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
s920243400
PaddleDetection
提交
d2ff3e49
P
PaddleDetection
项目概览
s920243400
/
PaddleDetection
与 Fork 源项目一致
Fork自
PaddlePaddle / PaddleDetection
通知
2
Star
0
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
0
列表
看板
标记
里程碑
合并请求
0
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
0
Issue
0
列表
看板
标记
里程碑
合并请求
0
合并请求
0
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
d2ff3e49
编写于
11月 15, 2016
作者:
L
LCY-Seso
提交者:
GitHub
11月 15, 2016
浏览文件
操作
浏览文件
下载
差异文件
Merge pull request #458 from luotao1/group
add layer check for recurrent_group
上级
c2aa8980
8c594376
变更
1
隐藏空白更改
内联
并排
Showing
1 changed file
with
22 addition
and
4 deletion
+22
-4
python/paddle/trainer_config_helpers/layers.py
python/paddle/trainer_config_helpers/layers.py
+22
-4
未找到文件。
python/paddle/trainer_config_helpers/layers.py
浏览文件 @
d2ff3e49
...
@@ -494,8 +494,7 @@ def scaling_projection(input, param_attr=None):
...
@@ -494,8 +494,7 @@ def scaling_projection(input, param_attr=None):
:return: A ScalingProjection object
:return: A ScalingProjection object
:rtype: ScalingProjection
:rtype: ScalingProjection
"""
"""
proj
=
ScalingProjection
(
input_layer_name
=
input
.
name
,
proj
=
ScalingProjection
(
input_layer_name
=
input
.
name
,
**
param_attr
.
attr
)
**
param_attr
.
attr
)
proj
.
origin
=
input
proj
.
origin
=
input
return
proj
return
proj
...
@@ -2783,7 +2782,12 @@ class SubsequenceInput(object):
...
@@ -2783,7 +2782,12 @@ class SubsequenceInput(object):
@
wrap_name_default
(
"recurrent_group"
)
@
wrap_name_default
(
"recurrent_group"
)
def
recurrent_group
(
step
,
input
,
reverse
=
False
,
name
=
None
,
targetInlink
=
None
):
def
recurrent_group
(
step
,
input
,
reverse
=
False
,
name
=
None
,
targetInlink
=
None
,
is_generating
=
False
):
"""
"""
Recurrent layer group is an extremely flexible recurrent unit in
Recurrent layer group is an extremely flexible recurrent unit in
PaddlePaddle. As long as the user defines the calculation done within a
PaddlePaddle. As long as the user defines the calculation done within a
...
@@ -2848,6 +2852,12 @@ def recurrent_group(step, input, reverse=False, name=None, targetInlink=None):
...
@@ -2848,6 +2852,12 @@ def recurrent_group(step, input, reverse=False, name=None, targetInlink=None):
:type targetInlink: LayerOutput|SubsequenceInput
:type targetInlink: LayerOutput|SubsequenceInput
:param is_generating: If is generating, none of input type should be LayerOutput;
else, for training or testing, one of the input type must
be LayerOutput.
: type is_generating: bool
:return: LayerOutput object.
:return: LayerOutput object.
:rtype: LayerOutput
:rtype: LayerOutput
"""
"""
...
@@ -2895,6 +2905,7 @@ def recurrent_group(step, input, reverse=False, name=None, targetInlink=None):
...
@@ -2895,6 +2905,7 @@ def recurrent_group(step, input, reverse=False, name=None, targetInlink=None):
seq_reversed
=
reverse
,
seq_reversed
=
reverse
,
target_inlinkname
=
targetInlinkName
)
target_inlinkname
=
targetInlinkName
)
in_args
=
[]
in_args
=
[]
has_LayerOutput
=
True
for
each_input
in
input
:
for
each_input
in
input
:
assert
is_single_input
(
each_input
)
assert
is_single_input
(
each_input
)
if
isinstance
(
each_input
,
LayerOutput
):
if
isinstance
(
each_input
,
LayerOutput
):
...
@@ -2902,6 +2913,7 @@ def recurrent_group(step, input, reverse=False, name=None, targetInlink=None):
...
@@ -2902,6 +2913,7 @@ def recurrent_group(step, input, reverse=False, name=None, targetInlink=None):
elif
isinstance
(
each_input
,
SubsequenceInput
):
elif
isinstance
(
each_input
,
SubsequenceInput
):
in_args
.
append
(
each_input
.
input
)
in_args
.
append
(
each_input
.
input
)
else
:
else
:
has_LayerOutput
=
False
mem_name
=
"__%s_memory__"
%
each_input
.
input
.
name
mem_name
=
"__%s_memory__"
%
each_input
.
input
.
name
mem
=
memory
(
mem
=
memory
(
name
=
mem_name
,
name
=
mem_name
,
...
@@ -2915,6 +2927,8 @@ def recurrent_group(step, input, reverse=False, name=None, targetInlink=None):
...
@@ -2915,6 +2927,8 @@ def recurrent_group(step, input, reverse=False, name=None, targetInlink=None):
mix
+=
identity_projection
(
mem
)
mix
+=
identity_projection
(
mem
)
in_args
.
append
(
mem
)
in_args
.
append
(
mem
)
assert
(
is_generating
!=
has_LayerOutput
)
layer_outs
=
step
(
*
in_args
)
layer_outs
=
step
(
*
in_args
)
if
isinstance
(
layer_outs
,
LayerOutput
):
if
isinstance
(
layer_outs
,
LayerOutput
):
...
@@ -3206,7 +3220,11 @@ def beam_search(step,
...
@@ -3206,7 +3220,11 @@ def beam_search(step,
return
predict
return
predict
tmp
=
recurrent_group
(
tmp
=
recurrent_group
(
step
=
__real_step__
,
input
=
real_input
,
reverse
=
False
,
name
=
name
)
step
=
__real_step__
,
input
=
real_input
,
reverse
=
False
,
name
=
name
,
is_generating
=
True
)
return
tmp
return
tmp
...
...
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