Deep Learning Concepts
Educational deep learning reference — algorithms, concepts, and common problems
Algorithms End-to-end implementations of Transformer, Diffusion, GANs, VAE, Q-Learning, Policy Gradient, and more — each with line-by-line annotations.
Maths and Stats Concepts Activation functions, loss functions, regularisation, optimisation primitives, information theory, and the mathematical tricks that make training work.
Common Problems Vanishing gradients, mode collapse, overfitting, reward hacking, numerical instability — what goes wrong and why.