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
c11afdb5
P
Paddle
项目概览
PaddlePaddle
/
Paddle
大约 1 年 前同步成功
通知
2299
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
c11afdb5
编写于
1月 28, 2019
作者:
X
Xin Pan
提交者:
GitHub
1月 28, 2019
浏览文件
操作
浏览文件
下载
差异文件
Merge pull request #15516 from panyx0718/imperative3
imperative supports multi grad ops
上级
b9191902
42e61af8
变更
4
隐藏空白更改
内联
并排
Showing
4 changed file
with
118 addition
and
85 deletion
+118
-85
paddle/fluid/imperative/layer.cc
paddle/fluid/imperative/layer.cc
+46
-37
paddle/fluid/imperative/layer.h
paddle/fluid/imperative/layer.h
+12
-6
paddle/fluid/imperative/tracer.cc
paddle/fluid/imperative/tracer.cc
+48
-42
python/paddle/fluid/tests/unittests/test_imperative.py
python/paddle/fluid/tests/unittests/test_imperative.py
+12
-0
未找到文件。
paddle/fluid/imperative/layer.cc
浏览文件 @
c11afdb5
...
@@ -204,59 +204,68 @@ framework::LoDTensor& VarBase::GradValue() {
...
@@ -204,59 +204,68 @@ framework::LoDTensor& VarBase::GradValue() {
}
}
std
::
map
<
std
::
string
,
std
::
vector
<
VarBase
*>>
OpBase
::
ApplyGrad
()
{
std
::
map
<
std
::
string
,
std
::
vector
<
VarBase
*>>
OpBase
::
ApplyGrad
()
{
if
(
!
grad_op_desc_
&&
backward_id_
<=
0
)
{
if
(
grad_op_descs_
.
empty
()
&&
backward_id_
<=
0
)
{
LOG
(
WARNING
)
<<
"op with no grad: "
<<
op_desc_
->
Type
();
LOG
(
WARNING
)
<<
"op with no grad: "
<<
op_desc_
->
Type
();
return
{};
return
{};
}
}
std
::
map
<
std
::
string
,
std
::
vector
<
framework
::
Variable
*>
>
grad_outputs
;
std
::
vector
<
framework
::
VariableValueMap
>
grad_outputs
;
if
(
backward_id_
>
0
)
{
if
(
backward_id_
>
0
)
{
VLOG
(
3
)
<<
"py_layer_grad"
;
VLOG
(
3
)
<<
"py_layer_grad"
;
grad_outputs
[
framework
::
GradVarName
(
PyLayer
::
kFwdOut
)]
=
PyLayer
::
ApplyGrad
(
grad_outputs
.
resize
(
1
);
backward_id_
,
grad_outputs
[
0
][
framework
::
GradVarName
(
PyLayer
::
kFwdOut
)]
=
grad_input_vars_
[
framework
::
GradVarName
(
PyLayer
::
kFwdInp
)]);
PyLayer
::
ApplyGrad
(
backward_id_
,
grad_input_vars_
[
0
][
framework
::
GradVarName
(
PyLayer
::
kFwdInp
)]);
}
else
{
}
else
{
VLOG
(
3
)
<<
"op grad "
<<
grad_op_desc_
->
Type
();
grad_outputs
.
resize
(
grad_op_descs_
.
size
());
for
(
auto
it
:
grad_output_vars_
)
{
for
(
size_t
k
=
0
;
k
<
grad_op_descs_
.
size
();
++
k
)
{
auto
&
outputs
=
grad_outputs
[
it
.
first
];
framework
::
OpDesc
*
grad_op_desc
=
grad_op_descs_
[
k
];
for
(
size_t
i
=
0
;
i
<
it
.
second
.
size
();
++
i
)
{
VLOG
(
3
)
<<
"op grad "
<<
grad_op_desc
->
Type
();
// Allocate a new variable
for
(
auto
it
:
grad_output_vars_
[
k
])
{
Variable
*
tmp_var
=
new
framework
::
Variable
();
auto
&
outputs
=
grad_outputs
[
k
][
it
.
first
];
tmp_var
->
GetMutable
<
framework
::
LoDTensor
>
();
for
(
size_t
i
=
0
;
i
<
it
.
second
.
size
();
++
i
)
{
outputs
.
push_back
(
tmp_var
);
// Allocate a new variable
Variable
*
tmp_var
=
new
framework
::
Variable
();
tmp_var
->
GetMutable
<
framework
::
LoDTensor
>
();
outputs
.
push_back
(
tmp_var
);
}
}
}
}
framework
::
RuntimeContext
ctx
(
grad_input_vars_
,
grad_outputs
);
framework
::
RuntimeContext
ctx
(
grad_input_vars_
[
k
],
grad_outputs
[
k
]
);
// No need to do compile time infer shape here.
