Atlas

Foundations

Training & Optimization

Backpropagation, gradients, and the descent to a good model.

You can jump straight in, but this star assumes Neural Networks. Not recommended as a first stop.

Four layers of depth

Each layer ends with a quiz. Finish layer 4 and you own this concept.

  1. L1IntuitionWhat 'training a model' actually means: nudging parameters downhill until predictions get good.5m
    40 XP
  2. L2MechanicsGradient descent variants, learning rate schedules, and the practical knobs that determine whether training succeeds.8m
    70 XP
  3. L3CodeImplementing SGD and Adam from scratch, then writing a full PyTorch training loop.12m
    110 XP
  4. L4FoundationsThe convex optimization theory behind gradient descent, convergence guarantees, and why loss landscapes for neural nets are so much harder.15m
    180 XP

Where this leads

33 stars in the atlas.