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
927767b6
P
Paddle
项目概览
PaddlePaddle
/
Paddle
大约 1 年 前同步成功
通知
2298
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
提交
927767b6
编写于
9月 30, 2017
作者:
Y
Yancey1989
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
add some checking
上级
a35e82a6
变更
2
显示空白变更内容
内联
并排
Showing
2 changed file
with
63 addition
and
36 deletion
+63
-36
paddle/operators/sequence_concat_op.cc
paddle/operators/sequence_concat_op.cc
+26
-18
paddle/operators/sequence_concat_op.h
paddle/operators/sequence_concat_op.h
+37
-18
未找到文件。
paddle/operators/sequence_concat_op.cc
浏览文件 @
927767b6
...
@@ -23,18 +23,19 @@ class SequenceConcatOp : public framework::OperatorWithKernel {
...
@@ -23,18 +23,19 @@ class SequenceConcatOp : public framework::OperatorWithKernel {
protected:
protected:
void
InferShape
(
framework
::
InferShapeContextBase
*
ctx
)
const
override
{
void
InferShape
(
framework
::
InferShapeContextBase
*
ctx
)
const
override
{
PADDLE_ENFORCE
_GT
(
ctx
->
Inputs
(
"X"
).
size
(),
0UL
,
PADDLE_ENFORCE
(
ctx
->
HasInputs
(
"X"
)
,
"Inputs(X) of SequenceConcatOp should not be empty
."
);
"Inputs(X) of SequenceConcatOp should not be null
."
);
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"Out"
),
PADDLE_ENFORCE
(
ctx
->
HasOutput
(
"Out"
),
"Output(Out) of SequenceConcatOp should not be null."
);
"Output(Out) of SequenceConcatOp should not be null."
);
const
size_t
level
=
static_cast
<
size_t
>
(
ctx
->
Attrs
().
Get
<
int
>
(
"level"
));
const
size_t
level
=
static_cast
<
size_t
>
(
ctx
->
Attrs
().
Get
<
int
>
(
"level"
));
const
size_t
axis
=
static_cast
<
size_t
>
(
ctx
->
Attrs
().
Get
<
int
>
(
"axis"
));
const
size_t
axis
=
static_cast
<
size_t
>
(
ctx
->
Attrs
().
Get
<
int
>
(
"axis"
));
PADDLE_ENFORCE
(
level
==
0UL
||
level
==
1UL
,
PADDLE_ENFORCE
(
level
==
0UL
||
level
==
1UL
,
"Sequence Concat Op only support one or two sequence now."
);
"The sequence_concat operator only accepts sequence "
"or a nested sequence as its input."
);
auto
ins_dims
=
ctx
->
GetInputsDim
(
"X"
);
auto
ins_dims
=
ctx
->
GetInputsDim
(
"X"
);
framework
::
DDim
out_dims
=
ins_dims
[
0
];
framework
::
DDim
out_dims
=
ins_dims
[
0
];
const
size_t
n
=
ins_dims
.
size
();
const
size_t
n
=
ins_dims
.
size
();
for
(
size_t
i
=
1
;
i
<
n
;
i
++
)
{
for
(
size_t
i
=
1
;
i
<
n
;
++
i
)
{
out_dims
[
axis
]
+=
ins_dims
[
i
][
axis
];
out_dims
[
axis
]
+=
ins_dims
[
i
][
axis
];
}
}
ctx
->
SetOutputDim
(
"Out"
,
out_dims
);
ctx
->
SetOutputDim
(
"Out"
,
out_dims
);
...
@@ -47,33 +48,40 @@ class SequenceConcatOpMaker : public framework::OpProtoAndCheckerMaker {
...
@@ -47,33 +48,40 @@ class SequenceConcatOpMaker : public framework::OpProtoAndCheckerMaker {
framework
::
OpAttrChecker
*
op_checker
)
framework
::
OpAttrChecker
*
op_checker
)
:
OpProtoAndCheckerMaker
(
proto
,
op_checker
)
{
:
OpProtoAndCheckerMaker
(
proto
,
op_checker
)
{
AddInput
(
"X"
,
AddInput
(
"X"
,
"
Multip LodTensors, the variable-length inputs of
"
"
The input Multip LoDTensors, which are variable-length
"
"
SequenceConcatOp
"
)
"
sequence or nested sequence.
