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
2e3ec66b
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
体验新版 GitCode,发现更多精彩内容 >>
提交
2e3ec66b
编写于
8月 27, 2019
作者:
J
joanna.wozna.intel
提交者:
Tao Luo
8月 27, 2019
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Add conv dequant squash for int8 (#18905)
上级
482ce818
变更
5
隐藏空白更改
内联
并排
Showing
5 changed file
with
129 addition
and
0 deletion
+129
-0
paddle/fluid/framework/ir/graph_pattern_detector.cc
paddle/fluid/framework/ir/graph_pattern_detector.cc
+18
-0
paddle/fluid/framework/ir/graph_pattern_detector.h
paddle/fluid/framework/ir/graph_pattern_detector.h
+17
-0
paddle/fluid/framework/ir/mkldnn/cpu_quantize_squash_pass.cc
paddle/fluid/framework/ir/mkldnn/cpu_quantize_squash_pass.cc
+33
-0
paddle/fluid/framework/ir/mkldnn/cpu_quantize_squash_pass.h
paddle/fluid/framework/ir/mkldnn/cpu_quantize_squash_pass.h
+5
-0
paddle/fluid/framework/ir/mkldnn/cpu_quantize_squash_pass_tester.cc
...id/framework/ir/mkldnn/cpu_quantize_squash_pass_tester.cc
+56
-0
未找到文件。
paddle/fluid/framework/ir/graph_pattern_detector.cc
浏览文件 @
2e3ec66b
...
...
@@ -1267,6 +1267,24 @@ PDNode *patterns::ConvRequant::operator()() {
return
requant_out
;
}
PDNode
*
patterns
::
ConvDequant
::
operator
()()
{
// Create Operators
auto
conv_op
=
pattern
->
NewNode
(
conv_op_repr
())
->
assert_is_op
(
"conv2d"
);
auto
dequant_op
=
pattern
->
NewNode
(
dequant_op_repr
())
->
assert_is_op
(
"dequantize"
);
auto
conv_out
=
pattern
->
NewNode
(
conv_out_repr
())
->
assert_is_op_output
(
"conv2d"
,
"Output"
);
auto
dequant_out
=
pattern
->
NewNode
(
dequant_out_repr
())
->
AsOutput
()
->
assert_is_op_output
(
"dequantize"
,
"Output"
);
conv_op
->
LinksTo
({
conv_out
});
dequant_op
->
LinksFrom
({
conv_out
}).
LinksTo
({
dequant_out
});
return
dequant_out
;
}
PDNode
*
patterns
::
PriorBox
::
operator
()()
{
auto
prior_box_op
=
pattern
->
NewNode
(
prior_box_op_repr
())
->
assert_is_op
(
"prior_box"
);
...
...
paddle/fluid/framework/ir/graph_pattern_detector.h
浏览文件 @
2e3ec66b
...
...
@@ -793,6 +793,23 @@ struct ConvRequant : public PatternBase {
PATTERN_DECL_NODE
(
requant_out
);
};
// Conv + Dequant
// named nodes:
// conv_op, conv_out
// dequant_op, dequant_out
struct
ConvDequant
:
public
PatternBase
{
ConvDequant
(
PDPattern
*
pattern
,
const
std
::
string
&
name_scope
)
:
PatternBase
(
pattern
,
name_scope
,
"conv_dequant"
)
{}
PDNode
*
operator
()();
PATTERN_DECL_NODE
(
conv_op
);
PATTERN_DECL_NODE
(
conv_out
);
PATTERN_DECL_NODE
(
dequant_op
);
PATTERN_DECL_NODE
(
dequant_out
);
};
// PriorBox operator
// operator: prior_box_op
// inputs: prior_box_input, prior_box_image
...
...
paddle/fluid/framework/ir/mkldnn/cpu_quantize_squash_pass.cc
浏览文件 @
2e3ec66b
...
...
