All Courses
6 lessons

CNNs & Vision

Filters, feature maps, and the architectures that taught machines to see

Lessons

  1. 01

    Convolution Operation

    The kernel that slides, multiplies, sums.

    Easy
  2. 02

    Pooling

    Max, average, and why we downsample.

    Easy
  3. 03

    Build a CNN from Scratch

    A LeNet-style conv net, one layer at a time in NumPy.

    Medium
  4. 04

    Image Classifier

    CIFAR-10 end-to-end — augmentation, training, evaluation.

    Medium
  5. 05

    ResNet & Skip Connections

    Residuals, identity paths, and why depth stopped hurting.

    Hard
  6. 06

    Vision Transformer (ViT)

    Treating image patches as tokens — the bridge to attention.

    Hard