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
PaddleDetection
提交
489f581d
P
PaddleDetection
项目概览
PaddlePaddle
/
PaddleDetection
大约 1 年 前同步成功
通知
695
Star
11112
Fork
2696
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
184
列表
看板
标记
里程碑
合并请求
40
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
P
PaddleDetection
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
184
Issue
184
列表
看板
标记
里程碑
合并请求
40
合并请求
40
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
489f581d
编写于
6月 13, 2018
作者:
T
tensor-tang
提交者:
GitHub
6月 13, 2018
浏览文件
操作
浏览文件
下载
差异文件
Merge pull request #11098 from mozga-intel/mozga-intel/Batch_norm_mkldnn_layout
MKLDNN layout: Support for batch norm operator
上级
f9b5365f
7d564356
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
193 addition
and
157 deletion
+193
-157
paddle/fluid/operators/batch_norm_mkldnn_op.cc
paddle/fluid/operators/batch_norm_mkldnn_op.cc
+180
-146
paddle/fluid/operators/batch_norm_op.cc
paddle/fluid/operators/batch_norm_op.cc
+13
-11
未找到文件。
paddle/fluid/operators/batch_norm_mkldnn_op.cc
浏览文件 @
489f581d
...
@@ -19,10 +19,17 @@ limitations under the License. */
...
@@ -19,10 +19,17 @@ limitations under the License. */
namespace
paddle
{
namespace
paddle
{
namespace
operators
{
namespace
operators
{
using
Tensor
=
framework
::
Tensor
;
using
batch_norm_bwd
=
mkldnn
::
batch_normalization_backward
;
using
batch_norm_fwd
=
mkldnn
::
batch_normalization_forward
;
using
framework
::
DataLayout
;
using
framework
::
Tensor
;
using
mkldnn
::
memory
;
using
mkldnn
::
primitive
;
using
mkldnn
::
reorder
;
using
mkldnn
::
stream
;
using
paddle
::
platform
::
MKLDNNDeviceContext
;
using
paddle
::
platform
::
MKLDNNDeviceContext
;
using
paddle
::
platform
::
MKLDNNMemDesc
;
using
paddle
::
platform
::
MKLDNNMemDesc
;
using
mkldnn
::
memory
;
using
platform
::
to_void_cast
;
template
<
typename
T
>
template
<
typename
T
>
using
EigenArrayMap
=
using
EigenArrayMap
=
...
@@ -64,21 +71,12 @@ void run_batch_norm_op(Args &&... args) {
...
@@ -64,21 +71,12 @@ void run_batch_norm_op(Args &&... args) {
mkldnn
::
stream
(
mkldnn
::
stream
::
kind
::
eager
).
submit
(
pipeline
).
wait
();
mkldnn
::
stream
(
mkldnn
::
stream
::
kind
::
eager
).
submit
(
pipeline
).
wait
();
}
}
template
<
typename
T
>
inline
void
*
cast_const_to_void
(
const
T
*
t
)
{
return
static_cast
<
void
*>
(
const_cast
<
T
*>
(
t
));
}
}
// namespace
}
// namespace
template
<
typename
T
>
template
<
typename
T
>
class
BatchNormMKLDNNOpKernel
:
public
paddle
::
framework
::
OpKernel
<
T
>
{
class
BatchNormMKLDNNOpKernel
:
public
paddle
::
framework
::
OpKernel
<
T
>
{
public:
public:
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
auto
data_layout_str
=
ctx
.
Attr
<
std
::
string
>
(
"data_layout"
);
auto
data_layout
=
framework
::
StringToDataLayout
(
data_layout_str
);
PADDLE_ENFORCE
(
data_layout
==
framework
::
DataLayout
::
kNCHW
,
"MKLDNN batch normalization handles only NCHW data layout"
);
const
float
epsilon
=
ctx
.
Attr
<
float
>
(
"epsilon"
);
const
float
epsilon
=
ctx
.
Attr
<
float
>
(
"epsilon"
);
const
float
momentum
=
ctx
.
Attr
<
float
>
(
"momentum"
);
const
float
momentum
=
ctx
.
Attr
<
float
>
(
"momentum"
);
const
bool
is_test
=
ctx
.
Attr
<
bool
>
(
"is_test"
);
const
bool
is_test
=
ctx
.
Attr
<
bool
>
(
"is_test"
);
...
@@ -99,41 +97,53 @@ class BatchNormMKLDNNOpKernel : public paddle::framework::OpKernel<T> {
...
@@ -99,41 +97,53 @@ class BatchNormMKLDNNOpKernel : public paddle::framework::OpKernel<T> {
const
auto
*
scale
=
ctx
.
Input
<
Tensor
>
(
"Scale"
);
const
auto
*
scale
=
ctx
.
Input
<
Tensor
>
(
"Scale"
);
const
auto
*
shift
=
ctx
.
Input
<
Tensor
>
(
"Bias"
);
const
auto
*
shift
=
ctx
.
Input
<
Tensor
>
(
"Bias"
);
y
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
PADDLE_ENFORCE
(
x
->
layout
()
==
DataLayout
::
kMKLDNN
&&
mean_out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
x
->
format
()
!=
memory
::
format
::
format_undef
,
variance_out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
"Wrong layout/format set for Input x tensor"
);
const
T
*
x_data
=
x
->
data
<
T
>
();
const
T
*
mean_data
=
mean
->
data
<
T
>
();
const
T
*
variance_data
=
variance
->
data
<
T
>
();
T
*
y_data
=
y
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
T
*
mean_out_data
=
mean_out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
T
*
variance_out_data
=
variance_out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
T
*
batch_mean_data
=
nullptr
;
T
*
batch_variance_data
=
nullptr
;
if
(
!
is_test
)
{
if
(
!
is_test
)
{
batch_mean
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
batch_mean
_data
=
batch_mean
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
batch_variance
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
batch_variance
_data
=
batch_variance
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
}
}
auto
propagation
=
is_test
==
true
?
mkldnn
::
prop_kind
::
forward_scoring
auto
propagation
=
is_test
==
true
?
mkldnn
::
prop_kind
::
forward_scoring
:
mkldnn
::
prop_kind
::
forward_training
;
:
mkldnn
::
prop_kind
::
forward_training
;
auto
dims
=
paddle
::
framework
::
vectorize2int
(
x
->
dims
());
auto
src_tz
=
paddle
::
framework
::
vectorize2int
(
x
->
dims
());
auto
scale_tz
=
paddle
::
framework
::
vectorize2int
(
scale
->
dims
());
auto
src_md
=
PADDLE_ENFORCE
(
scale_tz
.
size
()
==
1
,
"Dims of scale tensor is NOT 1"
);
MKLDNNMemDesc
(
dims
,
memory
::
data_type
::
f32
,
memory
::
format
::
nchw
);
const
unsigned
int
ic
=
scale_tz
[
0
];
auto
dst_md
=
MKLDNNMemDesc
(
dims
,
memory
::
data_type
::
f32
,
memory
::
format
::
nchw
);
auto
src_pd
=
mkldnn
::
memory
::
primitive_desc
{
src_md
,
mkldnn_engine
};
auto
dst_pd
=
mkldnn
::
memory
::
primitive_desc
{
dst_md
,
mkldnn_engine
};
auto
src
=
mkldnn
::
memory
{
src_pd
,
cast_const_to_void
(
x
->
data
<
T
>
())};
auto
dst
=
mkldnn
::
memory
{
dst_pd
,
y
->
data
<
T
>
()};
unsigned
flags
=
mkldnn
::
use_scale_shift
;
unsigned
flags
=
mkldnn
::
use_scale_shift
;
if
(
is_test
)
flags
|=
mkldnn
::
use_global_stats
;
if
(
is_test
)
flags
|=
mkldnn
::
use_global_stats
;
// create mkldnn memory from input x tensor
auto
src_memory
=
memory
({{{
src_tz
},
memory
::
data_type
::
f32
,
x
->
format
()},
mkldnn_engine
},
to_void_cast
(
x_data
));
// create primitive descriptor for batch norm forward
using
bn_fwd_types
=
bn_type_traits
<
mkldnn
::
batch_normalization_forward
>
;
using
bn_fwd_types
=
bn_type_traits
<
mkldnn
::
batch_normalization_forward
>
;
auto
batch_norm_fwd_desc
=
auto
batch_norm_fwd_desc
=
bn_fwd_types
::
op_desc
{
bn_fwd_types
::
op_desc
{
propagation
,
src_md
,
epsilon
,
flags
};
propagation
,
src_memory
.
get_primitive_desc
().
desc
(),
epsilon
,
flags
};
auto
batch_norm_fwd_pd
=
std
::
shared_ptr
<
batch_norm_fwd
::
primitive_desc
>
batch_norm_fwd_pd
=
bn_fwd_types
::
op_prim
{
batch_norm_fwd_desc
,
mkldnn_engine
};
std
::
shared_ptr
<
batch_norm_fwd
::
primitive_desc
>
(
new
batch_norm_fwd
::
primitive_desc
(
batch_norm_fwd_desc
,
mkldnn_engine
));
const
unsigned
int
ic
=
dims
[
1
];
// Save the pd to be used in backward pass
const
std
::
string
key
=
ctx
.
op
().
