Artificial IntelligenceIntermediateOne To One
Deep Learning and Modern AI
Neural networks, transformers and large language models, built and understood from the inside.
About this programme
Twelve sessions on how modern AI systems work. You build a neural network from scratch, train convolutional and transformer models in PyTorch, and finish by building a small application on top of a large language model. Every session ends with working code you wrote yourself, and you explain back what it does before we move on.
What you will be able to do
- Explain how a neural network learns, step by step
- Train and debug convolutional and transformer models in PyTorch
- Use pretrained models and fine tune them for your own task
- Build and evaluate an application that uses a large language model
Curriculum
4 modules · 12 lessons
- A neuron, a layer and backpropagation by hand1h
- Training loops, loss functions and optimisers in PyTorch1h
- Regularisation, dropout and batch normalisation1h
Before you start
- Python, including NumPy
- Basic machine learning: training, validation and overfitting