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
Crayon鑫
Paddle
提交
aa2635fe
P
Paddle
项目概览
Crayon鑫
/
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
aa2635fe
编写于
5月 09, 2018
作者:
K
Kexin Zhao
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
clean code
上级
cbf502e5
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
1 addition
and
14 deletion
+1
-14
paddle/fluid/operators/load_op.cc
paddle/fluid/operators/load_op.cc
+1
-10
paddle/fluid/operators/save_load_op_test.cc
paddle/fluid/operators/save_load_op_test.cc
+0
-4
未找到文件。
paddle/fluid/operators/load_op.cc
浏览文件 @
aa2635fe
...
...
@@ -12,7 +12,6 @@ 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 <fstream>
#include <iostream>
#include "paddle/fluid/framework/data_type_transform.h"
#include "paddle/fluid/framework/op_registry.h"
...
...
@@ -53,9 +52,6 @@ class LoadOp : public framework::OperatorBase {
auto
in_dtype
=
framework
::
ToDataType
(
tensor
->
type
());
auto
out_dtype
=
load_as_fp16
?
framework
::
proto
::
VarType
::
FP16
:
in_dtype
;
std
::
cout
<<
"In load op: "
<<
std
::
endl
;
std
::
cout
<<
"before conversion block"
<<
std
::
endl
;
if
(
in_dtype
!=
out_dtype
)
{
// convert to float16 tensor
auto
in_kernel_type
=
framework
::
OpKernelType
(
in_dtype
,
place
);
...
...
@@ -63,20 +59,15 @@ class LoadOp : public framework::OperatorBase {
framework
::
LoDTensor
fp16_tensor
;
// copy LoD info to the new tensor
fp16_tensor
.
set_lod
(
tensor
->
lod
());
std
::
cout
<<
"before TransDataType"
<<
std
::
endl
;
framework
::
TransDataType
(
in_kernel_type
,
out_kernel_type
,
*
tensor
,
&
fp16_tensor
);
std
::
cout
<<
"after TransDataType"
<<
std
::
endl
;
// reset output tensor
out_var
->
Clear
();
tensor
=
out_var
->
GetMutable
<
framework
::
LoDTensor
>
();
tensor
->
set_lod
(
fp16_tensor
.
lod
());
std
::
cout
<<
"before TransDataType"
<<
std
::
endl
;
tensor
->
ShareDataWith
(
fp16_tensor
);
std
::
cout
<<
"after TransDataType"
<<
std
::
endl
;
}
std
::
cout
<<
"Out of load op: "
<<
std
::
endl
;
}
};
...
...
paddle/fluid/operators/save_load_op_test.cc
浏览文件 @
aa2635fe
...
...
@@ -127,17 +127,13 @@ TEST(LoadFP16Op, CPU) {
auto
load_var
=
scope
.
Var
(
"out_var"
);
auto
load_op
=
paddle
::
framework
::
OpRegistry
::
CreateOp
(
"load"
,
{},
{{
"Out"
,
{
"out_var"
}}},
attrs
);
LOG
(
INFO
)
<<
"before load op run"
;
load_op
->
Run
(
scope
,
place
);
LOG
(
INFO
)
<<
"after load op run"
;
auto
target
=
load_var
->
Get
<
paddle
::
framework
::
LoDTensor
>
();
paddle
::
platform
::
float16
*
actual
=
target
.
data
<
paddle
::
platform
::
float16
>
();
LOG
(
INFO
)
<<
"after target->data"
;
for
(
int64_t
i
=
0
;
i
<
tensor
->
numel
();
++
i
)
{
EXPECT_EQ
(
expect
[
i
],
static_cast
<
float
>
(
actual
[
i
]));
}
LOG
(
INFO
)
<<
"after expect equal"
;
auto
&
actual_lod
=
target
.
lod
();
EXPECT_EQ
(
expect_lod
.
size
(),
actual_lod
.
size
());
...
...
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