@@ -160,6 +160,38 @@ void CPUQuantizeSquashPass::ConvRequantSquash(Graph* graph) const {
found_requant_squash_count
);
}
void
CPUQuantizeSquashPass
::
ConvDequantSquash
(
Graph
*
graph
)
const
{
GraphPatternDetector
gpd
;
patterns
::
ConvDequant
conv_dequant_pattern
{
gpd
.
mutable_pattern
(),
"conv_dequant"
};
conv_dequant_pattern
();
int
found_conv_dequant_squash_count
=
0
;
auto
handler
=
[
&
](
const
GraphPatternDetector
::
subgraph_t
&
subgraph
,
Graph
*
g
)
{
VLOG
(
4
)
<<
"squash conv-dequant ops pair"
;
GET_IR_NODE_FROM_SUBGRAPH
(
conv_op
,
conv_op
,
conv_dequant_pattern
);
GET_IR_NODE_FROM_SUBGRAPH
(
conv_out
,
conv_out
,
conv_dequant_pattern
);
GET_IR_NODE_FROM_SUBGRAPH
(
dequant_op
,
dequant_op
,
conv_dequant_pattern
);
GET_IR_NODE_FROM_SUBGRAPH
(
dequant_out
,
dequant_out
,
conv_dequant_pattern
);
// if conv2d has one output
if
(
conv_out
->
outputs
.
size
()
==
1
)
{
conv_op
->
Op
()
->
SetAttr
(
"force_fp32_output"
,
true
);
conv_op
->
Op
()
->
SetOutput
(
"Output"
,
std
::
vector
<
std
::
string
>
({
dequant_out
->
Name
()}));
IR_NODE_LINK_TO
(
conv_op
,
dequant_out
);
GraphSafeRemoveNodes
(
graph
,
{
conv_out
,
dequant_op
});
found_conv_dequant_squash_count
++
;
}
};
gpd
(
graph
,
handler
);
AddStatis
(
found_conv_dequant_squash_count
);
PrettyLogDetail
(
"--- squashed %d dequant with convs"
,
found_conv_dequant_squash_count
);
}
void
CPUQuantizeSquashPass
::
ApplyImpl
(
ir
::
Graph
*
graph
)
const
{
PADDLE_ENFORCE
(
graph
);
FusePassBase
::
Init
(
"cpu_quantize_squash_pass"
,
graph
);
...
...
@@ -168,6 +200,7 @@ void CPUQuantizeSquashPass::ApplyImpl(ir::Graph* graph) const {
FindNodesToKeep
(
graph
,
&
nodes_keep_counter
);
DequantQuantSquash
(
graph
,
&
nodes_keep_counter
);
ConvRequantSquash
(
graph
);
ConvDequantSquash
(
graph
);
}
}
// namespace ir
...
...
paddle/fluid/framework/ir/mkldnn/cpu_quantize_squash_pass.h
浏览文件 @
2e3ec66b
...
...
@@ -55,6 +55,11 @@ class CPUQuantizeSquashPass : public FusePassBase {
*/
void
ConvRequantSquash
(
Graph
*
graph
)
const
;
/*
* Squash conv2d with dequant when dequant is the only op after conv2d
*/
void
ConvDequantSquash
(
Graph
*
graph
)
const
;
const
std
::
string
name_scope_
{
"squash"
};
};
...
...
paddle/fluid/framework/ir/mkldnn/cpu_quantize_squash_pass_tester.cc
浏览文件 @
2e3ec66b
...
...
@@ -161,6 +161,36 @@ ProgramDesc BuildConvMultiRequantProgramDesc(bool use_mkldnn, float scale_out,
return
prog
;
}
// a->Conv1->b
// b->Dequant1(Scale1)->c
// c->Concat
ProgramDesc
BuildConvDequantConcatProgramDesc
(
bool
use_mkldnn
,
float
scale_out
,
float
scale
)
{
ProgramDesc
prog
;
for
(
auto
&
v
:
variable_names
)
{
prog
.
MutableBlock
(
0
)
->
Var
(
v
);
}
SetOp
(
&
prog
,
"conv2d"
,
"Conv1"
,
{
"a"
},
{
"b"
},
use_mkldnn
,
scale_out
);
SetOp
(
&
prog
,
"dequantize"
,
"Dequant1"
,
{
"b"
},
{
"c"
},
use_mkldnn
,
scale
);
SetOp
(
&
prog
,
"concat"
,
"Concat1"
,
{
"c"
},
{
"d"
},
use_mkldnn
);
return
prog
;
}
// a->Conv1->b
// b->Dequant1(Scale1)->c
// b->Conv2->d
ProgramDesc
BuildConvDequantConvProgramDesc
(
bool
use_mkldnn
,
float
scale_out
,
float
scale
)
{
ProgramDesc
prog
;
for
(
auto
&
v
:
variable_names
)
{
prog
.
