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
20659fc9
P
Paddle
项目概览
PaddlePaddle
/
Paddle
大约 2 年 前同步成功
通知
2325
Star
20933
Fork
5424
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
20659fc9
编写于
9月 03, 2018
作者:
T
tensor-tang
提交者:
GitHub
9月 03, 2018
浏览文件
操作
浏览文件
下载
差异文件
Merge pull request #13107 from tensor-tang/optimize/op/fusion_gru
Optimize fusion gru
上级
11bf6b26
c7adb99a
变更
3
显示空白变更内容
内联
并排
Showing
3 changed file
with
383 addition
and
168 deletion
+383
-168
paddle/fluid/operators/fusion_gru_op.cc
paddle/fluid/operators/fusion_gru_op.cc
+263
-163
paddle/fluid/operators/math/cpu_vec.h
paddle/fluid/operators/math/cpu_vec.h
+115
-0
paddle/fluid/operators/math/sequence2batch.h
paddle/fluid/operators/math/sequence2batch.h
+5
-5
未找到文件。
paddle/fluid/operators/fusion_gru_op.cc
浏览文件 @
20659fc9
...
@@ -13,16 +13,13 @@ See the License for the specific language governing permissions and
...
@@ -13,16 +13,13 @@ See the License for the specific language governing permissions and
limitations under the License. */
limitations under the License. */
#include "paddle/fluid/operators/fusion_gru_op.h"
#include "paddle/fluid/operators/fusion_gru_op.h"
#include <cstring> // for memcpy
#include <string>
#include <string>
#include "paddle/fluid/framework/eigen.h"
#include "paddle/fluid/operators/math/blas.h"
#include "paddle/fluid/operators/math/blas.h"
#include "paddle/fluid/operators/math/detail/activation_functions.h"
#include "paddle/fluid/operators/math/cpu_vec.h"
#include "paddle/fluid/operators/math/detail/gru_cpu_kernel.h"
#include "paddle/fluid/operators/math/detail/gru_kernel.h"
#include "paddle/fluid/operators/math/fc_compute.h"
#include "paddle/fluid/operators/math/fc_compute.h"
#include "paddle/fluid/operators/math/gru_compute.h"
#include "paddle/fluid/operators/math/math_function.h"
#include "paddle/fluid/operators/math/sequence2batch.h"
#include "paddle/fluid/operators/math/sequence2batch.h"
#include "paddle/fluid/platform/cpu_info.h"
namespace
paddle
{
namespace
paddle
{
namespace
operators
{
namespace
operators
{
...
@@ -35,12 +32,12 @@ void FusionGRUOp::InferShape(framework::InferShapeContext* ctx) const {
...
@@ -35,12 +32,12 @@ void FusionGRUOp::InferShape(framework::InferShapeContext* ctx) const {
"Input(WeightH) of GRU should not be null."
);
"Input(WeightH) of GRU should not be null."
);
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"XX"
),
"Output(XX) of GRU should not be null."
);
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"XX"
),
"Output(XX) of GRU should not be null."
);
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"
BatchedGate
"
),
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"
ReorderedH0
"
),
"Output(
BatchedGate
) of GRU should not be null."
);
"Output(
ReorderedH0
) of GRU should not be null."
);
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"Batch
ResetHiddenPrev
"
),
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"Batch
edInput
"
),
"Output(Batch
ResetHiddenPrev
) of GRU should not be null."
);
"Output(Batch
edInput
) of GRU should not be null."
);
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"Batched
Hidden
"
),
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"Batched
Out
"
),
"Output(Batched
Hidden
) of GRU should not be null."
);
"Output(Batched
Out
) of GRU should not be null."
);
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"Hidden"
),
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"Hidden"
),
"Output(Hidden) of GRU should not be null."
);
"Output(Hidden) of GRU should not be null."
);
...
@@ -83,12 +80,16 @@ void FusionGRUOp::InferShape(framework::InferShapeContext* ctx) const {
...
@@ -83,12 +80,16 @@ void FusionGRUOp::InferShape(framework::InferShapeContext* ctx) const {
}
}
framework
::
DDim
out_dims
({
x_dims
[
0
],
frame_size
});
framework
::
DDim
out_dims
({
x_dims
[
0
],
frame_size
});
ctx
->
SetOutputDim
(
"Hidden"
,
out_dims
);
ctx
->
SetOutputDim
(
"Hidden"
,
out_dims
);
ctx
->
SetOutputDim
(
"BatchedGate"
,
{
x_dims
[
0
],
wx_dims
[
1
]});
ctx
->
SetOutputDim
(
"BatchedInput"
,
{
x_dims
[
0
],
wx_dims
[
1
]});
ctx
->
SetOutputDim
(
"BatchedHidden"
,
out_dims
);
ctx
->
SetOutputDim
(
"BatchedOut"
,
out_dims
);
ctx
->
SetOutputDim
(
"BatchResetHiddenPrev"
,
out_dims
);
ctx
->
ShareLoD
(
"X"
,
"Hidden"
);
ctx
->
ShareLoD
(
"X"
,
"Hidden"
);
int
xx_width
=
x_dims
[
1
]
>
wx_dims
[
1
]
?
wx_dims
[
1
]
:
x_dims
[
1
];
int
xx_width
;
if
(
ctx
->
Attrs
().
Get
<
bool
>
(
"use_seq"
))
{
xx_width
=
wx_dims
[
1
];
}
else
{
xx_width
=
x_dims
[
1
]
>
wx_dims
[
1
]
?
wx_dims
[
1
]
:
x_dims
[
1
];
}
ctx
->
SetOutputDim
(
"XX"
,
{
x_dims
[
0
],
xx_width
});
ctx
->
SetOutputDim
(
"XX"
,
{
x_dims
[
0
],
xx_width
});
ctx
->
ShareLoD
(
"X"
,
"XX"
);
ctx
->
ShareLoD
(
"X"
,
"XX"
);
}
}
...
