GANS W2A C3 UNQ_C3 ExpandingBlock

Course Q&A Generative Adversarial Networks (GANS) Apply Generative Adversarial Networks
1

I’m working through this exercise and am pretty sure I did it right, but I’m getting an error that doesn’t make sense:

RuntimeError: invalid argument 0: Sizes of tensors must match except in dimension 1. Got 99 and 100 in dimension 2 at /pytorch/aten/src/TH/generic/THTensor.cpp:612

Does anyone monitor these forums and can help me out? Not sure I should post my code here, so feel free to delete this, but here’s what I have
#### START CODE HERE ####
self.conv1 = nn.Conv2d(input_channels, input_channels // 2, kernel_size=2) # 2x2
self.conv2 = nn.Conv2d(input_channels, input_channels // 2, kernel_size=3) # 3x3
self.conv3 = nn.Conv2d(input_channels // 2, input_channels // 2, kernel_size=3) # 3x3
#### END CODE HERE ####

1 Like
2

I’ve tried a couple other combinations, like putting input_channels // 2, as the first argument of the second self-conv2

1 Like
3

Nevermind, the error was in my crop() function

1 Like