Y

Deep learning · Intermediate · 6 h

Neural networks and the training loop

Tensors, layers, loss, gradients and optimisers, written as an explicit PyTorch training loop.

Why employers care: Interviews ask you to explain backpropagation and write a training loop. Frameworks hide it until something goes wrong.

In the roadmaps for: Data Scientist, AI/ML Engineer

Lessons, in order

  1. A neural network and its training loop in PyTorchRecognise handwritten digits with a two-layer network, and see every step of training instead of a magic fit().3 min

Next shelf: Training that generalises