"
)
.
AsDuplicable
();
.
AsDuplicable
();
AddOutput
(
"Out"
,
AddOutput
(
"Out"
,
"A
float Lod
Tensor, the variable-length output of "
"A
LoD
Tensor, the variable-length output of "
"
SequenceConcat
Op."
);
"
sequence_concat
Op."
);
AddAttr
<
int
>
(
"axis"
,
AddAttr
<
int
>
(
"axis"
,
"(int, default 0)"
"The axis which the inputs will be joined with."
"The axis which the inputs will be joined with."
"If axis is 0, the inputs will be joined with Lo
d
index."
)
"If axis is 0, the inputs will be joined with Lo
D
index."
)
.
SetDefault
(
0
);
.
SetDefault
(
0
);
AddAttr
<
int
>
(
"level"
,
AddAttr
<
int
>
(
"level"
,
"(int, default 0)"
"The level which the inputs will be joined with."
"The level which the inputs will be joined with."
"If level is 0, the inputs will be joined with word."
"If level is 0, the inputs will be joined with "
"If level is 1, the inputs will be joined with sentence."
)
"nested sequences."
"If level is 1, the inputs will be joined with sequences."
)
.
SetDefault
(
0
);
.
SetDefault
(
0
);
AddComment
(
R"DOC(
AddComment
(
R"DOC(
SequenceConcatOp concat multip LodTensors and only supports one or two levels.
The sequence_concat operator concatenates multiple LoDTensors.
It only supports sequences ( LoD Tensor with level=1)
or nested sequences (LoD tensor with level=0) as its inputs.
- Case1:
- Case1:
axis is 1, level is 1, the Lod
of Inputs are the same,
If the axis is 1, level is 1, the LoD
of Inputs are the same,
LoD(x0) = {{0,2,4},{0,1,2,3,4}}; Dims(x0) = (2,3,4)
LoD(x0) = {{0,2,4},{0,1,2,3,4}}; Dims(x0) = (2,3,4)
LoD(x1) = {{0,2,4},{0,1,2,3,4}}; Dims(x1) = (2,4,4)
LoD(x1) = {{0,2,4},{0,1,2,3,4}}; Dims(x1) = (2,4,4)
LoD(Out) = {{0,2,4},{01,2,3,4}}; Dims(Out) = (2,7,4)
LoD(Out) = {{0,2,4},{0
,
1,2,3,4}}; Dims(Out) = (2,7,4)
- Case2:
- Case2:
If
axis is 0, level is 1, the Lod
of inputs are different,
If
the axis is 0, level is 1, the LoD
of inputs are different,
LoD(x0) = {{0,2,4}, {0,1,2,3,4}}; Dims(x0) = (2,3,4)
LoD(x0) = {{0,2,4}, {0,1,2,3,4}}; Dims(x0) = (2,3,4)
LoD(x1) = {{0,3,5}, {0,1,3,4,5}}; Dims(x1) = (3,3,4)
LoD(x1) = {{0,3,5}, {0,1,3,4,5}}; Dims(x1) = (3,3,4)
LoD(Out) = {{0,5,9}, {0,1,2,4,5,6,7,8,9}}; Dims(Out) = (5,3,4)
LoD(Out) = {{0,5,9}, {0,1,2,4,5,6,7,8,9}}; Dims(Out) = (5,3,4)
NOTE: The level of all the inputs should be the same.
)DOC"
);
)DOC"
);
}
}
};
};
...
@@ -85,9 +93,9 @@ class SequenceConcatGradOp : public framework::OperatorWithKernel {
...
@@ -85,9 +93,9 @@ class SequenceConcatGradOp : public framework::OperatorWithKernel {
protected:
protected:
void
InferShape
(
framework
::
InferShapeContextBase
*
ctx
)
const
override
{
void
InferShape
(
framework
::
InferShapeContextBase
*
ctx
)
const
override
{
PADDLE_ENFORCE
(
ctx
->
HasInput
(
framework
::
GradVarName
(
"Out"
)),
PADDLE_ENFORCE
(
ctx
->
HasInput
(
framework
::
GradVarName
(
"Out"
)),
"
G
radient of Out should not be null."
);
"
The g
radient of Out should not be null."
