Dear friends, what is the core difference between the code of senet154 you published and the code of the original link you used to refer to? Is the pytorch version and usage of the original linked code the same as yours? Is there anything wrong with the original linked code

[xianglei3]

You mentioned that the original link is
https://github.com/Cadene/pretrained-models.pytorch/blob/master/pretrainedmodels/models/senet.py
I would like to know where your core improvement points are. I would like to thank you very much and look forward to your reply

In addition, when I used the original linked code, an error occurred as follows:
File "/home/mmdetection/mmdet/utils/registry.py", line 76, in build_from_cfg
return obj_cls(**args)
File "/home/mmdetection/mmdet/models/backbones/senet.py", line 292, in init
downsample_padding=0
File "/home/mmdetection/mmdet/models/backbones/senet.py", line 331, in _make_layer
if stride != 1 or self.inplanes != planes * block.expansion:
AttributeError: 'str' object has no attribute 'expansion'
Have you ever had this kind of problem, has it anything to do with the version or other factors?

Finally, how to use se ﹣ resnet50? Can I use your code directly or should I refer to the original link you used to refer to when you wrote senet154