Output
(
"SavedMean"
);
const
std
::
string
key_batch_norm_fwd_pd
=
key
+
"@bn_fwd_pd"
;
dev_ctx
.
SetBlob
(
key_batch_norm_fwd_pd
,
batch_norm_fwd_pd
);
// MKLDNN requires a single piece of memory for scale and shift/bias data
// MKLDNN requires a single piece of memory for scale and shift/bias data
const
size_t
scaleshift_size
=
2
*
ic
;
const
size_t
scaleshift_size
=
2
*
ic
;
...
@@ -143,73 +153,58 @@ class BatchNormMKLDNNOpKernel : public paddle::framework::OpKernel<T> {
...
@@ -143,73 +153,58 @@ class BatchNormMKLDNNOpKernel : public paddle::framework::OpKernel<T> {
copy_to_weights
(
scale
->
data
<
T
>
(),
scale
->
data
<
T
>
()
+
ic
,
shift
->
data
<
T
>
(),
copy_to_weights
(
scale
->
data
<
T
>
(),
scale
->
data
<
T
>
()
+
ic
,
shift
->
data
<
T
>
(),
shift
->
data
<
T
>
()
+
ic
,
&
scaleshift_data
);
shift
->
data
<
T
>
()
+
ic
,
&
scaleshift_data
);
auto
scaleshift_memory
=
mkldnn
::
memory
{
// crate mkldnn memory for weights(scale/shift)
batch_norm_fwd_pd
.
weights_primitive_desc
(),
scaleshift_data
.
data
()};
auto
scaleshift_memory
=
memory
(
batch_norm_fwd_pd
->
weights_primitive_desc
(),
scaleshift_data
.
data
());
if
(
is_test
)
{
// create mkldnn memory for output y tensor
auto
mean_memory
=
mkldnn
::
memory
{
batch_norm_fwd_pd
.
mean_primitive_desc
(),
auto
dst_memory
=
memory
(
batch_norm_fwd_pd
->
dst_primitive_desc
(),
y_data
);
cast_const_to_void
(
mean
->
data
<
T
>
())};
if
(
is_test
)
{
// create mkldnn memory for stats (as input)
auto
mean_memory
=
memory
(
batch_norm_fwd_pd
->
mean_primitive_desc
(),
to_void_cast
(
mean_data
));
auto
variance_memory
=
auto
variance_memory
=
m
kldnn
::
memory
{
batch_norm_fwd_pd
.
variance_primitive_desc
(),
m
emory
(
batch_norm_fwd_pd
->
variance_primitive_desc
(),
cast_const_to_void
(
variance
->
data
<
T
>
())}
;
to_void_cast
(
variance_data
))
;
run_batch_norm_op
<
typename
bn_fwd_types
::
op_type
>
(
run_batch_norm_op
<
typename
bn_fwd_types
::
op_type
>
(
batch_norm_fwd_pd
,
src
,
(
const
mkldnn
::
primitive
::
at
&
)
mean_memory
,
*
batch_norm_fwd_pd
,
src_memory
,
(
const
mkldnn
::
primitive
::
at
&
)
mean_memory
,
(
const
mkldnn
::
primitive
::
at
&
)
variance_memory
,
scaleshift_memory
,
(
const
mkldnn
::
primitive
::
at
&
)
variance_memory
,
scaleshift_memory
,
dst
);
dst
_memory
);
}
else
{
}
else
{
// create mkldnn memory for stats (as output)
auto
mean_memory
=
auto
mean_memory
=
mkldnn
::
memory
{
batch_norm_fwd_pd
.
mean_primitive_desc
(),
memory
(
batch_norm_fwd_pd
->
mean_primitive_desc
(),
batch_mean_data
);
cast_const_to_void
(
batch_mean
->
data
<
T
>
())};
auto
variance_memory
=
memory
(
batch_norm_fwd_pd
->
variance_primitive_desc
(),
batch_variance_data
);
auto
variance_memory
=
mkldnn
::
memory
{
batch_norm_fwd_pd
.
variance_primitive_desc
(),
cast_const_to_void
(
batch_variance
->
data
<
T
>
())};
run_batch_norm_op
<
bn_fwd_types
::
op_type
>
(
batch_norm_fwd_pd
,
src
,
run_batch_norm_op
<
bn_fwd_types
::
op_type
>
(
*
batch_norm_fwd_pd
,
src_memory
,
scaleshift_memory
,
dst
,
scaleshift_memory
,
dst
_memory
,
mean_memory
,
variance_memory
);
mean_memory
,
variance_memory
);
}
}
if
(
!
is_test
)
{
if
(
!
is_test
)
{
const
unsigned
int
in
=
dims
[
0
];
// mkldnn only compute stats for current batch
const
unsigned
int
sample_size
=
x
->
numel
()
/
in
/
ic
;
// so we need compute momentum stats via Eigen lib
EigenVectorArrayMap
<
T
>
batch_mean_e
(
batch_mean_data
,
ic
);
// saved_xx is use just in this batch of data
EigenVectorArrayMap
<
T
>
batch_variance_e
(
batch_variance_data
,
ic
);
EigenVectorArrayMap
<
T
>
saved_mean_e
(
ConstEigenVectorArrayMap
<
T
>
mean_e
(
mean_data
,
ic
);
batch_mean
->
mutable_data
<
T
>
(
ctx
.
GetPlace
()),
ic
);
ConstEigenVectorArrayMap
<
T
>
variance_e
{
variance_data
,
ic
};
EigenVectorArrayMap
<
T
>
saved_variance_e
(
batch_variance
->
mutable_data
<
T
>
(
ctx
.
GetPlace
()),
ic
);
EigenVectorArrayMap
<
T
>
running_mean_e
(
mean_out_data
,
ic
);
saved_mean_e
.
setZero
();
EigenVectorArrayMap
<
T
>
running_variance_e
(
variance_out_data
,
ic
);
saved_variance_e
.
setZero
();
const
unsigned
int
x_arr_size
=
in
*
ic
;
ConstEigenArrayMap
<
T
>
x_arr
(
x
->
data
<
T
>
(),
sample_size
,
x_arr_size
);
for
(
unsigned
int
nc
=
0
;
nc
<
x_arr_size
;
++
nc
)
{
saved_mean_e
(
nc
%
ic
)
+=
x_arr
.
col
(
nc
).
sum
();
}
saved_mean_e
/=
in
*
sample_size
;
for
(
unsigned
int
nc
=
0
;
nc
<
x_arr_size
;
++
nc
)
{
saved_variance_e
(
nc
%
ic
)
+=
(
x_arr
.
col
(
nc
)
-
saved_mean_e
(
nc
%
ic
)).
matrix
().
squaredNorm
();
}
saved_variance_e
/=
in
*
sample_size
;
ConstEigenVectorArrayMap
<
T
>
mean_arr
{
mean
->
data
<
T
>
(),
ic
};
ConstEigenVectorArrayMap
<
T
>
variance_arr
{
variance
->
data
<
T
>
(),
ic
};
EigenVectorArrayMap
<
T
>
running_mean_arr
(
mean_out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
()),
ic
);
EigenVectorArrayMap
<
T
>
running_var_arr
(
variance_out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
()),
ic
);
auto
one_minus_momentum
=
1.
-
momentum
;
auto
one_minus_momentum
=
1.
-
momentum
;
running_mean_arr
=
running_mean_e
=
mean_e
*
momentum
+
batch_mean_e
*
one_minus_momentum
;
mean_arr
*
momentum
+
saved_mean_e
*
one_minus_momentum
;
running_variance_e
=
running_var_arr
=
variance_e
*
momentum
+
batch_variance_e
*
one_minus_momentum
;
variance_arr
*
momentum
+
saved_variance_e
*
one_minus_momentum
;
}
}
y
->
set_layout
(
DataLayout
::
kMKLDNN
);
y
->
set_format
(
(
memory
::
format
)
dst_memory
.
get_primitive_desc
().
desc
().
data
.
format
);
}
}
};
};
...
@@ -217,11 +212,6 @@ template <typename T>
...
@@ -217,11 +212,6 @@ template <typename T>
class
BatchNormMKLDNNGradOpKernel
:
public
paddle
::
framework
::
OpKernel
<
T
>
{
class
BatchNormMKLDNNGradOpKernel
:
public
paddle
::
framework
::
OpKernel
<
T
>
{
public:
public:
void
Compute
(
const
paddle
::
framework
::
ExecutionContext
&
ctx
)
const
override
{
void
Compute
(
const
paddle
::
framework
::
ExecutionContext
&
ctx
)
const
override
{
auto
data_layout_str
=
ctx
.
