The Network Room
Seven short books that open the machine up: the forward pass worked by hand, the shapes that cause most bugs, convolution on a printable grid, attention on four tokens, the training failures and their causes, transfer learning as a decision, and the step from a notebook to something that runs.
The books
Each one stands alone and each costs the same. Start with the arithmetic if you want the inside of a network; start with the shape bugs if you have one failing right now.
Forty steps, each one doable by hand before you write a line of code.
$10.00 · 40 entries · 6 chapters
Thirty-eight shape failures, each with the error text and the fix.
$10.00 · 38 entries · 5 chapters
Thirty-four steps from one filter on one patch to a stack that recognises shapes.
$10.00 · 34 entries · 5 chapters
Thirty-two steps from one recurrent step to attention, on sequences you can follow.
$10.00 · 32 entries · 5 chapters
Forty-two training failures, each with the curve, the cause and the fix.
$10.00 · 42 entries · 6 chapters
Thirty steps on what to freeze, what to replace, and how little data you need.
$10.00 · 30 entries · 5 chapters
Thirty-six steps from cell 47 to something that runs and keeps running.
$10.00 · 36 entries · 5 chapters
This is educational material about how neural networks work. It is not a course, a certification, a bootcamp or a career programme, and it makes no promise about employment, salary or professional outcomes of any kind. The code and the worked figures are teaching examples, written for clarity rather than for production: read them, adapt them, and test anything you reuse. Library APIs change often, so the method is what carries over, not the exact call signature. Nothing here is professional, legal or financial advice. Models reproduce the patterns and the biases of the data they are trained on; deploying one that affects people carries responsibilities this book does not cover.