Skip to content
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.

[Numpy] fix test_operator_gpu.test_upsampling_bilinear_with_type #14557

Merged
merged 2 commits into from
Apr 4, 2019
Merged

[Numpy] fix test_operator_gpu.test_upsampling_bilinear_with_type #14557

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
a433742
Fix test_operator_gpu.test_upsampling_bilinear_with_type
junrushao Mar 29, 2019
1772880
Address comments
junrushao Mar 30, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions src/operator/nn/deconvolution-inl.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -143,9 +143,9 @@ struct DeconvolutionParam : public dmlc::Parameter<DeconvolutionParam> {
}
}
} else {
for (size_t i = 0; i < ndim; i++) {
o_pad[i] = pad[i];
o_adj[i] = adj[i];
for (int i = 0; i < (int) ndim; i++) {
o_pad[i] = i < pad.ndim() ? pad[i] : 0;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

According the initialization function, pad.ndim() should be equal to ndim here. Do you understand why i can be bigger than pad.ndim()?

Copy link
Member Author

@junrushao junrushao Mar 30, 2019 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The upsampling operator customized DeconvolutionParam here, which makes them not equal. It is kinda weird but I believe adj is not used in this case.

o_adj[i] = i < adj.ndim() ? adj[i] : 0;
}
}
}
Expand Down