C9
Neural Network Fundamentals
4 lessons3 hoursPlannedFree
Outcome
Build and train a multilayer perceptron from scratch in PyTorch.
Prerequisites
- Recommended: C1 Linear Algebra and Probability for ML
- Recommended: C4 Calculus for Backpropagation
Lessons
- L01Perceptrons and activations · coming soonExercise: Implement ReLU and GELU and compare gradients.
- L02Initialization (Xavier, Kaiming) · coming soonExercise: Observe activation variance across depth.
- L03Normalization (BatchNorm, LayerNorm) · coming soonExercise: Compare training with and without.
- L04Training loop in PyTorch · coming soonExercise: Train an MLP on MNIST.