);
PADDLE_ENFORCE
_GT
(
ctx
->
Outputs
(
framework
::
GradVarName
(
"X"
)).
size
(),
0UL
,
PADDLE_ENFORCE
(
ctx
->
HasOutputs
(
framework
::
GradVarName
(
"X"
))
,
"Gradient of X should not be empty."
)
"The gradient of X should not be empty."
);
ctx
->
SetOutputsDim
(
framework
::
GradVarName
(
"X"
),
ctx
->
GetInputsDim
(
"X"
));
ctx
->
SetOutputsDim
(
framework
::
GradVarName
(
"X"
),
ctx
->
GetInputsDim
(
"X"
));
}
}
};
};
...
...
paddle/operators/sequence_concat_op.h
浏览文件 @
927767b6
...
@@ -23,7 +23,7 @@ using Tensor = framework::Tensor;
...
@@ -23,7 +23,7 @@ using Tensor = framework::Tensor;
using
LoDTensor
=
framework
::
LoDTensor
;
using
LoDTensor
=
framework
::
LoDTensor
;
using
LoD
=
framework
::
LoD
;
using
LoD
=
framework
::
LoD
;
// Concat Lo
d, the initialized Lod
of Output is lod(x0),
// Concat Lo
D, the initialized LoD
of Output is lod(x0),
// if axis is not 0, the LoD(Out) will be the same as Inputs, if axis is 0:
// if axis is not 0, the LoD(Out) will be the same as Inputs, if axis is 0:
// Case1:
// Case1:
// There is one level, the Output LoD will be modified:
// There is one level, the Output LoD will be modified:
...
@@ -37,26 +37,26 @@ using LoD = framework::LoD;
...
@@ -37,26 +37,26 @@ using LoD = framework::LoD;
// LoD(x1) = {{0,3,5}, {0,1,3,4,5}}
// LoD(x1) = {{0,3,5}, {0,1,3,4,5}}
// LoD(Out) = {{0,5,9}, {0,1,2,4,5,6,7,8,9}}
// LoD(Out) = {{0,5,9}, {0,1,2,4,5,6,7,8,9}}
template
<
typename
T
>
template
<
typename
T
>
LoD
concatLo
d
(
const
std
::
vector
<
const
T
*>
ins
,
const
size_t
axis
,
LoD
concatLo
D
(
const
std
::
vector
<
const
T
*>
ins
,
const
size_t
axis
,
const
size_t
level
)
{
const
size_t
level
)
{
auto
out_lod
=
ins
[
0
]
->
lod
();
auto
out_lod
=
ins
[
0
]
->
lod
();
const
size_t
n
=
ins
.
size
();
const
size_t
n
=
ins
.
size
();
if
(
axis
==
0UL
)
{
if
(
axis
==
0UL
)
{
if
(
level
==
0
)
{
if
(
level
==
0
)
{
for
(
size_t
i
=
1
;
i
<
n
;
i
++
)
{
for
(
size_t
i
=
1
;
i
<
n
;
++
i
)
{
for
(
size_t
j
=
0
;
j
<
ins
[
i
]
->
lod
()[
0
].
size
();
j
++
)
{
for
(
size_t
j
=
0
;
j
<
ins
[
i
]
->
lod
()[
0
].
size
();
++
j
)
{
out_lod
[
0
][
j
]
+=
ins
[
i
]
->
lod
()[
0
][
j
];
out_lod
[
0
][
j
]
+=
ins
[
i
]
->
lod
()[
0
][
j
];
}
}
}
}
}
else
if
(
level
==
1
)
{
}
else
if
(
level
==
1
)
{
for
(
size_t
i
=
1
;
i
<
n
;
i
++
)
{
PADDLE_ENFORCE_EQ
(
ins
[
0
]
->
NumLevels
(),
2UL
,
PADDLE_ENFORCE_EQ
(
ins
[
i
]
->
NumLevels
(),
2UL
,
"If the level is 1, all of the inputs "
"All the LoDTensors of Inputs(X) should "
"should be the the nested sequence."
);
"have two level."
