Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
机器未来
Paddle
提交
09adb769
P
Paddle
项目概览
机器未来
/
Paddle
与 Fork 源项目一致
Fork自
PaddlePaddle / Paddle
通知
1
Star
1
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
1
列表
看板
标记
里程碑
合并请求
0
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
P
Paddle
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
1
Issue
1
列表
看板
标记
里程碑
合并请求
0
合并请求
0
Pages
分析
分析
仓库分析
DevOps
Wiki
0
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
提交
Issue看板
提交
09adb769
编写于
1月 17, 2018
作者:
W
wanghaoshuang
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Fix code style
上级
bfe7e242
变更
3
隐藏空白更改
内联
并排
Showing
3 changed file
with
25 addition
and
22 deletion
+25
-22
paddle/operators/block_expand_op.cc
paddle/operators/block_expand_op.cc
+10
-11
paddle/operators/block_expand_op.cu
paddle/operators/block_expand_op.cu
+5
-4
paddle/operators/block_expand_op.h
paddle/operators/block_expand_op.h
+10
-7
未找到文件。
paddle/operators/block_expand_op.cc
浏览文件 @
09adb769
...
...
@@ -57,16 +57,14 @@ class BlockExpandOp : public framework::OperatorWithKernel {
class
BlockExpandOpMaker
:
public
framework
::
OpProtoAndCheckerMaker
{
public:
BlockExpandOpMaker
(
framework
::
OpProto
*
proto
,
framework
::
OpAttrChecker
*
op_checker
)
BlockExpandOpMaker
(
OpProto
*
proto
,
OpAttrChecker
*
op_checker
)
:
OpProtoAndCheckerMaker
(
proto
,
op_checker
)
{
AddInput
(
"X"
,
R"DOC(
(Tensor)The input tensor has NCHW format.
N: batch size
C: channels
H: height
W: width
)DOC"
);
AddInput
(
"X"
,
"(Tensor)The input tensor has NCHW format."
"N: batch size"
"C: channels"
"H: height"
"W: width"
);
AddOutput
(
"Out"
,
"(LodTensor)The output data of block_expand op,"
);
AddAttr
<
int
>
(
"block_height"
,
"(int)height of block."
);
AddAttr
<
int
>
(
"block_width"
,
"(int)width of block."
);
...
...
@@ -155,7 +153,8 @@ namespace ops = paddle::operators;
REGISTER_OP
(
block_expand
,
ops
::
BlockExpandOp
,
ops
::
BlockExpandOpMaker
,
block_expand_grad
,
ops
::
BlockExpandGradOp
);
REGISTER_OP_CPU_KERNEL
(
block_expand
,
ops
::
BlockExpandKernel
<
paddle
::
platform
::
CPUPlace
,
float
>
);
block_expand
,
ops
::
BlockExpandKernel
<
paddle
::
platform
::
CPUDeviceContext
,
float
>
);
REGISTER_OP_CPU_KERNEL
(
block_expand_grad
,
ops
::
BlockExpandGradKernel
<
paddle
::
platform
::
CPU
Place
,
float
>
);
ops
::
BlockExpandGradKernel
<
paddle
::
platform
::
CPU
DeviceContext
,
float
>
);
paddle/operators/block_expand_op.cu
浏览文件 @
09adb769
...
...
@@ -17,8 +17,9 @@
namespace
ops
=
paddle
::
operators
;
REGISTER_OP_GPU_KERNEL
(
block_expand
,
ops
::
BlockExpandKernel
<
paddle
::
platform
::
GPUPlace
,
float
>
);
REGISTER_OP_GPU_KERNEL
(
REGISTER_OP_CUDA_KERNEL
(
block_expand
,
ops
::
BlockExpandKernel
<
paddle
::
platform
::
CUDADeviceContext
,
float
>
);
REGISTER_OP_CUDA_KERNEL
(
block_expand_grad
,
ops
::
BlockExpandGradKernel
<
paddle
::
platform
::
GPUPlace
,
float
>
);
ops
::
BlockExpandGradKernel
<
paddle
::
platform
::
CUDADeviceContext
,
float
>
);
paddle/operators/block_expand_op.h
浏览文件 @
09adb769
...
...
@@ -31,7 +31,7 @@ inline int get_output_size(int img_size, int block_size, int stride,
return
(
1
+
(
img_size
+
2
*
padding
-
block_size
+
stride
-
1
)
/
stride
);
}
template
<
typename
Place
,
typename
T
>
template
<
typename
DeviceContext
,
typename
T
>
class
BlockExpandKernel
:
public
framework
::
OpKernel
<
T
>
{
public:
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
...
...
@@ -71,8 +71,9 @@ class BlockExpandKernel : public framework::OpKernel<T> {
img_channels
,
block_height
,
block_width
});
math
::
Im2ColFunctor
<
math
::
ColFormat
::
kOCF
,
Place
,
T
>
f
;
f
(
ctx
.
device_context
(),
src
,
dilations
,
strides
,
paddings
,
&
dst
);
math
::
Im2ColFunctor
<
math
::
ColFormat
::
kOCF
,
DeviceContext
,
T
>
f
;
auto
&
dev_ctx
=
ctx
.
template
device_context
<
DeviceContext
>();
f
(
dev_ctx
,
src
,
dilations
,
strides
,
paddings
,
&
dst
);
}
out
->
Resize
(
out_dims
);
...
...
@@ -87,7 +88,7 @@ class BlockExpandKernel : public framework::OpKernel<T> {
}
};
template
<
typename
Place
,
typename
T
>
template
<
typename
DeviceContext
,
typename
T
>
class
BlockExpandGradKernel
:
public
framework
::
OpKernel
<
T
>
{
public:
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
...
...
@@ -98,7 +99,8 @@ class BlockExpandGradKernel : public framework::OpKernel<T> {
d_x
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
auto
x_v
=
framework
::
EigenVector
<
T
>::
Flatten
(
*
d_x
);
x_v
.
device
(
ctx
.
GetEigenDevice
<
Place
>
())
=
x_v
.
constant
(
0.0
);
auto
&
place
=
*
ctx
.
template
device_context
<
DeviceContext
>().
eigen_device
();
x_v
.
device
(
place
)
=
x_v
.
constant
(
0.0
);
auto
in_dim
=
in
->
dims
();
int
batch_size
=
in_dim
[
0
];
...
...
@@ -131,8 +133,9 @@ class BlockExpandGradKernel : public framework::OpKernel<T> {
const
Tensor
src
=
d_out
->
Slice
(
i
,
i
+
1
).
Resize
(
{
output_height
,
output_width
,
img_channels
,
block_height
,
block_width
});
math
::
Col2ImFunctor
<
math
::
ColFormat
::
kOCF
,
Place
,
T
>
f
;
f
(
ctx
.
device_context
(),
src
,
dilations
,
strides
,
paddings
,
&
dst
);
math
::
Col2ImFunctor
<
math
::
ColFormat
::
kOCF
,
DeviceContext
,
T
>
f
;
auto
&
dev_ctx
=
ctx
.
template
device_context
<
DeviceContext
>();
f
(
dev_ctx
,
src
,
dilations
,
strides
,
paddings
,
&
dst
);
}
d_out
->
Resize
(
d_out_dims
);
}
...
...
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录