Attr
<
std
::
string
>
(
"data_layout"
);
auto
data_layout
=
framework
::
StringToDataLayout
(
data_layout_str
);
PADDLE_ENFORCE
(
data_layout
==
framework
::
DataLayout
::
kNCHW
,
"MKLDNN batch normalization handles only NCHW data layout"
);
auto
&
dev_ctx
=
ctx
.
template
device_context
<
MKLDNNDeviceContext
>();
auto
&
dev_ctx
=
ctx
.
template
device_context
<
MKLDNNDeviceContext
>();
auto
mkldnn_engine
=
dev_ctx
.
GetEngine
();
auto
mkldnn_engine
=
dev_ctx
.
GetEngine
();
...
@@ -238,88 +228,132 @@ class BatchNormMKLDNNGradOpKernel : public paddle::framework::OpKernel<T> {
...
@@ -238,88 +228,132 @@ class BatchNormMKLDNNGradOpKernel : public paddle::framework::OpKernel<T> {
auto
*
diff_scale
=
ctx
.
Output
<
Tensor
>
(
framework
::
GradVarName
(
"Scale"
));
auto
*
diff_scale
=
ctx
.
Output
<
Tensor
>
(
framework
::
GradVarName
(
"Scale"
));
auto
*
diff_shift
=
ctx
.
Output
<
Tensor
>
(
framework
::
GradVarName
(
"Bias"
));
auto
*
diff_shift
=
ctx
.
Output
<
Tensor
>
(
framework
::
GradVarName
(
"Bias"
));
diff_x
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
PADDLE_ENFORCE
(
diff_y
->
layout
()
==
DataLayout
::
kMKLDNN
&&
diff_scale
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
diff_y
->
format
()
!=
memory
::
format
::
format_undef
,
diff_shift
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
"Wrong layout/format set for Input diff_y tensor"
);
const
T
*
x_data
=
x
->
data
<
T
>
();
const
T
*
diff_y_data
=
diff_y
->
data
<
T
>
();
const
T
*
batch_mean_data
=
batch_mean
->
data
<
T
>
();
const
T
*
batch_variance_data
=
batch_variance
->
data
<
T
>
();
const
T
*
scale_data
=
scale
->
data
<
T
>
();
const
T
*
shift_data
=
shift
->
data
<
T
>
();
T
*
diff_x_data
=
diff_x
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
T
*
diff_scale_data
=
diff_scale
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
T
*
diff_shift_data
=
diff_shift
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
auto
src_tz
=
paddle
::
framework
::
vectorize2int
(
x
->
dims
());
auto
diff_src_tz
=
src_tz
;
auto
dst_tz
=
src_tz
;
auto
diff_dst_tz
=
dst_tz
;
auto
scale_tz
=
paddle
::
framework
::
vectorize2int
(
scale
->
dims
());
PADDLE_ENFORCE
(
scale_tz
.
size
()
==
1
,
"Dims of scale tensor is NOT 1"
);
const
unsigned
int
ic
=
scale_tz
[
0
];
// Retrieve bn_fwd_pd from device context
const
std
::
string
key
=
ctx
.
op
().
Input
(
"SavedMean"
);
const
std
::
string
key_batch_norm_fwd_pd
=
key
+
"@bn_fwd_pd"
;
auto
batch_norm_fwd_pd
=
std
::
static_pointer_cast
<
batch_norm_fwd
::
primitive_desc
>
(
dev_ctx
.
GetBlob
(
key_batch_norm_fwd_pd
));
PADDLE_ENFORCE
(
batch_norm_fwd_pd
!=
nullptr
,
"Fail to find batch_norm_fwd_pd in device context"
);
auto
dims
=
paddle
::
framework
::
vectorize2int
(
x
->
dims
());
using
bn_bwd_types
=
bn_type_traits
<
mkldnn
::
batch_normalization_backward
>
;
unsigned
flags
=
mkldnn
::
use_scale_shift
|
!