@@ -115,22 +116,29 @@ void FusionGRUOpMaker::Make() {
...
@@ -115,22 +116,29 @@ void FusionGRUOpMaker::Make() {
"(Tensor) The FC weight with shape (M x 3D),"
"(Tensor) The FC weight with shape (M x 3D),"
"where M is the dim size of x, D is the hidden size. "
);
"where M is the dim size of x, D is the hidden size. "
);
AddInput
(
"WeightH"
,
AddInput
(
"WeightH"
,
"(Tensor) (D x 3D) Same as GRUOp, where D is the hidden size. "
);
"(Tensor) (D x 3D) Same as GRUOp, where D is the hidden size. "
"This weight is not exactly D x 3D as: {W_update, W_reset, W_state}"
"Acutally they are D x 2D and D x D two part weights."
"{W_update, W_reset; W_state}"
"{D x (D + D); D x D}"
);
AddInput
(
"Bias"
,
AddInput
(
"Bias"
,
"(Tensor, optional) (1 x 3D)."
"(Tensor, optional) (1 x 3D)."
"Almost same as GRUOp."
"Almost same as GRUOp."
"Note: if have FC bias it should be added on this bias."
)
"Note: if have FC bias it should be added on this bias."
)
.
AsDispensable
();
.
AsDispensable
();
AddOutput
(
"ReorderedH0"
,
"(Tensor) (N x D), which N is the min-batch size."
)
.
AsIntermediate
();
AddOutput
(
"XX"
,
AddOutput
(
"XX"
,
"(LoDTensor) the result after X * WeightX (size is T x
4
D)"
"(LoDTensor) the result after X * WeightX (size is T x
3
D)"
" or batched_X (size is T x M), this will be automatically chosen,"
" or batched_X (size is T x M), this will be automatically chosen,"
" where T is the total time steps in this mini-batch,"
" where T is the total time steps in this mini-batch,"
" D is the hidden size, M is the dim size of x input."
)
" D is the hidden size, M is the dim size of x input."
)
.
AsIntermediate
();
.
AsIntermediate
();
AddOutput
(
"BatchedGate"
,
"(LoDTensor) Same as GRUOp"
).
AsIntermediate
();
AddOutput
(
"BatchedInput"
,
AddOutput
(
"BatchResetHiddenPrev"
,
"(LoDTensor) (T x 3D) Same as GRUOp."
)
"(LoDTensor) This is the batched result of input X"
"or the batched result after fc, shape (T x 3D)"
)
.
AsIntermediate
();
.
AsIntermediate
();
AddOutput
(
"Batched
Hidden"
,
"(LoDTensor) (T X D) Same as GRUOp
."
)
AddOutput
(
"Batched
Out"
,
"(LoDTensor) (T X D) save batched hidden
."
)
.
AsIntermediate
();
.
AsIntermediate
();
AddOutput
(
"Hidden"
,
"(LoDTensor) (T x D) Same as GRUOp"
);
AddOutput
(
"Hidden"
,
"(LoDTensor) (T x D) Same as GRUOp"
);
AddAttr
<
std
::
string
>
(
"activation"
,
AddAttr
<
std
::
string
>
(
"activation"
,
...
@@ -146,6 +154,10 @@ void FusionGRUOpMaker::Make() {
...
@@ -146,6 +154,10 @@ void FusionGRUOpMaker::Make() {
"(bool, defalut: False) "
"(bool, defalut: False) "
"whether to compute reversed GRU."
)
"whether to compute reversed GRU."
)
.
SetDefault
(
false
);
.
SetDefault
(
false
);
AddAttr
<
bool
>
(
"use_seq"
,
"(bool, defalut: True) "
"whether to use seq mode to compute GRU."
)
.
SetDefault
(
true
);
AddComment
(
R"DOC(
AddComment
(
R"DOC(
The Fusion complete GRU Operator.
The Fusion complete GRU Operator.
This operator fuse the fully-connected operator into GRU,
This operator fuse the fully-connected operator into GRU,
...
@@ -153,172 +165,261 @@ more details can refer to GRU op.
...
@@ -153,172 +165,261 @@ more details can refer to GRU op.
)DOC"
);
)DOC"
);
}
}
template
<
typename
DeviceContext
,
typename
T
>
template
<
typename
T
>
inline
void
ReorderInitState
(
const
DeviceContext
&
ctx
,
const
framework
::
Tensor
&
src
,
framework
::
Vector
<
size_t
>
index_lod
,
framework
::
Tensor
*
dst
,
bool
indexed_src
)
{
math
::
CopyMatrixRowsFunctor
<
DeviceContext
,
T
>
row_shuffle
;
dst
->
mutable_data
<
T
>
(
src
.
dims
(),
ctx
.
GetPlace
());
row_shuffle
(
ctx
,
src
,
index_lod
,
dst
,
indexed_src
);
}
template
<
typename
DeviceContext
,
typename
T
>
class
FusionGRUKernel
:
public
framework
::
OpKernel
<
T
>
{
class
FusionGRUKernel
:
public
framework
::
OpKernel
<
T
>
{
public:
public:
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
auto
*
x
=
ctx
.
Input
<
LoDTensor
>
(
"X"
);
if
(
ctx
.
Attr
<
bool
>
(
"use_seq"
))
{
auto
*
wx
=
ctx
.
Input
<
Tensor
>
(
"WeightX"
);
SeqCompute
(
ctx
);
auto
*
wh
=
ctx
.
Input
<
Tensor
>
(
"WeightH"
);
}
else
{
auto
*
bias
=
ctx
.
Input
<
Tensor
>
(
"Bias"
);
BatchCompute
(
ctx
);
auto
*
h0
=
ctx
.
