Lesson 3 of 14 · Foundations

Training neural networks well

Losses, SGD, momentum and Adam, learning-rate schedules, initialisation, dropout, batch normalisation, weight decay and early stopping.

15 min read