Build a Diffusion Model That Draws Digits
GANs vs. Diffusion: How Image Generation Actually Works
Variational Autoencoders: Sampling From the Latent Space
Autoencoders: Compression as a Learning Signal
Deeper: Pooling, Batch Norm, and Residual Connections
Your First CNN: Beating the Dense Net on Digits
What Convolution Actually Does to an Image
Why Your Network Won’t Train: Init, Activations, and Regularization
Build a Digit Recognizer From Scratch in NumPy
Backprop by Hand: Deriving the Gradients You Keep Importing
The XOR Problem: Why One Layer Was Never Enough
What a Perceptron Actually Is