PyTorch module (nn.Module) questions

1

Hi, I printed the alexnet model…

model = models.alexnet(pretrained=True)

What is the idea behind features, classifier, and avgpool inside:

model._models

Just need little explain.

2

The Neural Networks has just been devided in 3 parts (you could also just see it as one big NN)

the features part is just applying the following layers:

nn.Sequential(
            nn.Conv2d(3, 64, kernel_size=11, stride=4, padding=2),
            nn.ReLU(inplace=True),
            nn.MaxPool2d(kernel_size=3, stride=2),
            nn.Conv2d(64, 192, kernel_size=5, padding=2),
            nn.ReLU(inplace=True),
            nn.MaxPool2d(kernel_size=3, stride=2),
            nn.Conv2d(192, 384, kernel_size=3, padding=1),
            nn.ReLU(inplace=True),
            nn.Conv2d(384, 256, kernel_size=3, padding=1),
            nn.ReLU(inplace=True),
            nn.Conv2d(256, 256, kernel_size=3, padding=1),
            nn.ReLU(inplace=True),
            nn.MaxPool2d(kernel_size=3, stride=2),
        )

Then perform an averagepooling,

and finally applying the following group of layers (here gathered under the name classifier):

nn.Sequential(
            nn.Dropout(),
            nn.Linear(256 * 6 * 6, 4096),
            nn.ReLU(inplace=True),
            nn.Dropout(),
            nn.Linear(4096, 4096),
            nn.ReLU(inplace=True),
            nn.Linear(4096, num_classes),
        )
1 Like
3

Thanks should look into the source code myself. :sunglasses: