Skip to content

Commit 6a02f85

Browse files
mthrokfacebook-github-bot
mthrok
authored and
facebook-github-bot
committed
Update to clang-format 11. (#3254)
Reviewed By: datumbox Differential Revision: D25954565 fbshipit-source-id: 5d65c47b86c56112197697fc652e14c0c15093dc
1 parent c49e346 commit 6a02f85

File tree

6 files changed

+233
-195
lines changed

6 files changed

+233
-195
lines changed

‎torchvision/csrc/ops/autograd/ps_roi_align_kernel.cpp

Lines changed: 7 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -62,12 +62,13 @@ class PSROIAlignFunction
6262
input_shape[2],
6363
input_shape[3]);
6464

65-
return {grad_in,
66-
torch::autograd::Variable(),
67-
torch::autograd::Variable(),
68-
torch::autograd::Variable(),
69-
torch::autograd::Variable(),
70-
torch::autograd::Variable()};
65+
return {
66+
grad_in,
67+
torch::autograd::Variable(),
68+
torch::autograd::Variable(),
69+
torch::autograd::Variable(),
70+
torch::autograd::Variable(),
71+
torch::autograd::Variable()};
7172
}
7273
};
7374

‎torchvision/csrc/ops/autograd/ps_roi_pool_kernel.cpp

Lines changed: 6 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -53,11 +53,12 @@ class PSROIPoolFunction : public torch::autograd::Function<PSROIPoolFunction> {
5353
input_shape[2],
5454
input_shape[3]);
5555

56-
return {grad_in,
57-
torch::autograd::Variable(),
58-
torch::autograd::Variable(),
59-
torch::autograd::Variable(),
60-
torch::autograd::Variable()};
56+
return {
57+
grad_in,
58+
torch::autograd::Variable(),
59+
torch::autograd::Variable(),
60+
torch::autograd::Variable(),
61+
torch::autograd::Variable()};
6162
}
6263
};
6364

‎torchvision/csrc/ops/autograd/roi_align_kernel.cpp

Lines changed: 8 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -57,13 +57,14 @@ class ROIAlignFunction : public torch::autograd::Function<ROIAlignFunction> {
5757
input_shape[3],
5858
ctx->saved_data["sampling_ratio"].toInt(),
5959
ctx->saved_data["aligned"].toBool());
60-
return {grad_in,
61-
torch::autograd::Variable(),
62-
torch::autograd::Variable(),
63-
torch::autograd::Variable(),
64-
torch::autograd::Variable(),
65-
torch::autograd::Variable(),
66-
torch::autograd::Variable()};
60+
return {
61+
grad_in,
62+
torch::autograd::Variable(),
63+
torch::autograd::Variable(),
64+
torch::autograd::Variable(),
65+
torch::autograd::Variable(),
66+
torch::autograd::Variable(),
67+
torch::autograd::Variable()};
6768
}
6869
};
6970

‎torchvision/csrc/ops/autograd/roi_pool_kernel.cpp

Lines changed: 6 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -53,11 +53,12 @@ class ROIPoolFunction : public torch::autograd::Function<ROIPoolFunction> {
5353
input_shape[2],
5454
input_shape[3]);
5555

56-
return {grad_in,
57-
torch::autograd::Variable(),
58-
torch::autograd::Variable(),
59-
torch::autograd::Variable(),
60-
torch::autograd::Variable()};
56+
return {
57+
grad_in,
58+
torch::autograd::Variable(),
59+
torch::autograd::Variable(),
60+
torch::autograd::Variable(),
61+
torch::autograd::Variable()};
6162
}
6263
};
6364

‎torchvision/csrc/ops/cpu/deform_conv2d_kernel.cpp

Lines changed: 103 additions & 86 deletions
Original file line numberDiff line numberDiff line change
@@ -634,44 +634,50 @@ std::tuple<at::Tensor, at::Tensor, at::Tensor> backward_gradient_inputs(
634634
input = input.reshape(
635635
{batch_sz / n_parallel_imgs, n_parallel_imgs, n_in_channels, in_h, in_w});
636636

637-
grad_offset = grad_offset.reshape({batch_sz / n_parallel_imgs,
638-
n_parallel_imgs,
639-
n_offset_grps * 2 * weight_h * weight_w,
640-
out_h,
641-
out_w});
642-
offset = offset.reshape({batch_sz / n_parallel_imgs,
643-
n_parallel_imgs,
644-
n_offset_grps * 2 * weight_h * weight_w,
645-
out_h,
646-
out_w});
637+
grad_offset = grad_offset.reshape(
638+
{batch_sz / n_parallel_imgs,
639+
n_parallel_imgs,
640+
n_offset_grps * 2 * weight_h * weight_w,
641+
out_h,
642+
out_w});
643+
offset = offset.reshape(
644+
{batch_sz / n_parallel_imgs,
645+
n_parallel_imgs,
646+
n_offset_grps * 2 * weight_h * weight_w,
647+
out_h,
648+
out_w});
647649