Input
<
Tensor
>
(
"H0"
);
}
}
auto
*
xx
=
ctx
.
Output
<
LoDTensor
>
(
"XX"
);
auto
*
batched_gate
=
ctx
.
Output
<
LoDTensor
>
(
"BatchedGate"
);
#define INIT_VEC_FUNC \
auto
*
batch_reset_hidden_prev
=
std::function<void(const int, const T *, T *)> act_gate, act_state; \
ctx
.
Output
<
LoDTensor
>
(
"BatchResetHiddenPrev"
);
std::function<void(const int, const T*, const T*, const T*, T*)> cross; \
auto
*
batch_hidden
=
ctx
.
Output
<
LoDTensor
>
(
"BatchedHidden"
);
auto& act_gate_str = ctx.Attr<std::string>("gate_activation"); \
auto
*
hidden_out
=
ctx
.
Output
<
LoDTensor
>
(
"Hidden"
);
auto& act_state_str = ctx.Attr<std::string>("activation"); \
if (platform::jit::MayIUse(platform::jit::avx)) { \
math::VecActivations<T, platform::jit::avx> act_functor; \
act_gate = act_functor(act_gate_str); \
act_state = act_functor(act_state_str); \
cross = math::vec_cross<T, platform::jit::avx>; \
} else { \
math::VecActivations<T, platform::jit::isa_any> act_functor; \
act_gate = act_functor(act_gate_str); \
act_state = act_functor(act_state_str); \
cross = math::vec_cross<T, platform::jit::isa_any>; \
}
#define INIT_BASE_INPUT_OUTPUT \
auto* h0 = ctx.Input<Tensor>("H0"); \
auto* wx = ctx.Input<Tensor>("WeightX"); \
auto* wh = ctx.Input<Tensor>("WeightH"); \
auto* bias = ctx.Input<Tensor>("Bias"); \
auto* xx = ctx.Output<LoDTensor>("XX"); \
auto* hidden_out = ctx.Output<LoDTensor>("Hidden"); \
bool is_reverse = ctx.Attr<bool>("is_reverse");
bool is_reverse = ctx.Attr<bool>("is_reverse");
#define INIT_BASE_SIZES \
auto x_dims = x->dims();
/* T x M*/
\
auto wh_dims = wh->dims();
/* D x 3D*/
\
const int total_T = x_dims[0]; \
const int M = x_dims[1]; \
const int D = wh_dims[0]; \
const int D3 = wh_dims[1]; \
const int D2 = D * 2;
void
SeqCompute
(
const
framework
::
ExecutionContext
&
ctx
)
const
{
using
DeviceContext
=
paddle
::
platform
::
CPUDeviceContext
;
auto
*
x
=
ctx
.
Input
<
LoDTensor
>
(
"X"
);
INIT_BASE_INPUT_OUTPUT
INIT_BASE_SIZES
INIT_VEC_FUNC
auto
x_lod
=
x
->
lod
();
const
int
N
=
x_lod
[
0
].
size
()
-
1
;
const
T
*
x_data
=
x
->
data
<
T
>
();
const
T
*
h0_data
=
h0
?
h0
->
data
<
T
>
()
:
nullptr
;
const
T
*
wx_data
=
wx
->
data
<
T
>
();
const
T
*
wh_data
=
wh
->
data
<
T
>
();
const
T
*
wh_state_data
=
wh_data
+
D
*
D2
;
T
*
xx_data
=
xx
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
T
*
xx_data
=
xx
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
T
*
batched_gate_data
=
batched_gate
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
T
*
hidden_out_data
=
hidden_out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
batch_reset_hidden_prev
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
batch_hidden
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
auto
blas
=
math
::
GetBlas
<
DeviceContext
,
T
>
(
ctx
);
hidden_out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
math
::
FCCompute
<
DeviceContext
,
T
>
(
blas
,
total_T
,
D3
,
M
,
x_data
,
wx_data
,
xx_data
,
bias
?
bias
->
data
<
T
>
()
:
nullptr
);
int
xx_offset
=
D3
;
int
gate_offset
=
D
;
if
(
is_reverse
)
{
const
int
offset
=
(
total_T
-
1
)
*
D
;
xx_data
=
xx_data
+
offset
*
3
;
hidden_out_data
=
hidden_out_data
+
offset
;
xx_offset
=
-
D3
;
gate_offset
=
-
D
;
}
auto
move_step
=
[
&
]()
{
xx_data
=
xx_data
+
xx_offset
;
hidden_out_data
=
hidden_out_data
+
gate_offset
;
};
for
(
int
i
=
0
;
i
<
N
;
++
i
)
{
int
bid
=
is_reverse
?
N
-
1
-
i
:
i
;
int
seq_len
=
x_lod
[
0
][
bid
+
1
]
-
x_lod
[
0
][
bid
];
const
T
*
prev_hidden_data
=
nullptr
;
int
tstart
=
0
;
if
(
h0_data
)
{
prev_hidden_data
=
h0_data
+
bid
*
D
;
}
else
{
// W: {W_update, W_reset; W_state}
// update gate
act_gate
(
D
,
xx_data
,
xx_data
);
// state gate
act_state
(
D
,
xx_data
+
D2
,
xx_data
+
D2
);
// out = a*b
blas
.
VMUL
(
D
,
xx_data
,
xx_data
+
D2
,
hidden_out_data
);
// save prev
prev_hidden_data
=
hidden_out_data
;
tstart
=
1
;
move_step
();
}
for
(
int
step
=
tstart
;
step
<
seq_len
;
++
step
)
{
// gemm prev * (Wu + Wr)
blas
.
GEMM
(
CblasNoTrans
,
CblasNoTrans
,
1
,
D2
,
D
,
static_cast
<
T
>
(
1
),
prev_hidden_data
,
D
,
wh_data
,
D2
,
static_cast
<
T
>
(
1
),
xx_data
,
D3
);
act_gate
(
D2
,
xx_data
,
xx_data
);
// rt = rt*ht_1 inplace result
blas
.
