PART 1
0/9Softmax, Loss, and Metrics Primitives
Build the numerically stable softmax, cross-entropy loss, and accuracy helpers used throughout the network.
Build a Trainable CNN from Scratch
Build a Trainable CNN from Scratch
Assemble a LeNet-style convolutional network with im2col convolutions, gradients, Adam, and a complete training loop.
Every source step is its own lesson with intuition, concepts, correctly rendered MathJax mathematics, implementation, tests, mistakes, and a checkpoint.
PART 2
0/7Initialization and Convolution Plumbing
Implement He initialization, zero biases, padding, output-shape math, and the im2col / col2im transforms that power efficient convolutions.
PART 3
0/17Layer Forward and Backward Passes
Code the forward and backward routines for convolution, max pooling, ReLU, flatten, and linear layers.
017conv2d forward+5 XP018conv2d grad input+5 XP019conv2d grad weights+5 XP020conv2d grad bias+5 XP021conv2d backward+5 XP022maxpool2d forward+5 XP023scatter grad window+5 XP024maxpool2d backward+5 XP025relu forward+5 XP026relu backward+5 XP027flatten forward+5 XP028flatten backward+5 XP029linear forward+5 XP030linear grad input+5 XP031linear grad weights+5 XP032linear grad bias+5 XP033linear backward+5 XP
PART 4
0/8Fused Loss and Optimizers
Fuse softmax with cross-entropy for stable training and implement both SGD and Adam parameter updates.
PART 5
0/10Assembling LeNet
Compose the layer primitives into convolutional and classifier blocks, wire up the full LeNet forward/backward pass, and add a predict helper.
PART 6
0/4Synthetic Data Pipeline
Generate a small synthetic image dataset and build shuffling, train/test splitting, and minibatch iteration utilities.
PART 7
0/4Training Loop and Evaluation
Tie everything together with a training step, epoch loop, full training driver, and a held-out evaluation routine.