648650
if (use_mask) {
649-
grad_mask = grad_mask.reshape({batch_sz / n_parallel_imgs,
650-
n_parallel_imgs,
651-
n_offset_grps * weight_h * weight_w,
652-
out_h,
653-
out_w});
654-
mask = mask.reshape({batch_sz / n_parallel_imgs,
655-
n_parallel_imgs,
656-
n_offset_grps * weight_h * weight_w,
657-
out_h,
658-
out_w});
651+
grad_mask = grad_mask.reshape(
652+
{batch_sz / n_parallel_imgs,
653+
n_parallel_imgs,
654+
n_offset_grps * weight_h * weight_w,
655+
out_h,
656+
out_w});
657+
mask = mask.reshape(
658+
{batch_sz / n_parallel_imgs,
659+
n_parallel_imgs,
660+
n_offset_grps * weight_h * weight_w,
661+
out_h,
662+
out_w});
659663
}
660664

661665
grad_out = grad_out
662-
.reshape({batch_sz / n_parallel_imgs,
663-
n_parallel_imgs,
664-
n_weight_grps,
665-
n_out_channels / n_weight_grps,
666-
out_h,
667-
out_w})
666+
.reshape(
667+
{batch_sz / n_parallel_imgs,
668+
n_parallel_imgs,
669+
n_weight_grps,
670+
n_out_channels / n_weight_grps,
671+
out_h,
672+
out_w})
668673
.permute({0, 2, 3, 1, 4, 5});
669674

670-
weight = weight.reshape({n_weight_grps,
671-
weight.size(0) / n_weight_grps,
672-
weight.size(1),
673-
weight.size(2),
674-
weight.size(3)});
675+
weight = weight.reshape(
676+
{n_weight_grps,
677+
weight.size(0) / n_weight_grps,
678+
weight.size(1),
679+
weight.size(2),
680+
weight.size(3)});
675681

676682
columns = columns.view(
677683
{n_weight_grps, columns.size(0) / n_weight_grps, columns.size(1)});
@@ -775,37 +781,41 @@ at::Tensor backward_gradient_parameters(
775781
}
776782

777783
at::Tensor grad_out_buf = grad_out
778-
.reshape({batch_sz / n_parallel_imgs,
779-
n_parallel_imgs,
780-
n_weight_grps,
781-
n_out_channels / n_weight_grps,
782-
out_h,
783-
out_w})
784+
.reshape(
785+
{batch_sz / n_parallel_imgs,
786+
n_parallel_imgs,
787+
n_weight_grps,
788+
n_out_channels / n_weight_grps,
789+
out_h,
790+
out_w})
784791
.permute({0, 2, 3, 1, 4, 5})
785792
.contiguous();
786793

787794
input = input.reshape(
788795
{batch_sz / n_parallel_imgs, n_parallel_imgs, n_in_channels, in_h, in_w});
789796

790-
offset = offset.reshape({batch_sz / n_parallel_imgs,
791-
n_parallel_imgs,
792-
n_offset_grps * 2 * weight_h * weight_w,
793-
out_h,
794-
out_w});
797+
offset = offset.reshape(
798+
{batch_sz / n_parallel_imgs,
799+
n_parallel_imgs,
800+
n_offset_grps * 2 * weight_h * weight_w,
801+
out_h,
802+
out_w});
795803

796804
if (use_mask) {
797-
mask = mask.reshape({batch_sz / n_parallel_imgs,
798-
n_parallel_imgs,
799-
n_offset_grps * weight_h * weight_w,
800-
out_h,
801-
out_w});
805+
mask = mask.reshape(
806+
{batch_sz / n_parallel_imgs,
807+
n_parallel_imgs,
808+
n_offset_grps * weight_h * weight_w,
809+
out_h,
810+
out_w});
802811
}
803812

804-
grad_weight = grad_weight.view({n_weight_grps,
805-
grad_weight.size(0) / n_weight_grps,
806-
grad_weight.size(1),
807-
grad_weight.size(2),
808-
grad_weight.size(3)});
813+
grad_weight = grad_weight.view(
814+
{n_weight_grps,
815+
grad_weight.size(0) / n_weight_grps,
816+
grad_weight.size(1),
817+
grad_weight.size(2),
818+
grad_weight.size(3)});
809819