VMUL
(
D
,
prev_hidden_data
,
xx_data
+
D
,
hidden_out_data
);
// gemm rt * Ws
blas
.
GEMM
(
CblasNoTrans
,
CblasNoTrans
,
1
,
D
,
D
,
static_cast
<
T
>
(
1
),
hidden_out_data
,
D
,
wh_state_data
,
D
,
static_cast
<
T
>
(
1
),
xx_data
+
D2
,
D3
);
act_state
(
D
,
xx_data
+
D2
,
xx_data
+
D2
);
// out = zt*ht~ + (1-zt)*ht_1
cross
(
D
,
xx_data
,
xx_data
+
D2
,
prev_hidden_data
,
hidden_out_data
);
// save prev
prev_hidden_data
=
hidden_out_data
;
move_step
();
}
}
}
void
BatchCompute
(
const
framework
::
ExecutionContext
&
ctx
)
const
{
using
DeviceContext
=
paddle
::
platform
::
CPUDeviceContext
;
auto
*
x
=
ctx
.
Input
<
LoDTensor
>
(
"X"
);
if
(
x
->
lod
()[
0
].
size
()
==
2
)
{
SeqCompute
(
ctx
);
return
;
}
INIT_BASE_INPUT_OUTPUT
INIT_BASE_SIZES
INIT_VEC_FUNC
auto
*
reordered_h0
=
ctx
.
Output
<
Tensor
>
(
"ReorderedH0"
);
auto
*
batched_input
=
ctx
.
Output
<
LoDTensor
>
(
"BatchedInput"
);
auto
*
batched_out
=
ctx
.
Output
<
LoDTensor
>
(
"BatchedOut"
);
const
T
*
x_data
=
x
->
data
<
T
>
();
const
T
*
x_data
=
x
->
data
<
T
>
();
const
T
*
wx_data
=
wx
->
data
<
T
>
();
const
T
*
wx_data
=
wx
->
data
<
T
>
();
const
T
*
wh_data
=
wh
->
data
<
T
>
();
const
T
*
wh_data
=
wh
->
data
<
T
>
();
auto
x_dims
=
x
->
dims
();
T
*
xx_data
=
xx
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
auto
wx_dims
=
wx
->
dims
();
T
*
batched_input_data
=
batched_input
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
T
*
batched_out_data
=
batched_out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
hidden_out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
auto
&
dev_ctx
=
ctx
.
template
device_context
<
DeviceContext
>();
auto
&
dev_ctx
=
ctx
.
template
device_context
<
DeviceContext
>();
auto
blas
=
math
::
GetBlas
<
DeviceContext
,
T
>
(
dev_ctx
);
auto
blas
=
math
::
GetBlas
<
DeviceContext
,
T
>
(
dev_ctx
);
math
::
LoDTensor2BatchFunctor
<
DeviceContext
,
T
>
to_batch
;
math
::
LoDTensor2BatchFunctor
<
DeviceContext
,
T
>
to_batch
;
if
(
x_dims
[
1
]
>
wx_dims
[
1
]
)
{
if
(
M
>
D3
)
{
math
::
FCCompute
<
DeviceContext
,
T
>
(
blas
,
x_dims
[
0
],
wx_dims
[
1
],
x_dims
[
1
]
,
math
::
FCCompute
<
DeviceContext
,
T
>
(
blas
,
total_T
,
D3
,
M
,
x_data
,
wx_data
,
x
_data
,
wx_data
,
x
x_data
,
xx_data
,
bias
?
bias
->
data
<
T
>
()
:
NULL
);
bias
?
bias
->
data
<
T
>
()
:
nullptr
);
to_batch
(
dev_ctx
,
*
xx
,
batched_
gate
,
true
,
is_reverse
);
to_batch
(
dev_ctx
,
*
xx
,
batched_
input
,
true
,
is_reverse
);
}
else
{
}
else
{
to_batch
(
dev_ctx
,
*
x
,
xx
,
true
,
is_reverse
);
to_batch
(
dev_ctx
,
*
x
,
xx
,
true
,
is_reverse
);
batched_
gate
->
set_lod
(
xx
->
lod
());
batched_
input
->
set_lod
(
xx
->
lod
());
math
::
FCCompute
<
DeviceContext
,
T
>
(
blas
,
x_dims
[
0
],
wx_dims
[
1
],
x_dims
[
1
]
,
math
::
FCCompute
<
DeviceContext
,
T
>
(
blas
,
total_T
,
D3
,
M
,
xx_data
,
wx_data
,
xx_data
,
wx_data
,
batched_gate
_data
,
batched_input
_data
,
bias
?
bias
->
data
<
T
>
()
:
NULL
);
bias
?
bias
->
data
<
T
>
()
:
nullptr
);
}
}
int
frame_size
=
static_cast
<
int
>
(
wx_dims
[
1
]
/
3
);
auto
batched_lod
=
batched_input
->
lod
();
math
::
GRUMetaValue
<
T
>
gru_value
;
const
auto
&
seq_order
=
batched_lod
[
2
];
gru_value
.
gate_weight
=
const_cast
<
T
*>
(
wh_data
);
const
int
max_bs
=
seq_order
.
size
();
gru_value
.
state_weight
=
reordered_h0
->
Resize
({
max_bs
,
D
});
const_cast
<
T
*>
(
wh_data
+
2
*
frame_size
*
frame_size
);
Tensor
ordered_h0
;
framework
::
Vector
<
size_t
>
order
(
batched_gate
->
lod
()[
2
]);
int
tstart
=
0
;
T
*
prev_hidden_data
=
nullptr
;
if
(
h0
)
{
if
(
h0
)
{
ReorderInitState
<
DeviceContext
,
T
>
(
// reorder h0
ctx
.
template
device_context
<
DeviceContext
>(),
*
h0
,
order
,
&
ordered_h0
,
T
*
reordered_h0_data
=
reordered_h0
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
true
);
const
T
*
h0_data
=
h0
->
data
<
T
>
();
gru_value
.
prev_out_value
=
ordered_h0
.
data
<
T
>
();
prev_hidden_data
=
reordered_h0_data
;
}
else
{
size_t
sz
=
sizeof
(
T
)
*
D
;
gru_value
.
prev_out_value
=
nullptr
;
for
(
int
i
=
0
;
i
<
max_bs
;
++
i
)
{
std
::
memcpy
(
reordered_h0_data
,
h0_data
+
seq_order
[
i
]
*
D
,
sz
);
reordered_h0_data
+=
D
;
}
}
auto
batch_starts
=
batched_gate
->
lod
()[
0
];
}
else
{
size_t
seq_len
=
batch_starts
.
size
()
-
1
;
// compute without h0
auto
active_node
=
T
*
cur_in_data
=
batched_input_data
;
math
::
detail
::
GetActivationType
(
ctx
.
Attr
<
std
::
string
>
(
"activation"
));
T
*
cur_out_data
=
batched_out_data
;
auto
active_gate
=
math
::
detail
::
GetActivationType
(
// W: {W_update, W_reset; W_state}
ctx
.
Attr
<
std
::
string
>
(
"gate_activation"
));
for
(
int
i
=
0
;
i
<
max_bs
;
++
i
)
{
// update gate
#ifdef PADDLE_WITH_MKLML
act_gate
(
D
,
cur_in_data
,
cur_in_data
);
// use MKL packed to speedup GEMM
// state gate
if
(
FLAGS_paddle_num_threads
>=
4
)
{
act_state
(
D
,
cur_in_data
+
D2
,
cur_in_data
+
D2
);
auto
blas
=
math
::
GetBlas
<
DeviceContext
,
T
>
(
dev_ctx
);
// out = a*b
T
*
packed_gate
=
blas
.
GEMM_ALLOC
(
CblasBMatrix
,
1
/*height of C*/
,
blas
.
VMUL
(
D
,
cur_in_data
,
cur_in_data
+
D2
,
cur_out_data
);
frame_size
*
2
/*width of weight*/
,
// add offset
frame_size
/*height of height*/
);
cur_in_data
+=
D3
;
PADDLE_ENFORCE
(
packed_gate
);
cur_out_data
+=
D
;
blas
.
GEMM_PACK
(
CblasBMatrix
,
CblasNoTrans
,
1
/*cur bs?*/
,
frame_size
*
2
,
frame_size
,
T
(
1.0
),
gru_value
.
gate_weight
,
frame_size
*
2
,
packed_gate
);
T
*
packed_state
=
blas
.
GEMM_ALLOC
(
CblasBMatrix
,
1
/*height of C*/
,
frame_size
/*width of weight*/
,
frame_size
/*height of height*/
);
PADDLE_ENFORCE
(
packed_state
);
blas
.
GEMM_PACK
(
CblasBMatrix
,
CblasNoTrans
,
1
/*cur bs?*/
,
frame_size
,
frame_size
,
T
(
1.0
),
gru_value
.
state_weight
,
frame_size
,
packed_state
);
for
(
size_t
n
=
0
;
n
<
seq_len
;
n
++
)
{
int
bstart
=
static_cast
<
int
>
(
batch_starts
[
n
]);
int
bend
=
static_cast
<
int
>
(
batch_starts
[
n
+
1
]);
int
cur_batch_size
=
bend
-
bstart
;
Tensor
gate_t
=
batched_gate
->
Slice
(
bstart
,
bend
);
Tensor
reset_hidden_prev_t
=
batch_reset_hidden_prev
->
Slice
(
bstart
,
bend
);
Tensor
hidden_t
=
batch_hidden
->
Slice
(
bstart
,
bend
);
gru_value
.
output_value
=
hidden_t
.
data
<
T
>
();
gru_value
.
gate_value
=
gate_t
.
data
<
T
>
();
gru_value
.
reset_output_value
=
reset_hidden_prev_t
.
data
<
T
>
();
if
(
gru_value
.
prev_out_value
)
{
blas
.
GEMM_COMPUTE
(
CblasNoTrans
,
CblasPacked
,
cur_batch_size
,
frame_size
*
2
,
frame_size
,
gru_value
.
prev_out_value
,
frame_size
,
packed_gate
,
frame_size
*
2
,
T
(
1
),
gru_value
.
gate_value
,
frame_size
*
3
);
}
}
tstart
=
1
;
math
::
detail
::
forward_reset_output
(
prev_hidden_data
=
batched_out_data
;
math
::
detail
::
forward
::
gru_resetOutput
<
T
>
(),
gru_value
,
frame_size
,
cur_batch_size
,
active_gate
);
if
(
gru_value
.
prev_out_value
)
{
blas
.
GEMM_COMPUTE
(
CblasNoTrans
,
CblasPacked
,
cur_batch_size
,
frame_size
,
frame_size
,
gru_value
.
reset_output_value
,
frame_size
,
packed_state
,
frame_size
,
T
(
1
),
gru_value
.
gate_value
+
frame_size
*
2
,
frame_size
*
3
);
}
}
// Then start from next
math
::
detail
::
forward_final_output
(
const
T
*
wh_state_data
=
wh_data
+
D
*
D2
;
math
::
detail
::
forward
::
gru_finalOutput
<
T
>
(),
gru_value
,
frame_size
,
const
auto
&
batch_starts
=
batched_lod
[
0
];
cur_batch_size
,
active_node
);
const
int
max_seq_len
=
batch_starts
.
size
()
-
1
;
batched_input_data
=
batched_input_data
+
tstart
*
max_bs
*
D3
;
gru_value
.
prev_out_value
=
gru_value
.
output_value
;
batched_out_data
=
batched_out_data
+
tstart
*
max_bs
*
D
;
for
(
int
step
=
tstart
;
step
<
max_seq_len
;
++
step
)
{
const
int
cur_bs
=
batch_starts
[
step
+
1
]
-
batch_starts
[
step
];
// gemm prev * (Wu + Wr)
blas
.