// No need to do compile time infer shape here.
// grad_op_desc_->InferShape(*block_);
// grad_op_desc_->InferShape(*block_);
grad_op_desc_
->
InferVarType
(
block_
);
grad_op_desc
->
InferVarType
(
block_
);
std
::
unique_ptr
<
framework
::
OperatorBase
>
opbase
=
std
::
unique_ptr
<
framework
::
OperatorBase
>
opbase
=
framework
::
OpRegistry
::
CreateOp
(
*
grad_op_desc_
);
framework
::
OpRegistry
::
CreateOp
(
*
grad_op_desc
);
framework
::
OperatorWithKernel
*
op_kernel
=
framework
::
OperatorWithKernel
*
op_kernel
=
dynamic_cast
<
framework
::
OperatorWithKernel
*>
(
opbase
.
get
());
dynamic_cast
<
framework
::
OperatorWithKernel
*>
(
opbase
.
get
());
PADDLE_ENFORCE_NOT_NULL
(
op_kernel
,
"only support op with kernel"
);
PADDLE_ENFORCE_NOT_NULL
(
op_kernel
,
"only support op with kernel"
);
framework
::
Scope
scope
;
framework
::
Scope
scope
;
PreparedOp
p
=
PreparedOp
::
Prepare
(
ctx
,
*
op_kernel
,
place_
);
PreparedOp
p
=
PreparedOp
::
Prepare
(
ctx
,
*
op_kernel
,
place_
);
p
.
op
.
RuntimeInferShape
(
scope
,
place_
,
ctx
);
p
.
op
.
RuntimeInferShape
(
scope
,
place_
,
ctx
);
p
.
func
(
framework
::
ExecutionContext
(
p
.
op
,
scope
,
*
p
.
dev_ctx
,
p
.
ctx
));
p
.
func
(
framework
::
ExecutionContext
(
p
.
op
,
scope
,
*
p
.
dev_ctx
,
p
.
ctx
));
}
}
}
for
(
auto
it
:
grad_output_vars_
)
{
for
(
size_t
k
=
0
;
k
<
grad_output_vars_
.
size
();
++
k
)
{
auto
&
outputs
=
grad_outputs
[
it
.
first
];
for
(
auto
it
:
grad_output_vars_
[
k
])
{
auto
&
origin_outputs
=
it
.
second
;
auto
&
outputs
=
grad_outputs
[
k
][
it
.
first
];
PADDLE_ENFORCE_EQ
(
outputs
.
size
(),
origin_outputs
.
size
());
auto
&
origin_outputs
=
it
.
second
;
PADDLE_ENFORCE_EQ
(
outputs
.
size
(),
origin_outputs
.
size
());
for
(
size_t
i
=
0
;
i
<
outputs
.
size
();
++
i
)
{
framework
::
Variable
*
grad
=
outputs
[
i
];
for
(
size_t
i
=
0
;
i
<
outputs
.
size
();
++
i
)
{
framework
::
Variable
*
orig_grad
=
origin_outputs
[
i
];
framework
::
Variable
*
grad
=
outputs
[
i
];
AddTo
(
grad
,
orig_grad
,
place_
);
framework
::
Variable
*
orig_grad
=
origin_outputs
[
i
];
delete
grad
;
AddTo
(
grad
,
orig_grad
,
place_
);
delete
grad
;
}
}
}
}
}
return
input_vars_
;
return
input_vars_
;
}
}
...
...
paddle/fluid/imperative/layer.h
浏览文件 @
c11afdb5
...
@@ -184,12 +184,13 @@ class OpBase {
...
@@ -184,12 +184,13 @@ class OpBase {
OpBase
()
OpBase
()
:
op_desc_
(
nullptr
),
:
op_desc_
(
nullptr
),
forward_id_
(
-
1
),
forward_id_
(
-
1
),
grad_op_desc_
(
nullptr
),
backward_id_
(
-
1
),
backward_id_
(
-
1
),
place_
(
platform
::
CPUPlace
())
{}
place_
(
platform
::
CPUPlace
())
{}
virtual
~
OpBase
()
{
virtual
~
OpBase
()
{
if
(
grad_op_desc_
)
delete
grad_op_desc_
;
for
(
framework
::
OpDesc
*
desc
:
grad_op_descs_
)
{
delete
desc
;
}
}
}
std
::
map
<
std
::
string
,
std
::
vector
<
VarBase
*>>
ApplyGrad
();
std
::
map
<
std
::
string
,
std
::
vector
<
VarBase
*>>
ApplyGrad
();
...
