DL03 · Convolutional Networks
A free, illustrated L2 course of 9 modules in the Deep Learning track. Modules include: Convolution and cross-correlation; Padding, stride, pooling and receptive fields; Classic architectures from LeNet to VGG.
DL03 L2 · Intermediate 9 modules · about 36 lessons Planned · Ring 2
Part of the Deep Learning track. Ships in Ring 2 (Spine) of the roadmap.
Modules
- Convolution and cross-correlation
- Padding, stride, pooling and receptive fields
- Classic architectures from LeNet to VGG
- Residual networks and after
- Transfer learning and fine-tuning
- Data augmentation
- Visualizing what a network learns
- Efficient architectures
- Detection and segmentation, a first look
Signature illustrations
The pictures a learner should still see in their head a year later:
Sliding-kernel animations, receptive fields growing with depth, galleries of feature maps, architecture block diagrams, saliency maps.
Status
This course is planned, not written yet. It ships with Ring 2 (Spine).
Every lesson will follow the same twelve-part blueprint with at least three purpose-made figures. Watch the repository to hear when modules ship, or help write and review it.