MEPX
Chapter 10 of 10All chapters
  1. 1The artificial neuron
  2. 2Layers and depth
  3. 3How it learns
  4. 4Activation functions
  5. 5Architectures
  6. 6Keeping it general
  7. 7Training in practice
  8. 8Transfer learning
  9. 9What they cannot do
  10. 10Learning it yourself

Chapter 10 of 10

Learning it yourself

A practical path.

Order of study

Build a small network by hand on a simple dataset before touching a framework. Understanding one training loop end to end explains more than reading about ten architectures.

  • PyTorch and TensorFlow dominate; either is fine to learn first.
  • A laptop is enough for the fundamentals; rented hardware handles the rest.

Where to be careful

The field moves fast and the fundamentals move slowly. Gradient descent, loss functions and generalisation have not changed, and they explain most of what new architectures do.