CNN Input size to fully connected layer

Hi, trying to make a CNN work but i am getting errors regarding the input to the fc. I understand that it is the input that is wrong but i cant seem to get it right any help would be much appreciated.

Here is the network

class Net(nn.Module):
    def __init__(self):
      super(SecondCNN, self).__init__()
      self.conv1 = nn.Conv2d(3, 6, 5)
      self.pool = nn.MaxPool2d(1, 1)
      self.conv2 = nn.Conv2d(6, 16, 5)
      self.conv3 = nn.Conv2d(16, 32, 5)
      self.fc1 = nn.Linear(32 * 20 * 20 , 120)
      self.fc2 = nn.Linear(120, 84)
      self.fc3 = nn.Linear(84, 10)

    def forward(self, x):

      x = self.pool(F.relu(self.conv1(x)))
      print(x.shape)
      x = self.pool(F.relu(self.conv2(x)))
      print(x.shape)
      x = self.pool(F.relu(self.conv3(x)))
      print(x.shape)
      x = F.relu(self.fc1(x))
      print(x.shape)
      x = F.relu(self.fc2(x))
      x = self.fc3(x)
      return x

Here is the size of the first input to the network: torch.Size([4, 3, 32, 32])

And as you can see im printing out the shape along the way as im trying to debug this, i only get the first three as it throws the runtime error before the fc.

Here are the shapes:

torch.Size([4, 6, 28, 28])
torch.Size([4, 16, 24, 24])
torch.Size([4, 32, 20, 20])

Your model architecture looks good and also the debugging steps printing the shapes are great.
The issue is raised because you are not flattening the activation tensor via x = x.view(x.size(0), -1) before passing it to self.fc1. Add this line of code and it should work.