MutableBlock
(
0
)
->
Var
(
v
);
}
SetOp
(
&
prog
,
"conv2d"
,
"Conv1"
,
{
"a"
},
{
"b"
},
use_mkldnn
,
scale_out
);
SetOp
(
&
prog
,
"dequantize"
,
"Dequant1"
,
{
"b"
},
{
"c"
},
use_mkldnn
,
scale
);
SetOp
(
&
prog
,
"conv2d"
,
"Conv2"
,
{
"b"
},
{
"d"
},
use_mkldnn
);
return
prog
;
}
void
InitTensorHolder
(
Scope
*
scope
,
const
paddle
::
platform
::
Place
&
place
,
const
char
*
var_name
)
{
auto
x
=
scope
->
Var
(
var_name
);
...
...
@@ -217,6 +247,7 @@ void EqualScaleOutTest(const ProgramDesc& prog, const std::string& name,
void
CheckRequantScalesTest
(
const
ProgramDesc
&
prog
,
float
scale_in
,
float
scale_out
)
{
std
::
unique_ptr
<
ir
::
Graph
>
graph
(
new
ir
::
Graph
(
prog
));
PrepareGraph
(
&
graph
,
prog
);
RegisterPass
(
&
graph
);
...
...
@@ -238,6 +269,7 @@ TEST(CpuQuantizeSquashPass, equal_scales) {
auto
use_mkldnn
=
true
;
// Remove 4 nodes: Dequant, Quant, e, f
auto
remove_nodes
=
4
;
CountNodeTest
(
BuildConvRequantProgramDesc
(
use_mkldnn
,
scale_out
,
scale
,
scale
),
remove_nodes
);
...
...
@@ -253,6 +285,7 @@ TEST(CpuQuantizeSquashPass, unequal_scales) {
auto
use_mkldnn
=
true
;
// Remove 4 nodes: Dequant, Quant, e, d
auto
remove_nodes
=
4
;
CountNodeTest
(
BuildConvRequantProgramDesc
(
use_mkldnn
,
scale_out
,
scale1
,
scale2
),
remove_nodes
);
...
...
@@ -280,6 +313,7 @@ TEST(CpuQuantizeSquashPass, branch_to_equal_unequal_and_fp32) {
// Remove 3 nodes: Quant1, c, Quant2,
// Insert 1 node: Requant
auto
remove_nodes
=
2
;
CountNodeTest
(
BuildConvMultiOutputProgramDesc
(
use_mkldnn
,
scale_out
,
scale
,
scale
,
scale2
),
remove_nodes
);
...
...
@@ -322,6 +356,7 @@ TEST(CpuQuantizeSquashPass,
// Remove 3 nodes: Dequant1, c, Quant
// Insert 1 node: Requant
auto
remove_nodes
=
2
;
CountNodeTest
(
BuildConcatDequantQuantProgramDesc
(
use_mkldnn
,
scale_out
,
scale
,
scale2
),
remove_nodes
);
...
...
@@ -345,6 +380,27 @@ TEST(CpuQuantizeSquashPass, more_than_one_conv_out_outputs) {
remove_nodes
);
}
// a->Conv1->c->Concat
TEST
(
CpuQuantizeSquashPass
,
conv_dequant_only_one_output
)
{
auto
scale_out
=
1.0
f
;
auto
scale
=
1.2345
f
;
auto
use_mkldnn
=
true
;
// remove 2 nodes: Dequant1, c
auto
remove_nodes
=
2
;
CountNodeTest
(
BuildConvDequantConcatProgramDesc
(
use_mkldnn
,
scale_out
,
scale
),
remove_nodes
);
}
TEST
(
CpuQuantizeSquashPass
,
conv_dequant_more_than_one_op_after_conv
)
{
auto
scale_out
=
1.0
f
;
auto
scale
=
1.2345
f
;
auto
use_mkldnn
=
true
;
// nothing change
auto
remove_nodes
=
0
;
CountNodeTest
(
BuildConvDequantConvProgramDesc
(
use_mkldnn
,
scale_out
,
scale
),
remove_nodes
);
}
}
// namespace ir
}
// namespace framework
}
// namespace paddle
...
...
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