResNet 3 Transfer Learning 05 From RNNs and LSTMs to self-attention — the architecture behind every modern LLM.
3 lessons 1 RNNs and LSTMs 2 The Attention Mechanism 3 The Transformer Architecture 06 Debugging training, tracking experiments, and one capstone from raw data to an evaluated model.
3 lessons 1 Debugging Neural Networks 2 Tracking Experiments 3 Capstone: an Image Classifier
Try a simulation from this track The real one, from a lesson — it runs in your browser. No account needed.
Try it Neural network playground
Save with a career-path bundle This track is part of a larger path. Buy the whole path as a bundle and every track in it costs 40% less.
Career-path bundle 40% off
1 Python for Data Science 2 Data Analytics: Statistics & Data Visualization 3 Introduction to Machine Learning 4 Introduction to Deep Learning & Neural Networks 5 MLOps: Deploying Machine Learning & Deep Learning Models Career-path bundle 40% off
1 Python for Data Science 2 Introduction to Deep Learning & Neural Networks 3 LLMs, Prompt Engineering & Context Engineering 4 Building AI Agents: Harness Engineering 5 MLOps: Deploying Machine Learning & Deep Learning Models 6 Computer Vision with Deep Learning 7 Natural Language Processing (NLP) 
Python
PyTorch
NumPy