@@ -198,9 +199,11 @@ class OpBase {
...
@@ -198,9 +199,11 @@ class OpBase {
// For pure python PyLayer, use `forward_id_`, otherwise, use op_desc_.
// For pure python PyLayer, use `forward_id_`, otherwise, use op_desc_.
framework
::
OpDesc
*
op_desc_
;
framework
::
OpDesc
*
op_desc_
;
int
forward_id_
;
int
forward_id_
;
// When has backward, one of `grad_op_desc_` or `backward_id_` is set,
// When has backward, one of `grad_op_descs_` or `backward_id_` is set,
// not both.
// not both.
framework
::
OpDesc
*
grad_op_desc_
;
// Note: each fwd op corresponds to a vector of bwd ops.
std
::
vector
<
framework
::
OpDesc
*>
grad_op_descs_
;
int
backward_id_
;
int
backward_id_
;
platform
::
Place
place_
;
platform
::
Place
place_
;
...
@@ -210,8 +213,11 @@ class OpBase {
...
@@ -210,8 +213,11 @@ class OpBase {
OpBasePtrMap
pre_ops_
;
OpBasePtrMap
pre_ops_
;
std
::
map
<
std
::
string
,
std
::
vector
<
int
>>
pre_ops_out_idx_
;
std
::
map
<
std
::
string
,
std
::
vector
<
int
>>
pre_ops_out_idx_
;
framework
::
VariableValueMap
grad_input_vars_
;
// Inputs to a vector of bwd ops.
framework
::
VariableValueMap
grad_output_vars_
;
std
::
vector
<
framework
::
VariableValueMap
>
grad_input_vars_
;
// Outputs to a vector of bwd ops.
std
::
vector
<
framework
::
VariableValueMap
>
grad_output_vars_
;
framework
::
BlockDesc
*
block_
;
framework
::
BlockDesc
*
block_
;
};
};
...
...
paddle/fluid/imperative/tracer.cc
浏览文件 @
c11afdb5
...
@@ -24,15 +24,16 @@ namespace imperative {
...
@@ -24,15 +24,16 @@ namespace imperative {
void
CreateGradOp
(
const
framework
::
OpDesc
&
op_desc
,
void
CreateGradOp
(
const
framework
::
OpDesc
&
op_desc
,
const
std
::
unordered_set
<
std
::
string
>&
no_grad_set
,
const
std
::
unordered_set
<
std
::
string
>&
no_grad_set
,
const
std
::
vector
<
framework
::
BlockDesc
*>&
grad_sub_block
,
const
std
::
vector
<
framework
::
BlockDesc
*>&
grad_sub_block
,
framework
::
OpDesc
**
grad_op_desc
,
std
::
vector
<
framework
::
OpDesc
*>*
grad_op_descs
,
std
::
unordered_map
<
std
::
string
,
std
::
string
>*
grad_to_var
)
{
std
::
unordered_map
<
std
::
string
,
std
::
string
>*
grad_to_var
)
{
std
::
vector
<
std
::
unique_ptr
<
framework
::
OpDesc
>>
grad_op_descs
=
PADDLE_ENFORCE
(
grad_op_descs
->
empty
());
std
::
vector
<
std
::
unique_ptr
<
framework
::
OpDesc
>>
descs
=
framework
::
OpInfoMap
::
Instance
()
framework
::
OpInfoMap
::
Instance
()
.
Get
(
op_desc
.
Type
())
.
Get
(
op_desc
.
Type
())
.
GradOpMaker
()(
op_desc
,
no_grad_set
,
grad_to_var
,
grad_sub_block
);
.
GradOpMaker
()(
op_desc
,
no_grad_set
,
grad_to_var
,
grad_sub_block
);
PADDLE_ENFORCE
(
grad_op_descs
.
size
()
==
1
,
"Only support 1 grad op now."
);
for
(
auto
&
desc
:
descs
)
{
// TODO(panyx0718): Leak?
grad_op_descs
->
emplace_back
(
desc
.
release
());
*
grad_op_desc
=
grad_op_descs
[
0
].
release
();
}
}
}
void
InitVar
(
framework
::
Variable
*
var
,
framework
::
Variable
*
grad_var
,
void
InitVar
(
framework
::
Variable
*
var
,
framework
::
Variable
*
grad_var
,
...
