dafdc9b5cb8fa4c65285aad22b1429549d06d71a,networks/resample2d_package/functions/resample2d.py,Resample2dFunction,backward,#Resample2dFunction#Any#,27

Before Change



        assert(gradOutput.is_contiguous() == True)
        
        with torch.cuda.device_of(input1):
            b, c, h, w = input1.size()
            gradInput1 = input1.new().resize_(b,c,h,w).zero_()

            b, c, h, w = input2.size()
            gradInput2 = input2.new().resize_(b,c,h,w).zero_()

            resample2d.Resample2d_cuda_backward(input1, input2, gradOutput, gradInput1, gradInput2, self.kernel_size)

        return gradInput1, gradInput2

After Change


                                            grad_input1.data, grad_input2.data,
                                            ctx.kernel_size)

        return grad_input1, grad_input2, None
Italian Trulli
In pattern: SUPERPATTERN

Frequency: 3

Non-data size: 9

Instances


Project Name: NVIDIA/flownet2-pytorch
Commit Name: dafdc9b5cb8fa4c65285aad22b1429549d06d71a
Time: 2018-02-04
Author: chenkaidev@gmail.com
File Name: networks/resample2d_package/functions/resample2d.py
Class Name: Resample2dFunction
Method Name: backward


Project Name: NVIDIA/flownet2-pytorch
Commit Name: dafdc9b5cb8fa4c65285aad22b1429549d06d71a
Time: 2018-02-04
Author: chenkaidev@gmail.com
File Name: networks/channelnorm_package/functions/channelnorm.py
Class Name: ChannelNormFunction
Method Name: backward


Project Name: zhanghang1989/PyTorch-Encoding
Commit Name: 2d21747abdbf1968a9d565a4090d5d6297ade71a
Time: 2018-06-04
Author: zhang.hang@rutgers.edu
File Name: encoding/functions/syncbn.py
Class Name: _sum_square
Method Name: forward