code does'nt match its image in neural_networks_tutorial

[nipeone]

conv1 kernel size in the code is 3x3,but feature map size after conv1 is 28x28 in the above image,it should be 30x30,or change the kernel size to 5x5.

[aimuch]

I also find self.conv1 = nn.Conv2d(1, 6, 3) should self.conv1 = nn.Conv2d(1, 6, 5) and self.conv2 has the same problem.

[aimuch]

I also find self.conv1 = nn.Conv2d(1, 6, 3) should self.conv1 = nn.Conv2d(1, 6, 5) and self.conv2 has the same problem.

Official has fix this problem:

x = F.max_pool2d(F.relu(self.conv1(x)), (2, 2))

[holly1238]

Kernel size is 5x5. This was fixed in the tutorial in #1257.