mkldnn
::
use_global_stats
;
auto
src_md
=
// create mkldnn memory from input diff_y tensor
MKLDNNMemDesc
(
dims
,
memory
::
data_type
::
f32
,
memory
::
format
::
nchw
);
auto
user_diff_dst_memory
=
auto
dst_md
=
memory
({{{
diff_dst_tz
},
memory
::
data_type
::
f32
,
diff_y
->
format
()},
MKLDNNMemDesc
(
dims
,
memory
::
data_type
::
f32
,
memory
::
format
::
nchw
);
mkldnn_engine
},
auto
diff_src_md
=
to_void_cast
(
diff_y_data
));
MKLDNNMemDesc
(
dims
,
memory
::
data_type
::
f32
,
memory
::
format
::
nchw
);
auto
diff_dst_md
=
MKLDNNMemDesc
(
dims
,
memory
::
data_type
::
f32
,
memory
::
format
::
nchw
);
using
bn_bwd_types
=
bn_type_traits
<
mkldnn
::
batch_normalization_backward
>
;
// create mkldnn memory from input x tensor
using
bn_fwd_types
=
bn_type_traits
<
mkldnn
::
batch_normalization_forward
>
;
auto
src_memory
=
memory
({{{
src_tz
},
memory
::
data_type
::
f32
,
x
->
format
()},
mkldnn_engine
},
to_void_cast
(
x_data
));
auto
batch_norm_fwd_desc
=
bn_fwd_types
::
op_desc
{
// for diff_dst, try to use same format as dst in forward pass
mkldnn
::
prop_kind
::
forward_training
,
src_md
,
epsilon
,
flags
};
auto
diff_dst_pd
=
batch_norm_fwd_pd
.
get
()
->
dst_primitive_desc
();
auto
batch_norm_fwd_pd
=
auto
diff_dst_md
=
diff_dst_pd
.
desc
();
bn_fwd_types
::
op_prim
{
batch_norm_fwd_desc
,
mkldnn_engine
};
// create primitive descriptor for batch norm backward
unsigned
flags
=
mkldnn
::
use_scale_shift
;
auto
batch_norm_bwd_desc
=
bn_bwd_types
::
op_desc
{
auto
batch_norm_bwd_desc
=
bn_bwd_types
::
op_desc
{
mkldnn
::
prop_kind
::
backward
,
diff_dst_md
,
dst_md
,
epsilon
,
flags
};
mkldnn
::
prop_kind
::
backward
,
diff_dst_md
,
src_memory
.
get_primitive_desc
().
desc
(),
epsilon
,
flags
};
auto
batch_norm_bwd_pd
=
bn_bwd_types
::
op_prim
{
auto
batch_norm_bwd_pd
=
bn_bwd_types
::
op_prim
{
batch_norm_bwd_desc
,
mkldnn_engine
,
batch_norm_fwd_pd
};
batch_norm_bwd_desc
,
mkldnn_engine
,
*
batch_norm_fwd_pd
};
auto
src
=
mkldnn
::
memory
{{
src_md
,
mkldnn_engine
},
// reorder user_diff_dst if it's not in preferred format
cast_const_to_void
(
x
->
data
<
T
>
())};
auto
diff_dst_memory
=
user_diff_dst_memory
;
primitive
reorder_diff_dst
;
auto
mean
=
mkldnn
::
memory
{
batch_norm_bwd_pd
.
mean_primitive_desc
(),
bool
is_diff_dst_reordered
=
false
;
cast_const_to_void
(
batch_mean
->
data
<
T
>
())};
if
(
diff_dst_pd
!=
user_diff_dst_memory
.
get_primitive_desc
())
{
diff_dst_memory
=
memory
(
diff_dst_pd
);
auto
variance
=
reorder_diff_dst
=
reorder
(
user_diff_dst_memory
,
diff_dst_memory
);
mkldnn
::
memory
{
batch_norm_bwd_pd
.
variance_primitive_desc
(),
is_diff_dst_reordered
=
true
;
cast_const_to_void
(
batch_variance
->
data
<
T
>
())};
}
auto
diff_dst
=
mkldnn
::
memory
{{
diff_dst_md
,
mkldnn_engine
},
cast_const_to_void
(
diff_y
->
data
<
T
>
())};
const
unsigned
int
ic
=
dims
[
1
];
// create mkldnn memory for input tensors (src/mean/variance)
auto
mean_memory
=
memory
(
batch_norm_bwd_pd
.
mean_primitive_desc
(),
to_void_cast
(
batch_mean_data
));
auto
variance_memory
=
memory
(
batch_norm_bwd_pd
.
variance_primitive_desc
(),
to_void_cast
(
batch_variance_data
));
// MKLDNN requires a single piece of memory for scale and shift/bias data
const
size_t
scaleshift_size
=
2
*
ic
;
const
size_t
scaleshift_size
=
2
*
ic
;
std
::
vector
<
T
>
scaleshift_data
;
std
::
vector
<
T
>
scaleshift_data
;
scaleshift_data
.
reserve
(
scaleshift_size
);
scaleshift_data
.
reserve
(
scaleshift_size
);
copy_to_weights
(
scale
->
data
<
T
>
(),
scale
->
data
<
T
>
()
+
ic
,
shift
->
data
<
T
>
()
,
copy_to_weights
(
scale
_data
,
scale_data
+
ic
,
shift_data
,
shift_data
+
ic
,
shift
->
data
<
T
>
()
+
ic
,
&
scaleshift_data
);
&
scaleshift_data
);
auto
scaleshift_memory
=
mkldnn
::
memory
{
// create mkldnn memory for input tensors (scale/shift)
batch_norm_bwd_pd
.