);
for
(
size_t
i
=
1
;
i
<
n
;
++
i
)
{
for
(
size_t
j
=
0
;
j
<
ins
[
i
]
->
lod
()[
0
].
size
();
j
++
)
{
for
(
size_t
j
=
0
;
j
<
ins
[
i
]
->
lod
()[
0
].
size
();
++
j
)
{
out_lod
[
0
].
push_back
(
ins
[
i
]
->
lod
()[
0
][
j
]);
out_lod
[
0
].
push_back
(
ins
[
i
]
->
lod
()[
0
][
j
]);
}
}
for
(
size_t
j
=
0
;
j
<
ins
[
i
]
->
lod
()[
1
].
size
();
j
++
)
{
for
(
size_t
j
=
0
;
j
<
ins
[
i
]
->
lod
()[
1
].
size
();
++
j
)
{
out_lod
[
1
][
j
]
+=
ins
[
i
]
->
lod
()[
1
][
j
];
out_lod
[
1
][
j
]
+=
ins
[
i
]
->
lod
()[
1
][
j
];
}
}
}
}
...
@@ -66,7 +66,7 @@ LoD concatLod(const std::vector<const T*> ins, const size_t axis,
...
@@ -66,7 +66,7 @@ LoD concatLod(const std::vector<const T*> ins, const size_t axis,
}
}
template
<
typename
Place
,
typename
T
>
template
<
typename
Place
,
typename
T
>
class
SequenceConcatOpKernel
:
public
framework
::
OpKernel
{
class
SequenceConcatOpKernel
:
public
framework
::
OpKernel
<
T
>
{
public:
public:
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
auto
ins
=
ctx
.
MultiInput
<
LoDTensor
>
(
"X"
);
auto
ins
=
ctx
.
MultiInput
<
LoDTensor
>
(
"X"
);
...
@@ -74,18 +74,37 @@ class SequenceConcatOpKernel : public framework::OpKernel {
...
@@ -74,18 +74,37 @@ class SequenceConcatOpKernel : public framework::OpKernel {
const
size_t
axis
=
static_cast
<
size_t
>
(
ctx
.
Attr
<
int
>
(
"axis"
));
const
size_t
axis
=
static_cast
<
size_t
>
(
ctx
.
Attr
<
int
>
(
"axis"
));
const
size_t
level
=
static_cast
<
size_t
>
(
ctx
.
Attr
<
int
>
(
"level"
));
const
size_t
level
=
static_cast
<
size_t
>
(
ctx
.
Attr
<
int
>
(
"level"
));
const
size_t
n
=
ins
.
size
();
const
size_t
n
=
ins
.
size
();
for
(
size_t
i
=
1
;
i
<
n
;
++
i
)
{
PADDLE_ENFORCE_EQ
(
ins
[
0
]
->
NumLevels
(),
ins
[
i
]
->
NumLevels
(),
"The level number of all the input LoDTensors "
"should be the same."
);
PADDLE_ENFORCE_EQ
(
ins
[
0
]
->
dims
().
size
(),
ins
[
i
]
->
dims
().
size
(),
"The dimensions size of all the input LoDTensors "
"should be the same."
);
const
size_t
dims_size
=
ins
[
i
]
->
dims
().
size
();
for
(
size_t
j
=
0
;
j
<
dims_size
;
++
j
)
{
if
(
j
==
axis
)
continue
;
PADDLE_ENFORCE_EQ
(
ins
[
0
]
->
dims
()[
j
],
ins
[
i
]
->
dims
()[
j
],
"The dimensions of all the input LoDTensors "
"except for the specify axis should be "
"matched exactly."
);
}
}
out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
out
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
auto
out_lod
=
concatLo
d
<
LoDTensor
>
(
ins
,
axis
,
level
);
auto
out_lod
=
concatLo
D
<
LoDTensor
>
(
ins
,
axis
,
level
);
out
->
set_lod
(
out_lod
);
out
->
set_lod
(
out_lod
);
auto
out_lod_level
=
out_lod
[
level
];
auto
out_lod_level
=
out_lod
[
level
];
for
(
size_t
i
=
0
;
i
<
out_lod_level
.
size
()
-
1
;
i
++
)
{
for
(
size_t
i
=
0
;
i
<
out_lod_level
.
size
()
-
1
;
++
i
)
{
Tensor
out_t
=
out
->
Slice
<
T
>
(
static_cast
<
int
>
(
out_lod_level
[
i
]),
Tensor
out_t
=
out
->
Slice
<
T
>
(
static_cast
<
int
>
(
out_lod_level
[
i
]),
static_cast
<
int
>
(
out_lod_level
[
i
+
1
]));
static_cast
<
int
>
(
out_lod_level
[
i
+
1
]));
auto
out_stride
=
framework
::
stride
(
out_t
.
