miraclewkf / senet-pytorch Goto Github PK

I just recode and change:
self.conv1 = conv3x3(inplanes, planes * 2, stride)
self.bn1 = nn.BatchNorm2d(planes * 2)
self.conv2 = conv3x3(planes * 2, planes * 2)
self.bn2 = nn.BatchNorm2d(planes * 2)

but failed:
File "/home/zufall/Anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 325, in call
result = self.forward(*input, **kwargs)
File "/media/zufall/workspace1/Recognize_Commodity_all/train/se_resnet.py", line 286, in forward
x = self.layer1(x)
File "/home/zufall/Anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 325, in call
result = self.forward(*input, **kwargs)
File "/home/zufall/Anaconda3/lib/python3.6/site-packages/torch/nn/modules/container.py", line 67, in forward
input = module(input)
File "/home/zufall/Anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 325, in call
result = self.forward(*input, **kwargs)
File "/media/zufall/workspace1/Recognize_Commodity_all/train/se_resnet.py", line 56, in forward
out = self.fc1(out)
File "/home/zufall/Anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 325, in call
result = self.forward(*input, **kwargs)
File "/home/zufall/Anaconda3/lib/python3.6/site-packages/torch/nn/modules/linear.py", line 55, in forward
return F.linear(input, self.weight, self.bias)
File "/home/zufall/Anaconda3/lib/python3.6/site-packages/torch/nn/functional.py", line 835, in linear
return torch.addmm(bias, input, weight.t())
RuntimeError: size mismatch at /pytorch/torch/lib/THC/generic/THCTensorMathBlas.cu:243

Don't have test code?

What should I do to deal with 448*448 pictures with se_resnet50?

hello,
i wanna try my custom datasets which has 21classes (20+1) ,but i finlly find that output Tensor is a logical value , why no SOFTMAX ?

In the code, the global average pooling in the SE structure is implemented through average pooling, but the output size is not like 1X1XC in the paper. Is there a problem with the code? I try to use A to implement it. Is it feasible?

in your code:
if planes == 64:
self.globalAvgPool = nn.AvgPool2d(56, stride=1)
elif planes == 128:
self.globalAvgPool = nn.AvgPool2d(28, stride=1)
elif planes == 256:
self.globalAvgPool = nn.AvgPool2d(14, stride=1)
elif planes == 512:
self.globalAvgPool = nn.AvgPool2d(7, stride=1)

my code:
self.avgpool = nn.AdaptiveMaxPool2d((1, 1))

Hi，
Before seethe original paper, I have saw the channel-wise in 'SCA-CNN', I think the two structures are vary similar, could you tell me the difference between them?
Thank you very much!

Does this version support pytorch0.4.0?

hello ，Do you have se-renext151 pre training parameters?