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
OPTHREE
Paddle
提交
67ce5864
P
Paddle
项目概览
OPTHREE
/
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
体验新版 GitCode,发现更多精彩内容 >>
提交
67ce5864
编写于
3月 21, 2018
作者:
W
wangyang59
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
gpu implementation of bilinear interp
上级
f67f0cae
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
96 addition
and
1 deletion
+96
-1
paddle/fluid/operators/bilinear_interp_op.cu
paddle/fluid/operators/bilinear_interp_op.cu
+95
-0
paddle/fluid/operators/bilinear_interp_op.h
paddle/fluid/operators/bilinear_interp_op.h
+1
-1
未找到文件。
paddle/fluid/operators/bilinear_interp_op.cu
0 → 100644
浏览文件 @
67ce5864
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License. */
#include "hl_cnn.h"
#include "paddle/fluid/operators/bilinear_interp_op.h"
namespace
paddle
{
namespace
operators
{
template
<
typename
T
>
class
BilinearInterpOpCUDAKernel
:
public
framework
::
OpKernel
<
T
>
{
public:
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
PADDLE_ENFORCE
(
platform
::
is_gpu_place
(
ctx
.
GetPlace
()),
"This kernel only runs on GPU device."
);
auto
*
input_t
=
ctx
.
Input
<
Tensor
>
(
"X"
);
// float tensor
auto
*
output_t
=
ctx
.
Output
<
Tensor
>
(
"Out"
);
// float tensor
auto
*
input
=
input_t
->
data
<
T
>
();
auto
*
output
=
output_t
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
int
out_h
=
ctx
.
Attr
<
int
>
(
"out_h"
);
int
out_w
=
ctx
.
Attr
<
int
>
(
"out_w"
);
int
batch_size
=
input_t
->
dims
()[
0
];
int
channels
=
input_t
->
dims
()[
1
];
int
in_h
=
input_t
->
dims
()[
2
];
int
in_w
=
input_t
->
dims
()[
3
];
int
in_hw
=
in_h
*
in_w
;
int
out_hw
=
out_h
*
out_w
;
int
in_chw
=
channels
*
in_hw
;
int
out_chw
=
channels
*
out_hw
;
T
ratio_h
=
(
out_h
>
1
)
?
static_cast
<
T
>
(
in_h
-
1
)
/
(
out_h
-
1
)
:
0.
f
;
T
ratio_w
=
(
out_w
>
1
)
?
static_cast
<
T
>
(
in_w
-
1
)
/
(
out_w
-
1
)
:
0.
f
;
if
(
in_h
==
out_h
&&
in_w
==
out_w
)
{
memcpy
(
output
,
input
,
input_t
->
numel
()
*
sizeof
(
T
));
}
else
{
hl_bilinear_forward
(
input
,
in_h
,
in_w
,
batch_size
,
in_chw
,
output
,
out_h
,
out_w
,
batch_size
,
out_chw
,
channels
,
ratio_h
,
ratio_w
);
}
}
};
template
<
typename
T
>
class
BilinearInterpGradOpCUDAKernel
:
public
framework
::
OpKernel
<
T
>
{
public:
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
auto
*
d_input_t
=
ctx
.
Output
<
Tensor
>
(
framework
::
GradVarName
(
"X"
));
auto
*
d_output_t
=
ctx
.
Input
<
Tensor
>
(
framework
::
GradVarName
(
"Out"
));
auto
*
d_input
=
d_input_t
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
auto
*
d_output
=
d_output_t
->
data
<
T
>
();
int
out_h
=
ctx
.
Attr
<
int
>
(
"out_h"
);
int
out_w
=
ctx
.
Attr
<
int
>
(
"out_w"
);
int
batch_size
=
d_input_t
->
dims
()[
0
];
int
channels
=
d_input_t
->
dims
()[
1
];
int
in_h
=
d_input_t
->
dims
()[
2
];
int
in_w
=
d_input_t
->
dims
()[
3
];
int
in_hw
=
in_h
*
in_w
;
int
out_hw
=
out_h
*
out_w
;
int
in_chw
=
channels
*
in_hw
;
int
out_chw
=
channels
*
out_hw
;
T
ratio_h
=
(
out_h
>
1
)
?
static_cast
<
T
>
(
in_h
-
1
)
/
(
out_h
-
1
)
:
0.
f
;
T
ratio_w
=
(
out_w
>
1
)
?
static_cast
<
T
>
(
in_w
-
1
)
/
(
out_w
-
1
)
:
0.
f
;
if
(
in_h
==
out_h
&&
in_w
==
out_w
)
{
memcpy
(
d_input
,
d_output
,
d_input_t
->
numel
()
*
sizeof
(
T
));
}
else
{
hl_bilinear_backward
(
d_input
,
in_h
,
in_w
,
batch_size
,
in_chw
,
d_output
,
out_h
,
out_w
,
batch_size
,
out_chw
,
channels
,
ratio_h
,
ratio_w
);
}
}
};
}
// namespace operators
}
// namespace paddle
namespace
ops
=
paddle
::
operators
;
REGISTER_OP_CUDA_KERNEL
(
bilinear_interp
,
ops
::
BilinearInterpOpCUDAKernel
<
float
>
);
REGISTER_OP_CUDA_KERNEL
(
bilinear_interp_grad
,
ops
::
BilinearInterpGradOpCUDAKernel
<
float
>
);
\ No newline at end of file
paddle/fluid/operators/bilinear_interp_op.h
浏览文件 @
67ce5864
...
...
@@ -105,7 +105,7 @@ class BilinearInterpGradKernel : public framework::OpKernel<T> {
T
ratio_w
=
(
out_w
>
1
)
?
static_cast
<
T
>
(
in_w
-
1
)
/
(
out_w
-
1
)
:
0.
f
;
if
(
in_h
==
out_h
&&
in_w
==
out_w
)
{
memcpy
(
d_input
,
d_output
,
product
(
d_input_t
->
dims
()
)
*
sizeof
(
T
));
memcpy
(
d_input
,
d_output
,
d_input_t
->
numel
(
)
*
sizeof
(
T
));
}
else
{
for
(
int
k
=
0
;
k
<
batch_size
;
++
k
)
{
// loop for batches
for
(
int
i
=
0
;
i
<
out_h
;
++
i
)
{
// loop for images
...
...
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