CHAPTER 31 / 40
Deep Learning in Practice
3 lessons · Explore, experiment and understand.
DOJO PETYour little coding companion
DOJO / YOUR GUIDE
Let’s figure it out.
Ask me about this course or where to start.
Chapter lessons
3 lessons
- Step 131~15m3D · 113★Most cinematic
Backpropagation & Reverse-Mode Autodiff
Training a neural network reduces to one question repeated millions of times: how should I nudge each parameter to make the loss smaller? Th…
Open lesson - Step 132~15m3D · 113★Most cinematic
From SGD to Adam: Gradient-Descent Optimizers
Training a model means minimizing a loss over parameters . The workhorse is gradient descent: take a step opposite the gradient, , where is…
Open lesson - Step 133~15m3D · 113★Most cinematic
Regularization and Generalization
The central problem of learning is generalization: we minimize a training loss but care about the expected loss on unseen data, . We only ev…
Open lesson