CNN Architecture Deep Dive — Stride, Padding & VGGNet | Visual ML

About this lesson

Go beyond basic convolution: learn stride, padding, parameter sharing, and how VGGNet stacks 16 layers. Based on Stanford CS231n. 🎯 What you'll see: • Output size formula: (W-F+2P)/S + 1 • Stride 1 vs stride 2 side-by-side demo • Zero-padding preserving spatial dimensions • Parameter sharing: 105M → 35K weights • Hero: VGGNet volume shrinking 224→112→56→28→14→7 • Receptive field growth with stacked 3x3 filters Perfect companion to our CNN Convolution video. 🔗 More at: https://8gwifi.org/math #CNN #VGGNet #convolution #stride #padding #deeplearning #machinelearning #CS231n #education

Original Description