GEMM
(
CblasNoTrans
,
CblasNoTrans
,
cur_bs
,
D2
,
D
,
static_cast
<
T
>
(
1
),
prev_hidden_data
,
D
,
wh_data
,
D2
,
static_cast
<
T
>
(
1
),
batched_input_data
,
D3
);
T
*
cur_batched_data
=
batched_input_data
;
T
*
cur_out_data
=
batched_out_data
;
T
*
cur_prev_hidden_data
=
prev_hidden_data
;
for
(
int
i
=
0
;
i
<
cur_bs
;
++
i
)
{
act_gate
(
D2
,
cur_batched_data
,
cur_batched_data
);
// rt = rt*ht_1 inplace result
blas
.
VMUL
(
D
,
cur_prev_hidden_data
,
cur_batched_data
+
D
,
cur_out_data
);
cur_batched_data
+=
D3
;
cur_prev_hidden_data
+=
D
;
cur_out_data
+=
D
;
}
}
blas
.
GEMM_FREE
(
packed_gate
);
cur_batched_data
=
batched_input_data
;
blas
.
GEMM_FREE
(
packed_state
);
cur_out_data
=
batched_out_data
;
}
else
{
blas
.
GEMM
(
CblasNoTrans
,
CblasNoTrans
,
cur_bs
,
D
,
D
,
static_cast
<
T
>
(
1
),
#endif
cur_out_data
,
D
,
wh_state_data
,
D
,
static_cast
<
T
>
(
1
),
for
(
size_t
n
=
0
;
n
<
seq_len
;
n
++
)
{
cur_batched_data
+
D2
,
D3
);
int
bstart
=
static_cast
<
int
>
(
batch_starts
[
n
]);
int
bend
=
static_cast
<
int
>
(
batch_starts
[
n
+
1
]);
cur_prev_hidden_data
=
prev_hidden_data
;
int
cur_batch_size
=
bend
-
bstart
;
for
(
int
i
=
0
;
i
<
cur_bs
;
++
i
)
{
// ht~ = act_state(...)
Tensor
gate_t
=
batched_gate
->
Slice
(
bstart
,
bend
);
act_state
(
D
,
cur_batched_data
+
D2
,
cur_batched_data
+
D2
);
Tensor
reset_hidden_prev_t
=
// out = zt*ht~ + (1-zt)*ht_1
batch_reset_hidden_prev
->
Slice
(
bstart
,
bend
);
cross
(
D
,
cur_batched_data
,
cur_batched_data
+
D2
,
cur_prev_hidden_data
,
Tensor
hidden_t
=
batch_hidden
->
Slice
(
bstart
,
bend
);
cur_out_data
);
gru_value
.
output_value
=
hidden_t
.
data
<
T
>
();
gru_value
.
gate_value
=
gate_t
.
data
<
T
>
();
cur_batched_data
+=
D3
;
gru_value
.
reset_output_value
=
reset_hidden_prev_t
.
data
<
T
>
();
cur_prev_hidden_data
+=
D
;
cur_out_data
+=
D
;
math
::
GRUUnitFunctor
<
DeviceContext
,
T
>::
compute
(
dev_ctx
,
gru_value
,
frame_size
,
cur_batch_size
,
active_node
,
active_gate
);
gru_value
.
prev_out_value
=
gru_value
.
output_value
;
}
}
#ifdef PADDLE_WITH_MKLML
prev_hidden_data
=
batched_out_data
;
batched_out_data
=
cur_out_data
;
batched_input_data
=
cur_batched_data
;
}
}
#endif
math
::
Batch2LoDTensorFunctor
<
DeviceContext
,
T
>
to_seq
;
math
::
Batch2LoDTensorFunctor
<
DeviceContext
,
T
>
to_seq
;
batch
_hidden
->
set_lod
(
batched_gate
->
lod
()
);
batch
ed_out
->
set_lod
(
batched_lod
);
to_seq
(
dev_ctx
,
*
batch
_hidden
,
hidden_out
);
to_seq
(
dev_ctx
,
*
batch
ed_out
,
hidden_out
);
}
}
#undef INIT_VEC_FUNC
#undef INIT_BASE_SIZES
#undef INIT_BASE_INPUT_OUTPUT
};
};
}
// namespace operators
}
// namespace operators
...
@@ -327,6 +428,5 @@ class FusionGRUKernel : public framework::OpKernel<T> {
...
@@ -327,6 +428,5 @@ class FusionGRUKernel : public framework::OpKernel<T> {
namespace
ops
=
paddle
::
operators
;
namespace
ops
=
paddle
::
operators
;
REGISTER_OPERATOR
(
fusion_gru
,
ops
::
FusionGRUOp
,
ops
::
FusionGRUOpMaker
,
REGISTER_OPERATOR
(
fusion_gru
,
ops
::
FusionGRUOp
,
ops
::
FusionGRUOpMaker
,
paddle
::
framework
::
DefaultGradOpDescMaker
<
true
>
);
paddle
::
framework
::
DefaultGradOpDescMaker
<
true
>
);
REGISTER_OP_CPU_KERNEL
(
REGISTER_OP_CPU_KERNEL
(
fusion_gru
,
ops
::
FusionGRUKernel
<
float
>
,
fusion_gru
,
ops
::
FusionGRUKernel
<
paddle
::
platform
::
CPUDeviceContext
,
float
>
,
ops
::
FusionGRUKernel
<
double
>
);
ops
::
FusionGRUKernel
<
paddle
::
platform
::
CPUDeviceContext
,
double
>
);
paddle/fluid/operators/math/cpu_vec.h
浏览文件 @
20659fc9
...
