Hi Matthew,
Alexnet is fully accessible programmatically via the ‘alexnet’ function. When you call the function, MATLAB loads a pretrained DAGNetwork object that includes all layer definitions, weights, and connections. While the internal training process and some implementation-level details are abstracted, you can easily inspect and modify the architecture.
To view the source-like definition of the network, you can use:
net = alexnet;
layers = net.Layers;
lgraph = layerGraph(net);
For more details regarding ‘alexnet’, you can refer to the following documentations:
- alexnet: https://www.mathworks.com/help/deeplearning/ref/alexnet.html
- layer graph: https://www.mathworks.com/help/deeplearning/ref/nnet.cnn.layergraph.html
Hope this helps!
