Neural Network initialized with random weights always returns the same output with random inputs

Question

I have a problem with pytorch in Spyder. A randomly initialized Neural Network returns always the same output also for random input tensor. I am currently using local GPU with Spyder. I made sure that the initialization of the weights is random and not all zeros.

Example:

        x = torch.rand(1, 3, 360, 640)
        x = self.stage_1(x)
        x = self.stage_2(x)
        x = self.stage_3(x)
        x = self.stage_4(x)
        x = self.stage_5(x)
        x = self.stage_6(x)
        x = torch.flatten(x, start_dim=1)

        y = torch.rand(1, 3, 360, 640)
        y = self.stage_1(y)
        y = self.stage_2(y)
        y = self.stage_3(y)
        y = self.stage_4(y)
        y = self.stage_5(y)
        y = self.stage_6(y)
        y = torch.flatten(y, start_dim=1)

This code returns always y == x

This is the stage class:

class VggStage(nn.Module):
   def __init__(self,
                input_channels: int,
                output_channels: int) -> None:
    """
    

    Parameters
    ----------
    input_channels : int
        DESCRIPTION.
    output_channels : int
        DESCRIPTION.

    Returns
    -------
    None
        DESCRIPTION.

    """
       super().__init__()
    
       self.conv1 = nn.Conv2d(in_channels=input_channels,
                              out_channels=output_channels, 
                              kernel_size=(3, 3))
       self.conv2 = nn.Conv2d(in_channels=output_channels,
                              out_channels=output_channels, 
                              kernel_size=(3, 3))
       self.max_pool = nn.MaxPool2d(kernel_size=(2, 2),
                                    stride=(2, 2))
    
   def forward(self,
               x: torch.Tensor) -> torch.Tensor:

       x = F.relu(self.conv1(x))
       x = F.relu(self.conv2(x))
       x = self.max_pool(x)
    
       return x