@@ -132,6 +132,121 @@ inline void vec_scal<float, platform::jit::avx512_common>(const int n,
...
@@ -132,6 +132,121 @@ inline void vec_scal<float, platform::jit::avx512_common>(const int n,
vec_scal
<
float
,
platform
::
jit
::
avx2
>
(
n
,
a
,
x
,
y
);
vec_scal
<
float
,
platform
::
jit
::
avx2
>
(
n
,
a
,
x
,
y
);
}
}
template
<
typename
T
,
platform
::
jit
::
cpu_isa_t
isa
=
platform
::
jit
::
isa_any
>
inline
void
vec_bias_sub
(
const
int
n
,
const
T
a
,
const
T
*
x
,
T
*
y
)
{
for
(
int
i
=
0
;
i
<
n
;
++
i
)
{
y
[
i
]
=
a
-
x
[
i
];
}
}
template
<
>
inline
void
vec_bias_sub
<
float
,
platform
::
jit
::
avx
>
(
const
int
n
,
const
float
a
,
const
float
*
x
,
float
*
y
)
{
#ifdef __AVX__
constexpr
int
block
=
AVX_FLOAT_BLOCK
;
if
(
n
<
block
)
{
vec_bias_sub
<
float
,
platform
::
jit
::
isa_any
>
(
n
,
a
,
x
,
y
);
return
;
}
const
int
rest
=
n
%
block
;
const
int
end
=
n
-
rest
;
int
i
=
0
;
__m256
bias
=
_mm256_set1_ps
(
a
);
__m256
tmp
;
#define MOVE_ONE_STEP \
tmp = _mm256_loadu_ps(x + i); \
tmp = _mm256_sub_ps(bias, tmp); \
_mm256_storeu_ps(y + i, tmp)
for
(
i
=
0
;
i
<
end
;
i
+=
block
)
{
MOVE_ONE_STEP
;
}
#undef MOVE_ONE_STEP
if
(
rest
==
0
)
{
return
;
}
// can not continue move step if src and dst are inplace
for
(
i
=
n
-
rest
;
i
<
n
;
++
i
)
{
y
[
i
]
=
a
-
x
[
i
];
}
#else
vec_bias_sub
<
float
,
platform
::
jit
::
isa_any
>
(
n
,
a
,
x
,
y
);
#endif
}
template
<
>
inline
void
vec_bias_sub
<
float
,
platform
::
jit
::
avx2
>
(
const
int
n
,
const
float
a
,
const
float
*
x
,
float
*
y
)
{
vec_bias_sub
<
float
,
platform
::
jit
::
avx
>
(
n
,
a
,
x
,
y
);
}
template
<
>
inline
void
vec_bias_sub
<
float
,
platform
::
jit
::
avx512_common
>
(
const
int
n
,
const
float
a
,
const
float
*
x
,
float
*
y
)
{
// TODO(TJ): enable me
vec_bias_sub
<
float
,
platform
::
jit
::
avx2
>
(
n
,
a
,
x
,
y
);
}
// out = x*y + (1-x)*z
template
<
typename
T
,
platform
::
jit
::
cpu_isa_t
isa
=
platform
::
jit
::
isa_any
>
inline
void
vec_cross
(
const
int
n
,
const
T
*
x
,
const
T
*
y
,
const
T
*
z
,
T
*
out
)
{
for
(
int
i
=
0
;
i
<
n
;
++
i
)
{
out
[
i
]
=
x
[
i
]
*
y
[
i
]
+
(
static_cast
<
T
>
(
1
)
-
x
[
i
])
*
z
[
i
];
}
}
template
<
>
inline
void
vec_cross
<
float
,
platform
::
jit
::
avx
>
(
const
int
n
,
const
float
*
x
,
const
float
*
y
,
const
float
*
z
,
float
*
out
)
{
#ifdef __AVX__
constexpr
int
block
=
AVX_FLOAT_BLOCK
;
if
(
n
<
block
)
{
vec_cross
<
float
,
platform
::
jit
::
isa_any
>
(
n
,
x
,
y
,
z
,
out
);
return
;
}
const
int
rest
=
n
%
block
;
const
int
end
=
n
-
rest
;
int
i
=
0
;
__m256
bias
=
_mm256_set1_ps
(
1.
f
);
__m256
tmpx
,
tmpy
,
tmpz
;
for
(
i
=
0
;
i
<
end
;
i
+=
block
)
{
tmpx
=
_mm256_loadu_ps
(
x
+
i
);
tmpy
=
_mm256_loadu_ps
(
y
+
i
);
tmpz
=
_mm256_loadu_ps
(
z
+
i
);
tmpy
=
_mm256_mul_ps
(
tmpx
,
tmpy
);
tmpx
=
_mm256_sub_ps
(
bias
,
tmpx
);
tmpz
=
_mm256_mul_ps
(
tmpx
,
tmpz
);
tmpz
=
_mm256_add_ps
(
tmpy
,
tmpz
);
_mm256_storeu_ps
(
out
+
i
,
tmpz
);
}
if
(
rest
==
0
)
{
return
;
}
// can not continue move step if src and dst are inplace
for
(
i
=
n
-
rest
;
i
<
n
;
++
i
)
{
out
[
i
]
=
x
[
i
]
*
y
[
i
]
+
(
1.