@@ -138,49 +139,52 @@ void Tracer::Trace(OpBase* op, const VarBasePtrMap& inputs,
...
@@ -138,49 +139,52 @@ void Tracer::Trace(OpBase* op, const VarBasePtrMap& inputs,
prepared_op
.
op
,
scope
,
*
prepared_op
.
dev_ctx
,
prepared_op
.
ctx
));
prepared_op
.
op
,
scope
,
*
prepared_op
.
dev_ctx
,
prepared_op
.
ctx
));
if
(
!
stop_gradient
)
{
if
(
!
stop_gradient
)
{
framework
::
OpDesc
*
grad_op_desc
;
// TODO(panyx): Is this leaked?
std
::
unique_ptr
<
std
::
unordered_map
<
std
::
string
,
std
::
string
>>
grad_to_var
(
std
::
unique_ptr
<
std
::
unordered_map
<
std
::
string
,
std
::
string
>>
grad_to_var
(
new
std
::
unordered_map
<
std
::
string
,
std
::
string
>
());
new
std
::
unordered_map
<
std
::
string
,
std
::
string
>
());
CreateGradOp
(
*
op_desc
,
{},
{
block
},
&
grad_op_desc
,
grad_to_var
.
get
());
CreateGradOp
(
*
op_desc
,
{},
{
block
},
&
op
->
grad_op_descs_
,
grad_to_var
.
get
());
op
->
grad_op_desc_
=
grad_op_desc
;
op
->
grad_input_vars_
.
resize
(
op
->
grad_op_descs_
.
size
());
for
(
auto
it
:
grad_op_desc
->
Inputs
())
{
op
->
grad_output_vars_
.
resize
(
op
->
grad_op_descs_
.
size
());
auto
&
grad_in_vars
=
op
->
grad_input_vars_
[
it
.
first
];
for
(
size_t
i
=
0
;
i
<
op
->
grad_op_descs_
.
size
();
++
i
)
{
for
(
const
std
::
string
&
grad_invar
:
it
.
second
)
{
framework
::
OpDesc
*
grad_op_desc
=
op
->
grad_op_descs_
[
i
];
block
->
FindRecursiveOrCreateVar
(
grad_invar
);
for
(
auto
it
:
grad_op_desc
->
Inputs
())
{
auto
var_it
=
grad_to_var
->
find
(
grad_invar
);
auto
&
grad_in_vars
=
op
->
grad_input_vars_
[
i
][
it
.
first
];
if
(
var_it
==
grad_to_var
->
end
())
{
for
(
const
std
::
string
&
grad_invar
:
it
.
second
)
{
auto
fwd_var_it
=
vars
.
find
(
grad_invar
);
block
->
FindRecursiveOrCreateVar
(
grad_invar
);
PADDLE_ENFORCE
(
fwd_var_it
!=
vars
.
end
());
auto
var_it
=
grad_to_var
->
find
(
grad_invar
);
// Forward inputs or outputs.
if
(
var_it
==
grad_to_var
->
end
())
{
grad_in_vars
.
push_back
(
fwd_var_it
->
second
->
var_
);
auto
fwd_var_it
=
vars
.
find
(
grad_invar
);
}
else
{
PADDLE_ENFORCE
(
fwd_var_it
!=
vars
.
end
());
// Forward inputs or outputs.
grad_in_vars
.
push_back
(
fwd_var_it
->
second
->
var_
);
}
else
{
VarBase
*
var
=
vars
[
var_it
->
second
];
if
(
!
var
->
grads_
->
var_
->
IsInitialized
())
{
InitVar
(
var
->
var_
,
var
->
grads_
->
var_
,
prepared_op
.
GetDeviceContext
());
}
// Douts.
grad_in_vars
.
push_back
(
var
->
grads_
->
var_
);
}
}
}
for
(
auto
it
:
grad_op_desc
->
Outputs
())
{
auto
&
grad_out_vars
=
op
->
grad_output_vars_
[
i
][
it
.
first
];
for
(
const
std
::
string
&
grad_outvar
:
it
.
second
)
{
block
->
FindRecursiveOrCreateVar
(
grad_outvar
);
auto
var_it
=
grad_to_var
->
find
(
grad_outvar
);
PADDLE_ENFORCE
(
var_it
!=
grad_to_var
->
end
(),
"Could not found the grad op output var, should this "
"operator %s's stop gradient be True"
,
op_desc
->
Type
());
VarBase
*
var
=
vars
[
var_it
->
second
];
VarBase
*
var
=
vars
[
var_it
->
second
];
if
(
!
var
->
grads_
->
var_
->
IsInitialized
())
{
if
(
!
var
->
grads_
->
var_
->
IsInitialized
())
{
InitVar
(
var
->
var_
,
var
->
grads_
->
var_
,
InitVar
(
var
->
var_
,
var
->
grads_
->
var_
,
prepared_op
.