810820
auto columns = at::empty(
811821
{n_weight_grps,
@@ -846,10 +856,11 @@ at::Tensor backward_gradient_parameters(
846856
}
847857
}
848858

849-
grad_weight = grad_weight.view({grad_weight.size(0) * grad_weight.size(1),
850-
grad_weight.size(2),
851-
grad_weight.size(3),
852-
grad_weight.size(4)});
859+
grad_weight = grad_weight.view(
860+
{grad_weight.size(0) * grad_weight.size(1),
861+
grad_weight.size(2),
862+
grad_weight.size(3),
863+
grad_weight.size(4)});
853864
return grad_weight;
854865
}
855866

@@ -976,26 +987,29 @@ at::Tensor deform_conv2d_forward_kernel(
976987
}
977988

978989
// Separate batches into blocks
979-
out = out.view({batch_sz / n_parallel_imgs,
980-
n_parallel_imgs,
981-
out_channels,
982-
out_h,
983-
out_w});
990+
out = out.view(
991+
{batch_sz / n_parallel_imgs,
992+
n_parallel_imgs,
993+
out_channels,
994+
out_h,
995+
out_w});
984996
input_c = input_c.view(
985997
{batch_sz / n_parallel_imgs, n_parallel_imgs, n_in_channels, in_h, in_w});
986998

987-
offset_c = offset_c.view({batch_sz / n_parallel_imgs,
988-
n_parallel_imgs,
989-
n_offset_grps * 2 * weight_h * weight_w,
990-
out_h,
991-
out_w});
999+
offset_c = offset_c.view(
1000+
{batch_sz / n_parallel_imgs,
1001+
n_parallel_imgs,
1002+
n_offset_grps * 2 * weight_h * weight_w,
1003+
out_h,
1004+
out_w});
9921005

9931006
if (use_mask) {
994-
mask_c = mask_c.view({batch_sz / n_parallel_imgs,
995-
n_parallel_imgs,
996-
n_offset_grps * weight_h * weight_w,
997-
out_h,
998-
out_w});
1007+
mask_c = mask_c.view(
1008+
{batch_sz / n_parallel_imgs,
1009+
n_parallel_imgs,
1010+
n_offset_grps * weight_h * weight_w,
1011+
out_h,
1012+
out_w});
9991013
}
10001014

10011015
at::Tensor out_buf = at::zeros(
@@ -1006,16 +1020,18 @@ at::Tensor deform_conv2d_forward_kernel(
10061020
out.options());
10071021

10081022
// Separate channels into convolution groups
1009-
out_buf = out_buf.view({out_buf.size(0),
1010-
n_weight_grps,
1011-
out_buf.size(1) / n_weight_grps,
1012-
out_buf.size(2),
1013-
out_buf.size(3)});
1014-
weight_c = weight_c.view({n_weight_grps,
1015-
weight_c.size(0) / n_weight_grps,
1016-
weight_c.size(1),
1017-
weight_c.size(2),
1018-
weight_c.size(3)});
1023+
out_buf = out_buf.view(
1024+
{out_buf.size(0),
1025+
n_weight_grps,
1026+
out_buf.size(1) / n_weight_grps,
1027+
out_buf.size(2),
1028+
out_buf.size(3)});
1029+
weight_c = weight_c.view(
1030+
{n_weight_grps,
1031+
weight_c.size(0) / n_weight_grps,
1032+
weight_c.size(1),
1033+
weight_c.size(2),
1034+
weight_c.size(3)});
10191035

10201036
// Sample points and perform convolution
10211037
auto columns = at::zeros(
@@ -1056,11 +1072,12 @@ at::Tensor deform_conv2d_forward_kernel(
10561072
columns.view({columns.size(0) * columns.size(1), columns.size(2)});
10571073
}
10581074

1059-
out_buf = out_buf.view({batch_sz / n_parallel_imgs,
1060-
out_channels,
1061-
n_parallel_imgs,
1062-
out_h,
1063-
out_w});
1075+
out_buf = out_buf.view(
1076+
{batch_sz / n_parallel_imgs,
1077+
out_channels,
1078+
n_parallel_imgs,
1079+
out_h,
1080+
out_w});
10641081
out_buf.transpose_(1, 2);
10651082
out.copy_(out_buf);
10661083
out = out.view({batch_sz, out_channels, out_h, out_w});

0 commit comments

Comments
 (0)