deformable convolution

[tqyunwuxin]

hello,
do you Implement the deformable convolution by yourself？What's the difference between it and that in EDVR? I noticed that number of channels of the featuremap and offset can be different，but it should be same in EDVR.
if i want to use it in my own project, Should I copy the directories ”src“ ”modules" "functions" "_ext" ,and then run the make.sh?
thank you!

[YapengTian]

Both TDAN and EDVR used deformable convolution implemented by others. EDVR used deconv-v2, which can be compiled by pytorch-1.0+.

If you use it in your project, please copy files and run make.sh. Please check this repo: https://github.com/chengdazhi/Deformable-Convolution-V2-PyTorch/tree/pytorch_1.0.0.

[Mambatttrj]

ImportError: libpython3.6m.so.1.0: cannot open shared object file: No such file or directory

hi~ How can I do to deal with this problem

[YapengTian]

Could you install pytorch0.3.1 and run make.sh to compile the dconv?

[Mambatttrj]

Could you install pytorch0.3.1 and run make.sh to compile the dconv?

I install pytorch0.3.1. And I run make.sh, too
How could I know compile successfully.

[YapengTian]

If there are no errors, it should be good.

[CDisbettermen]

hi i want to know what size of train datasize,i make some train data ,resolution 180*144,channels 3,but they can not train this model.

[YapengTian]

hi i want to know what size of train datasize,i make some train data ,resolution 180*144,channels 3,but they can not train this model.

Sorry for the late response. I used 48x48 resolution as the input size.