GetDeviceContext
());
prepared_op
.
GetDeviceContext
());
}
}
// Douts.
grad_out_vars
.
push_back
(
var
->
grads_
->
var_
);
grad_in_vars
.
push_back
(
var
->
grads_
->
var_
);
}
}
}
for
(
auto
it
:
grad_op_desc
->
Outputs
())
{
auto
&
grad_out_vars
=
op
->
grad_output_vars_
[
it
.
first
];
for
(
const
std
::
string
&
grad_outvar
:
it
.
second
)
{
block
->
FindRecursiveOrCreateVar
(
grad_outvar
);
auto
var_it
=
grad_to_var
->
find
(
grad_outvar
);
PADDLE_ENFORCE
(
var_it
!=
grad_to_var
->
end
(),
"Could not found the grad op output var, should this "
"operator %s's stop gradient be True"
,
op_desc
->
Type
());
VarBase
*
var
=
vars
[
var_it
->
second
];
if
(
!
var
->
grads_
->
var_
->
IsInitialized
())
{
InitVar
(
var
->
var_
,
var
->
grads_
->
var_
,
prepared_op
.
GetDeviceContext
());
}
}
grad_out_vars
.
push_back
(
var
->
grads_
->
var_
);
}
}
}
}
}
}
...
@@ -209,10 +213,12 @@ std::vector<VarBase*> Tracer::PyTrace(OpBase* op,
...
@@ -209,10 +213,12 @@ std::vector<VarBase*> Tracer::PyTrace(OpBase* op,
out
->
TrackPreOp
(
op
,
PyLayer
::
kFwdOut
,
i
,
stop_gradient
);
out
->
TrackPreOp
(
op
,
PyLayer
::
kFwdOut
,
i
,
stop_gradient
);
}
}
if
(
!
stop_gradient
)
{
if
(
!
stop_gradient
)
{
op
->
grad_input_vars_
.
resize
(
1
);
op
->
grad_output_vars_
.
resize
(
1
);
auto
&
grad_input_vars
=
auto
&
grad_input_vars
=
op
->
grad_input_vars_
[
framework
::
GradVarName
(
PyLayer
::
kFwdInp
)];
op
->
grad_input_vars_
[
0
][
framework
::
GradVarName
(
PyLayer
::
kFwdInp
)];
auto
&
grad_output_vars
=
auto
&
grad_output_vars
=
op
->
grad_output_vars_
[
framework
::
GradVarName
(
PyLayer
::
kFwdOut
)];
op
->
grad_output_vars_
[
0
][
framework
::
GradVarName
(
PyLayer
::
kFwdOut
)];
for
(
const
VarBase
*
inp
:
inputs
)
{
for
(
const
VarBase
*
inp
:
inputs
)
{
grad_input_vars
.
push_back
(
inp
->
var_
);
grad_input_vars
.
push_back
(
inp
->
var_
);
...
...
python/paddle/fluid/tests/unittests/test_imperative.py
浏览文件 @
c11afdb5
...
@@ -67,6 +67,18 @@ class MLP(fluid.imperative.Layer):
...
@@ -67,6 +67,18 @@ class MLP(fluid.imperative.Layer):
class
TestImperative
(
unittest
.
TestCase
):
class
TestImperative
(
unittest
.
TestCase
):
def
test_sum_op
(
self
):
x
=
np
.
ones
([
2
,
2
],
np
.
float32
)
with
fluid
.
imperative
.
guard
():
inputs
=
[]
for
_
in
range
(
10
):
inputs
.
append
(
fluid
.
imperative
.
base
.
to_variable
(
x
))
ret
=
fluid
.
layers
.
sums
(
inputs
)
loss
=
fluid
.
layers
.
reduce_sum
(
ret
)
loss
.
_backward
()
self
.
assertTrue
(
np
.
allclose
(
ret
.
_numpy
(),
x
*
10
))
self
.
assertTrue
(
np
.
allclose
(
inputs
[
0
].
_gradient
(),
x
))
def
test_layer
(
self
):
def
test_layer
(
self
):
with
fluid
.
imperative
.
guard
():
with
fluid
.
imperative
.
guard
():
cl
=
core
.
Layer
()
cl
=
core
.
Layer
()
...
...
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