Free public beta until December 4, 2026: every course is open, no card needed. After that, the first chapter of each course stays free and full access is a one-time $39 ($19 for up to 100 founding members chosen for helpful feedback). Your progress is kept. How the beta works

deep-learning

Diagnosing a network that isn't learning

taught

Taught in

A fixed diagnostic order
Deep Learning / Neural Networks
Loss that won't decrease
Deep Learning / Neural Networks
Loss that explodes
Deep Learning / Neural Networks
Accuracy that's stuck
Deep Learning / Neural Networks
Checking labels and data leakage
Deep Learning / Neural Networks
Checking tensor shapes and gradients
Deep Learning / Neural Networks
Numerical instability
Deep Learning / Neural Networks

Requires

vanishing-exploding-gradients required
overfitting-and-underfitting required

Required by

Overfitting a tiny batch as a first diagnostic required
The end-to-end model decision chain required

Local knowledge map

Local knowledge map

Click to expand

Solid arrows are required prerequisites; dashed are recommended or helpful. Lower rows are more foundational. Click any box to jump to that concept.