Something has gone wrong...

[tuji-sjp]

When I tried to run this code, the following error occurred:

Fast Gradient Sign Method
Model: inception_v3

Traceback (most recent call last):
File "fgsm.py", line 62, in
out = model(inp)
File "/home/jinping/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in call
result = self.forward(*input, **kwargs)
File "/home/jinping/anaconda3/lib/python3.6/site-packages/torchvision/models/inception.py", line 117, in forward
x = F.avg_pool2d(x, kernel_size=8)
RuntimeError: Given input size: (2048x5x5). Calculated output size: (2048x0x0). Output size is too small at /home/jinping/project/pytorch/aten/src/THCUNN/generic/SpatialAveragePooling.cu:63

[sarathknv]

Image size should be 299x299 for Inception. After you make this change in the code, you'll encounter a different error.

Traceback (most recent call last): File "fgsm.py", line 77, in <module> loss.backward() File "/usr/local/lib/python3.5/dist-packages/torch/tensor.py", line 96, in backward torch.autograd.backward(self, gradient, retain_graph, create_graph) File "/usr/local/lib/python3.5/dist-packages/torch/autograd/__init__.py", line 90, in backward allow_unreachable=True) # allow_unreachable flag RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation

There's some problem with Inception model.

[tuji-sjp]

Image size should be 299x299 for Inception. After you make this change in the code, you'll encounter a different error.

Traceback (most recent call last): File "fgsm.py", line 77, in <module> loss.backward() File "/usr/local/lib/python3.5/dist-packages/torch/tensor.py", line 96, in backward torch.autograd.backward(self, gradient, retain_graph, create_graph) File "/usr/local/lib/python3.5/dist-packages/torch/autograd/__init__.py", line 90, in backward allow_unreachable=True) # allow_unreachable flag RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation

There's some problem with Inception model.

ok, I will check it.