examples: RuntimeError: invalid argument 2: size '[-1 x 300]' for SNLI example

Had kind of the same issue with the same error but my padding was correct. With my issue I was flattening incorrectly and was able to resolve by verifying the size of x before flattening the tensor and corrected my x.view and the fc1 layer accordingly.

class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        # convolutional layer
        self.conv1 = nn.Conv2d(3, 6, 3, padding=1)
        self.conv2 = nn.Conv2d(6, 16, 3, padding=1)
        self.conv3 = nn.Conv2d(16, 32, 3, padding=1)
        self.pool = nn.MaxPool2d(2, 2)
        ->self.fc1 = nn.Linear(32 * 4 * 4, 120)
        self.fc2 = nn.Linear(120, 10)

    def forward(self, x):
        # add sequence of convolutional and max pooling layers
        x = self.pool(F.relu(self.conv1(x)))
        x = self.pool(F.relu(self.conv2(x)))
        x = self.pool(F.relu(self.conv3(x)))
        ->print(x.size())
        ->x = x.view(-1, 32 * 4 * 4)
        x = F.relu(self.fc1(x))
        x = F.relu(self.fc2(x))
        
        return x

I got the similar error, can you elaborate a little bit more on “You need to calculate the input size of this layer according to the input size of the net.”

I got the following error:

RuntimeError: invalid argument 2: size '[-1 x 1024]' is invalid for input with 11520 elements at 
..\aten\src\TH\THStorage.cpp:80

Full error information:

RuntimeError                              Traceback (most recent call last)
<ipython-input-9-6694bd56733d> in <module>()
     21         optimizer.zero_grad()
     22         # forward pass: compute predicted outputs by passing inputs to the model
---> 23         output = model(data)
     24         # calculate the batch loss
     25         loss = criterion(output, target)

~\Anaconda3\lib\site-packages\torch\nn\modules\module.py in __call__(self, *input, **kwargs)
    475             result = self._slow_forward(*input, **kwargs)
    476         else:
--> 477             result = self.forward(*input, **kwargs)
    478         for hook in self._forward_hooks.values():
    479             hook_result = hook(self, input, result)

<ipython-input-7-971843679cb8> in forward(self, x)
     21         x = self.pool(F.relu(self.conv2(x)))
     22         x = self.pool(F.relu(self.conv3(x)))
---> 23         x = x.view(-1, 64*4*4)
     24         x = self.dropout(x)
     25         x = F.relu(self.fc1(x))

My code block:

import torch.nn as nn
import torch.nn.functional as F

# define the CNN architecture
class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()

        self.conv1 = nn.Conv2d(3, 16, kernel_size=3, padding=1) # convolutional layer
        self.conv2 = nn.Conv2d(16, 32, kernel_size=3)
        self.conv3 = nn.Conv2d(32, 64, kernel_size=3, padding=1)
        self.pool = nn.MaxPool2d(2, 2) # max pooling layer
        #self.drop2d = nn.Dropout2d(0.2)
        self.fc1 = nn.Linear(64*4*4, 500) # densely connected layer
        self.fc2 = nn.Linear(500, 10)
        self.dropout = nn.Dropout(0.2)

    def forward(self, x):
        # add sequence of convolutional and max pooling layers
        x = self.pool(F.relu(self.conv1(x)))
        x = self.pool(F.relu(self.conv2(x)))
        x = self.pool(F.relu(self.conv3(x)))
        x = x.view(-1, 64*4*4)
        x = self.dropout(x)
        x = F.relu(self.fc1(x))
        x = self.dropout(x)
        x = self.fc2(x)
        return x

# create a complete CNN
model = Net()

Found the cause: “parameter padding=1” was missing in one layer.

Wrong code:

    self.conv1 = nn.Conv2d(3, 16, kernel_size=3, padding=1) # convolutional layer
    self.conv2 = nn.Conv2d(16, 32, kernel_size=3)
    self.conv3 = nn.Conv2d(32, 64, kernel_size=3, padding=1)

Right code:

    self.conv1 = nn.Conv2d(3, 16, kernel_size=3, padding=1) # convolutional layer
    self.conv2 = nn.Conv2d(16, 32, kernel_size=3, padding=1)
    self.conv3 = nn.Conv2d(32, 64, kernel_size=3, padding=1)