13RanNeural nets & vision

AlexNet in Keras

I rebuilt an AlexNet-style network layer by layer in Keras for 227 by 227 colour images, compiled it, printed the summary and drew it with visualkeras. It was never trained.

Original tutorial by Aman Kharwal: AlexNet Neural Network Architecture (opens in a new tab)

What it printed in 2022

Numbers read from the saved notebook outputs. Cell numbers count every cell in the notebook, Markdown included, starting at 1.
Output	Value	Cell #
AlexNet parameters	41,546,506	8

Cell # counts every cell from the top of the notebook, Markdown included, starting at 1.

41.5 million parameters, and over 90% of them sit in the first dense layer, where 9,216 flattened features meet 4,096 units.

The data

No data. The notebook is about the architecture.

What I'd do differently

Train it on something small, for example CIFAR-10 resized, so the architecture meets data.
Work through the output-size formula for each layer by hand, which was the useful part of the exercise.