Deep Learning with PyTorch

Building a CNN model from scratch

For this example, let's build our own architecture from scratch. Our network architecture will contain a combination of different layers, namely:

• Conv2d
• MaxPool2d
• Rectified linear unit (ReLU)
• View
• Linear layer

Let's look at a pictorial representation of the architecture we are going to implement:

Let's implement this architecture in PyTorch and then walk through what each individual layer does:

class Net(nn.Module):    def __init__(self):        super().__init__()        self.conv1 = nn.Conv2d(1, 10, kernel_size=5)        self.conv2 = nn.Conv2d(10, 20, kernel_size=5)        self.conv2_drop = nn.Dropout2d() self.fc1 = nn.Linear(320, 50) ...