Complete course notes from CMPUT 328. Each lesson includes HTML study guides and Anki flashcard decks.
Click a topic to study →
Linear + logistic regression on MNIST with optimizer notes and evaluation flow. Covers gradient descent, cross-entropy loss, and model evaluation metrics.
FFNN architectures, CIFAR-10 training schedule, and regularization checklist. Deep dive into multi-layer perceptrons and activation functions.
Losses, optimizers, derivatives, and debugging flow for assignment prep. Essential mathematics and computational techniques for neural networks.
Convolutions, augmentation, training pitfalls, and evaluation playbook. Understanding kernels, pooling, and convolutional architectures.
Patch embeddings, attention math, and low-rank adaptation recipes. Modern transformer architectures applied to computer vision tasks.
R-CNN to YOLO, anchor math, IoU metrics, and MNISTDD-RGB tuning tips. Complete guide to detecting and localizing objects in images.
FCNs, U-Net, Mask R-CNN, loss functions, and dense prediction metrics. Pixel-level classification and instance differentiation techniques.
Classical AE, VAE, VQ-VAE, gradient tricks, and reconstruction diagnostics. Generative models and unsupervised representation learning.
GANs architecture, training dynamics, generator-discriminator interplay, loss functions, optimization strategies, and modern variants. Deep dive into adversarial training and practical applications.
Forward/reverse diffusion processes, denoising techniques, score matching, DDPM/DDIM architectures, conditioning methods, and state-of-the-art implementations in generative image synthesis.