f
-
x
[
i
])
*
z
[
i
];
}
#else
vec_cross
<
float
,
platform
::
jit
::
isa_any
>
(
n
,
x
,
y
,
z
,
out
);
#endif
}
template
<
>
inline
void
vec_cross
<
float
,
platform
::
jit
::
avx2
>
(
const
int
n
,
const
float
*
x
,
const
float
*
y
,
const
float
*
z
,
float
*
out
)
{
vec_cross
<
float
,
platform
::
jit
::
avx
>
(
n
,
x
,
y
,
z
,
out
);
}
template
<
>
inline
void
vec_cross
<
float
,
platform
::
jit
::
avx512_common
>
(
const
int
n
,
const
float
*
x
,
const
float
*
y
,
const
float
*
z
,
float
*
out
)
{
// TODO(TJ): enable me
vec_cross
<
float
,
platform
::
jit
::
avx
>
(
n
,
x
,
y
,
z
,
out
);
}
template
<
typename
T
,
platform
::
jit
::
cpu_isa_t
isa
=
platform
::
jit
::
isa_any
>
template
<
typename
T
,
platform
::
jit
::
cpu_isa_t
isa
=
platform
::
jit
::
isa_any
>
inline
void
vec_add_bias
(
const
int
n
,
const
T
a
,
const
T
*
x
,
T
*
y
)
{
inline
void
vec_add_bias
(
const
int
n
,
const
T
a
,
const
T
*
x
,
T
*
y
)
{
for
(
int
i
=
0
;
i
<
n
;
++
i
)
{
for
(
int
i
=
0
;
i
<
n
;
++
i
)
{
...
...
paddle/fluid/operators/math/sequence2batch.h
浏览文件 @
20659fc9
...
@@ -92,7 +92,7 @@ class LoDTensor2BatchFunctor {
...
@@ -92,7 +92,7 @@ class LoDTensor2BatchFunctor {
// Calculate the start position of each batch.
// Calculate the start position of each batch.
// example: sequences = {s0, s1, s2}
// example: sequences = {s0, s1, s2}
// s0: 0 0 0 0, s1: 1 1 1 1 1, s2: 2 2 2
// s0: 0 0 0 0, s1: 1 1 1 1 1, s2: 2 2 2
//
num_batch
= 5,
//
max_seqlen
= 5,
// batchIndex = {b0, b1, b2, b3, b4}
// batchIndex = {b0, b1, b2, b3, b4}
// b0: 1 0 2, b1: 1 0 2, b2: 1 0 2, b3: 1 0, b4: 1
// b0: 1 0 2, b1: 1 0 2, b2: 1 0 2, b3: 1 0, b4: 1
// batch_start_positions[6] = {0, 3, 6, 9, 11, 12}
// batch_start_positions[6] = {0, 3, 6, 9, 11, 12}
...
@@ -109,7 +109,7 @@ class LoDTensor2BatchFunctor {
...
@@ -109,7 +109,7 @@ class LoDTensor2BatchFunctor {
// where 1 is the second sequence,
// where 1 is the second sequence,
// 0 is the first sequence,
// 0 is the first sequence,
// 2 is the third sequence.
// 2 is the third sequence.
// The
num_batch
represents batch size after rearranging the
// The
max_seqlen
represents batch size after rearranging the
// input LodTensor. It is also the maximum length of input sequence.
// input LodTensor. It is also the maximum length of input sequence.
paddle
::
framework
::
LoD
batch_lods
;
paddle
::
framework
::
LoD
batch_lods
;
...
@@ -118,8 +118,8 @@ class LoDTensor2BatchFunctor {
...
@@ -118,8 +118,8 @@ class LoDTensor2BatchFunctor {
batch_lods
.
emplace_back
(
std
::
vector
<
size_t
>
{
0
});
batch_lods
.
emplace_back
(
std
::
vector
<
size_t
>
{
0
});
// batch_lods[0] is the start positions for batch LoDTensor
// batch_lods[0] is the start positions for batch LoDTensor
int
num_batch
=
seq_info
[
0
].
length
;
int
max_seqlen
=
seq_info
[
0
].
length
;
batch_lods
[
0
].
resize
(
static_cast
<
size_t
>
(
num_batch
+
1
));
batch_lods
[
0
].
resize
(
static_cast
<
size_t
>
(
max_seqlen
+
1
));
// batch_lods[1] is the raw index in the input LoDTensor
// batch_lods[1] is the raw index in the input LoDTensor
batch_lods
[
1
].
resize
(
static_cast
<
size_t
>
(
lod_tensor
.
dims
()[
0
]));
batch_lods
[
1
].
resize
(
static_cast
<
size_t
>
(
lod_tensor
.
dims
()[
0
]));
// batch_lods[2] is the sort order for the input LoDTensor.
// batch_lods[2] is the sort order for the input LoDTensor.
...
@@ -128,7 +128,7 @@ class LoDTensor2BatchFunctor {
...
@@ -128,7 +128,7 @@ class LoDTensor2BatchFunctor {
size_t
*
batch_starts
=
batch_lods
[
0
].
data
();
size_t
*
batch_starts
=
batch_lods
[
0
].
data
();
size_t
*
seq2batch_idx
=
batch_lods
[
1
].
data
();
size_t
*
seq2batch_idx
=
batch_lods
[
1
].
data
();
batch_starts
[
0
]
=
0
;
batch_starts
[
0
]
=
0
;
for
(
int
n
=
0
;
n
<
num_batch
;
n
++
)
{
for
(
int
n
=
0
;
n
<
max_seqlen
;
n
++
)
{
auto
batch_id
=
static_cast
<
int
>
(
batch_starts
[
n
]);
auto
batch_id
=
static_cast
<
int
>
(
batch_starts
[
n
]);
for
(
size_t
i
=
0
;
i
<
seq_info
.
size
();
++
i
)
{
for
(
size_t
i
=
0
;
i
<
seq_info
.
size
();
++
i
)
{
int
seq_len
=
seq_info
[
i
].
length
;
int
seq_len
=
seq_info
[
i
].
length
;
...
...
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