weights_primitive_desc
(),
scaleshift_data
.
data
()};
auto
scaleshift_memory
=
memory
(
batch_norm_bwd_pd
.
weights_primitive_desc
(),
scaleshift_data
.
data
());
// create mkldnn memory for output diff weights (combined scale/shift)
std
::
vector
<
T
>
diff_scaleshift_data
;
std
::
vector
<
T
>
diff_scaleshift_data
;
diff_scaleshift_data
.
reserve
(
scaleshift_size
);
diff_scaleshift_data
.
reserve
(
scaleshift_size
);
copy_to_weights
(
diff_scale
->
data
<
T
>
(),
diff_scale
->
data
<
T
>
()
+
ic
,
diff_shift
->
data
<
T
>
(),
diff_shift
->
data
<
T
>
()
+
ic
,
&
diff_scaleshift_data
);
auto
diff_scaleshift_memory
=
auto
diff_scaleshift_memory
=
mkldnn
::
memory
{
batch_norm_bwd_pd
.
diff_weights_primitive_desc
(),
memory
(
batch_norm_bwd_pd
.
diff_weights_primitive_desc
(),
diff_scaleshift_data
.
data
()};
diff_scaleshift_data
.
data
());
auto
diff_src
=
mkldnn
::
memory
{{
diff_src_md
,
mkldnn_engine
},
// here assume diff_src is in the same format of src
static_cast
<
void
*>
(
diff_x
->
data
<
T
>
())};
auto
diff_src_memory
=
memory
(
src_memory
.
get_primitive_desc
(),
diff_x_data
);
run_batch_norm_op
<
bn_bwd_types
::
op_type
>
(
// finally create batch_norm backward primitive
batch_norm_bwd_pd
,
src
,
mean
,
variance
,
diff_dst
,
scaleshift_memory
,
auto
batch_norm_bwd_prim
=
diff_src
,
diff_scaleshift_memory
);
batch_norm_bwd
(
batch_norm_bwd_pd
,
src_memory
,
mean_memory
,
variance_memory
,
diff_dst_memory
,
scaleshift_memory
,
diff_src_memory
,
diff_scaleshift_memory
);
// execute optional reorder and batch_norm backward primitive
std
::
vector
<
primitive
>
pipeline
;
if
(
is_diff_dst_reordered
)
pipeline
.
push_back
(
reorder_diff_dst
);
pipeline
.
push_back
(
batch_norm_bwd_prim
);
stream
(
stream
::
kind
::
eager
).
submit
(
pipeline
).
wait
();
// copy back diff sacle/shift to output tensors (diff scale/shift)
diff_scaleshift_data
.
resize
(
scaleshift_size
);
auto
it
=
std
::
begin
(
diff_scaleshift_data
);
auto
it
=
std
::
begin
(
diff_scaleshift_data
);
std
::
copy
(
it
,
std
::
next
(
it
,
ic
),
diff_scale
->
data
<
T
>
()
);
std
::
copy
(
it
,
std
::
next
(
it
,
ic
),
diff_scale
_data
);
std
::
copy
(
std
::
next
(
it
,
ic
),
std
::
end
(
diff_scaleshift_data
),
std
::
copy
(
std
::
next
(
it
,
ic
),
std
::
end
(
diff_scaleshift_data
),
diff_shift
->
data
<
T
>
());
diff_shift_data
);
// set layout/format of output tensors
diff_x
->
set_layout
(
DataLayout
::
kMKLDNN
);
diff_x
->
set_format
((
memory
::
format
)
diff_src_memory
.
get_primitive_desc
()
.
desc
()
.
data
.
format
);
}
}
};
};
}
// namespace operators
}
// namespace operators
}
// namespace paddle
}
// namespace paddle
namespace
ops
=
paddle
::
operators
;
namespace
ops
=
paddle
::
operators
;
REGISTER_OP_KERNEL
(
batch_norm
,
MKLDNN
,
paddle
::
platform
::
CPUPlace
,
REGISTER_OP_KERNEL
(
batch_norm
,
MKLDNN
,
::
paddle
::
platform
::
CPUPlace
,
ops
::
BatchNormMKLDNNOpKernel
<
float
>
);
ops
::
BatchNormMKLDNNOpKernel
<
float
>
);
REGISTER_OP_KERNEL
(
batch_norm_grad
,
MKLDNN
,
paddle
::
platform
::
CPUPlace
,
REGISTER_OP_KERNEL
(
batch_norm_grad
,
MKLDNN
,
::
paddle
::
platform
::
CPUPlace
,
ops
::
BatchNormMKLDNNGradOpKernel
<
float
>
);
ops
::
BatchNormMKLDNNGradOpKernel
<
float
>
);
paddle/fluid/operators/batch_norm_op.cc
浏览文件 @
489f581d
...
