diff options
| author | Roy Li <royboy@fb.com> | 2018-12-17 15:44:23 -0800 |
|---|---|---|
| committer | Facebook Github Bot <facebook-github-bot@users.noreply.github.com> | 2018-12-17 15:46:37 -0800 |
| commit | e0b261a35b1483e15c7b81a476851abfdb69602a (patch) | |
| tree | 3a6bd2b06252a6c7837ed3d85055727af61b5280 /tools | |
| parent | c66adfc16b8c3e486c2fbcc18de2f0cffd73b208 (diff) | |
| download | pytorch-e0b261a35b1483e15c7b81a476851abfdb69602a.tar.gz pytorch-e0b261a35b1483e15c7b81a476851abfdb69602a.tar.bz2 pytorch-e0b261a35b1483e15c7b81a476851abfdb69602a.zip | |
Port nn fold and unfold to c++
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14597
Reviewed By: ezyang
Differential Revision: D13272227
fbshipit-source-id: 6eccab5ff5830a977398a96393b778095120edc6
Diffstat (limited to 'tools')
| -rw-r--r-- | tools/autograd/derivatives.yaml | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/tools/autograd/derivatives.yaml b/tools/autograd/derivatives.yaml index 3f2e8952eb..42c89da38f 100644 --- a/tools/autograd/derivatives.yaml +++ b/tools/autograd/derivatives.yaml @@ -1106,6 +1106,12 @@ - name: thnn_conv_dilated3d_backward(Tensor grad_output, Tensor self, Tensor weight, IntList kernel_size, IntList stride, IntList padding, IntList dilation, Tensor columns, Tensor ones, std::array<bool,3> output_mask) grad_output, self, weight: _convolution_double_backward(grads[0], grads[1], grads[2], grad_output, weight, self, stride, padding, dilation, false, {{0, 0, 0}}, 1, false, false, false, grad_input_mask) +- name: thnn_col2im(Tensor self, IntList output_size, IntList kernel_size, IntList dilation, IntList padding, IntList stride) + self: thnn_col2im_backward(grad, kernel_size, dilation, padding, stride) + +- name: thnn_im2col(Tensor self, IntList kernel_size, IntList dilation, IntList padding, IntList stride) + self: thnn_im2col_backward(grad, {self.size(2), self.size(3)}, kernel_size, dilation, padding, stride) + # NN double backwards support - name: adaptive_avg_pool2d_backward(Tensor grad_output, Tensor self) |