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
PaddlePaddle
Paddle
提交
d7985052
P
Paddle
项目概览
PaddlePaddle
/
Paddle
1 年多 前同步成功
通知
2302
Star
20931
Fork
5422
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
1423
列表
看板
标记
里程碑
合并请求
543
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
1,423
Issue
1,423
列表
看板
标记
里程碑
合并请求
543
合并请求
543
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
未验证
提交
d7985052
编写于
9月 10, 2021
作者:
C
ceci3
提交者:
GitHub
9月 10, 2021
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
fix bn/in/squeeze/syncbn extra (#35502)
* fix bn/in/squeeze/syncbn extra * update bn * update * update
上级
3896bdbd
变更
4
隐藏空白更改
内联
并排
Showing
4 changed file
with
16 addition
and
8 deletion
+16
-8
paddle/fluid/operators/batch_norm_op.cc
paddle/fluid/operators/batch_norm_op.cc
+6
-3
paddle/fluid/operators/instance_norm_op.cc
paddle/fluid/operators/instance_norm_op.cc
+4
-2
paddle/fluid/operators/mkldnn/batch_norm_mkldnn_op.cc
paddle/fluid/operators/mkldnn/batch_norm_mkldnn_op.cc
+3
-1
paddle/fluid/operators/squeeze_op.cc
paddle/fluid/operators/squeeze_op.cc
+3
-2
未找到文件。
paddle/fluid/operators/batch_norm_op.cc
浏览文件 @
d7985052
...
...
@@ -257,13 +257,16 @@ void BatchNormOpMaker::Make() {
AddOutput
(
"ReserveSpace"
,
"Reserve GPU space for triggering the new semi-persistent "
"NHWC kernel"
)
.
AsDispensable
();
.
AsDispensable
()
.
AsExtra
();
AddAttr
<
bool
>
(
"use_mkldnn"
,
"(bool, default false) Only used in mkldnn kernel"
)
.
SetDefault
(
false
);
.
SetDefault
(
false
)
.
AsExtra
();
AddAttr
<
bool
>
(
"fuse_with_relu"
,
"(bool, default false) Only used in mkldnn kernel"
)
.
SetDefault
(
false
);
.
SetDefault
(
false
)
.
AsExtra
();
AddAttr
<
bool
>
(
"use_global_stats"
,
"(bool, default false) Whether to use global mean and "
"variance. In inference or test mode, set use_global_stats "
...
...
paddle/fluid/operators/instance_norm_op.cc
浏览文件 @
d7985052
...
...
@@ -149,11 +149,13 @@ void InstanceNormOpMaker::Make() {
AddOutput
(
"SavedMean"
,
"Mean of the current mini batch, "
"will apply to output when training"
)
.
AsIntermediate
();
.
AsIntermediate
()
.
AsExtra
();
AddOutput
(
"SavedVariance"
,
"Variance of the current mini batch, "
"will apply to output when training"
)
.
AsIntermediate
();
.
AsIntermediate
()
.
AsExtra
();
AddComment
(
R"DOC(
Instance Normalization.
...
...
paddle/fluid/operators/mkldnn/batch_norm_mkldnn_op.cc
浏览文件 @
d7985052
...
...
@@ -47,7 +47,9 @@ class BatchNormMKLDNNHandler : public platform::MKLDNNHandlerNoCachingT<
mkldnn
::
batch_normalization_backward
>
(
mkldnn_engine
,
ctx
.
GetPlace
())
{
const
float
epsilon
=
ctx
.
Attr
<
float
>
(
"epsilon"
);
const
bool
fuse_with_relu
=
ctx
.
Attr
<
bool
>
(
"fuse_with_relu"
);
const
bool
fuse_with_relu
=
ctx
.
HasAttr
(
"fuse_with_relu"
)
?
ctx
.
Attr
<
bool
>
(
"fuse_with_relu"
)
:
false
;
std
::
vector
<
std
::
string
>
DataLayout_error_msg
=
{
"kNHWC"
,
"kNCHW"
,
"kAnyLayout"
,
"kMKLDNN"
};
...
...
paddle/fluid/operators/squeeze_op.cc
浏览文件 @
d7985052
...
...
@@ -225,7 +225,7 @@ class Squeeze2Op : public framework::OperatorWithKernel {
ctx
->
ShareLoD
(
"X"
,
"Out"
);
}
OP_INOUT_CHECK
(
ctx
->
HasOutput
(
"XShape"
),
"Output"
,
"XShape"
,
"Squeeze2"
)
;
if
(
!
ctx
->
HasOutput
(
"XShape"
))
return
;
std
::
vector
<
int64_t
>
xshape_dims
(
x_dims
.
size
()
+
1
);
xshape_dims
[
0
]
=
0
;
...
...
@@ -323,7 +323,8 @@ class Squeeze2OpMaker : public SqueezeOpMaker {
AddOutput
(
"XShape"
,
"XShape is just used to store the shape and lod of X, which will "
"be used in SqueezeGradOp."
)
.
AsIntermediate
();
.
AsIntermediate
()
.
AsExtra
();
}
};
...
...
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