dims
());
auto
out_stride
=
framework
::
stride
(
out_t
.
dims
());
size_t
offset
=
0
;
size_t
offset
=
0
;
for
(
size_t
j
=
0
;
j
<
n
;
j
++
)
{
for
(
size_t
j
=
0
;
j
<
n
;
++
j
)
{
auto
in_lod_level
=
ins
[
j
]
->
lod
()[
level
];
auto
in_lod_level
=
ins
[
j
]
->
lod
()[
level
];
auto
in_stride
=
framework
::
stride
(
ins
[
j
]
->
dims
());
auto
in_stride
=
framework
::
stride
(
ins
[
j
]
->
dims
());
Tensor
in_t
=
ins
[
j
]
->
Slice
<
T
>
(
static_cast
<
int
>
(
in_lod_level
[
i
]),
Tensor
in_t
=
ins
[
j
]
->
Slice
<
T
>
(
static_cast
<
int
>
(
in_lod_level
[
i
]),
...
@@ -100,7 +119,7 @@ class SequenceConcatOpKernel : public framework::OpKernel {
...
@@ -100,7 +119,7 @@ class SequenceConcatOpKernel : public framework::OpKernel {
};
};
template
<
typename
Place
,
typename
T
>
template
<
typename
Place
,
typename
T
>
class
SequenceConcatGradOpKernel
:
public
framework
::
OpKernel
{
class
SequenceConcatGradOpKernel
:
public
framework
::
OpKernel
<
T
>
{
public:
public:
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
void
Compute
(
const
framework
::
ExecutionContext
&
ctx
)
const
override
{
auto
ins
=
ctx
.
MultiInput
<
framework
::
LoDTensor
>
(
"X"
);
auto
ins
=
ctx
.
MultiInput
<
framework
::
LoDTensor
>
(
"X"
);
...
@@ -118,17 +137,17 @@ class SequenceConcatGradOpKernel : public framework::OpKernel {
...
@@ -118,17 +137,17 @@ class SequenceConcatGradOpKernel : public framework::OpKernel {
x_grads
[
i
]
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
x_grads
[
i
]
->
mutable_data
<
T
>
(
ctx
.
GetPlace
());
}
}
auto
out_lod
=
concatLo
d
<
LoDTensor
>
(
ins
,
axis
,
level
);
auto
out_lod
=
concatLo
D
<
LoDTensor
>
(
ins
,
axis
,
level
);
auto
out_lod_level
=
out_lod
[
level
];
auto
out_lod_level
=
out_lod
[
level
];
for
(
size_t
i
=
0
;
i
<
out_lod_level
.
size
()
-
1
;
i
++
)
{
for
(
size_t
i
=
0
;
i
<
out_lod_level
.
size
()
-
1
;
++
i
)
{
Tensor
out_grad_t
=
Tensor
out_grad_t
=
out_grad
->
Slice
<
T
>
(
static_cast
<
int
>
(
out_lod_level
[
i
]),
out_grad
->
Slice
<
T
>
(
static_cast
<
int
>
(
out_lod_level
[
i
]),
static_cast
<
int
>
(
out_lod_level
[
i
+
1
]));
static_cast
<
int
>
(
out_lod_level
[
i
+
1
]));
auto
out_grad_stride
=
framework
::
stride
(
out_grad_t
.
dims
());
auto
out_grad_stride
=
framework
::
stride
(
out_grad_t
.
dims
());
size_t
offset
=
0
;
size_t
offset
=
0
;
for
(
size_t
j
=
0
;
j
<
n
;
j
++
)
{
for
(
size_t
j
=
0
;
j
<
n
;
++
j
)
{
auto
x_grad_lod_level
=
x_grads
[
j
]
->
lod
()[
level
];
auto
x_grad_lod_level
=
x_grads
[
j
]
->
lod
()[
level
];
auto
x_grad_stride
=
framework
::
stride
(
x_grads
[
j
]
->
dims
());
auto
x_grad_stride
=
framework
::
stride
(
x_grads
[
j
]
->
dims
());
Tensor
x_grad_t
=
Tensor
x_grad_t
=
...
...
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