Offset output channel

deform_conv3d_pytorch_op/deform2d_double/test_model.py

Line 22 in 03985c6

c_in // channel_per_group * 2 * kernel * kernel,

Your implementation is not the same as the original paper described in section 2.1. Offset output channel dimension should be 2 x N, where N is the conv kernel size k*k in 2D case. That is, the offset is shared across the channel dimension. You could refer to the implementation of https://github.com/ChunhuanLin/deform_conv_pytorch

If you set the channel_per_group the same as c_in, it will be your case.
You could refer to the official implementation: https://github.com/msracver/Deformable-ConvNets/blob/a2977507d0980bafab7e4161ea78138fe510cb65/rfcn/operator_cxx/nn/deformable_im2col.cuh#L232