@@ -110,19 +110,19 @@ class BatchNormOp : public framework::OperatorWithKernel {
...
@@ -110,19 +110,19 @@ class BatchNormOp : public framework::OperatorWithKernel {
ctx
.
Input
<
Tensor
>
(
"Variance"
)
->
type
()),
ctx
.
Input
<
Tensor
>
(
"Variance"
)
->
type
()),
"Variance input should be of float type"
);
"Variance input should be of float type"
);
framework
::
LibraryType
library_
{
framework
::
LibraryType
::
kPlain
};
// TODO(pzelazko-intel): enable MKLDNN layout when it's ready
// TODO(pzelazko-intel): enable MKLDNN layout when it's ready
framework
::
LibraryType
library
=
framework
::
LibraryType
::
kPlain
;
framework
::
DataLayout
layout
=
framework
::
DataLayout
::
kAnyLayout
;
framework
::
DataLayout
layout
=
framework
::
DataLayout
::
kAnyLayout
;
#ifdef PADDLE_WITH_MKLDNN
#ifdef PADDLE_WITH_MKLDNN
if
(
library
_
==
framework
::
LibraryType
::
kPlain
&&
if
(
library
==
framework
::
LibraryType
::
kPlain
&&
platform
::
CanMKLDNNBeUsed
(
ctx
))
{
platform
::
CanMKLDNNBeUsed
(
ctx
))
{
library
_
=
framework
::
LibraryType
::
kMKLDNN
;
library
=
framework
::
LibraryType
::
kMKLDNN
;
layout
=
framework
::
DataLayout
::
kMKLDNN
;
layout
=
framework
::
DataLayout
::
kMKLDNN
;
}
}
#endif
#endif
return
framework
::
OpKernelType
(
input_data_type
,
ctx
.
GetPlace
(),
layout
,
return
framework
::
OpKernelType
(
input_data_type
,
ctx
.
GetPlace
(),
layout
,
library
_
);
library
);
}
}
};
};
...
@@ -370,19 +370,21 @@ class BatchNormGradOp : public framework::OperatorWithKernel {
...
@@ -370,19 +370,21 @@ class BatchNormGradOp : public framework::OperatorWithKernel {
PADDLE_THROW
(
"can't find Y@GRAD"
);
PADDLE_THROW
(
"can't find Y@GRAD"
);
}
}
framework
::
LibraryType
library_
{
framework
::
LibraryType
::
kPlain
};
// TODO(pzelazko-intel): enable MKLDNN layout when it's ready
// TODO(pzelazko-intel): enable MKLDNN layout when it's ready
framework
::
DataLayout
layout_
=
framework
::
DataLayout
::
kAnyLayout
;
framework
::
LibraryType
library
=
framework
::
LibraryType
::
kPlain
;
framework
::
DataLayout
layout
=
framework
::
DataLayout
::
kAnyLayout
;
#ifdef PADDLE_WITH_MKLDNN
#ifdef PADDLE_WITH_MKLDNN
if
(
library
_
==
framework
::
LibraryType
::
kPlain
&&
if
(
library
==
framework
::
LibraryType
::
kPlain
&&
platform
::
CanMKLDNNBeUsed
(
ctx
))
{
platform
::
CanMKLDNNBeUsed
(
ctx
))
{
library
_
=
framework
::
LibraryType
::
kMKLDNN
;
library
=
framework
::
LibraryType
::
kMKLDNN
;
layout
_
=
framework
::
DataLayout
::
kMKLDNN
;
layout
=
framework
::
DataLayout
::
kMKLDNN
;
}
}
#endif
#endif
return
framework
::
OpKernelType
(
return
framework
::
OpKernelType
(
framework
::
ToDataType
(
ctx
.
Input
<
Tensor
>
(
"X"
)
->
type
()),
ctx
.
GetPlace
(),
framework
::
ToDataType
(
ctx
.
Input
<
Tensor
>
(
"X"
)
->
type
()),
ctx
.
GetPlace
(),
layout
_
,
library_
);
layout
,
library
);
}
}
};
};
...
...
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