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
夏有凉风,冬有雪
Paddle
提交
66ae0a8c
P
Paddle
项目概览
夏有凉风,冬有雪
/
Paddle
与 Fork 源项目一致
Fork自
PaddlePaddle / Paddle
通知
1
Star
1
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
1
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
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
Issue
1
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
体验新版 GitCode,发现更多精彩内容 >>
提交
66ae0a8c
编写于
1月 03, 2018
作者:
Y
yangyaming
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Enhence shrink_rnn_memory_op.
上级
25af35d8
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
20 addition
and
9 deletion
+20
-9
paddle/operators/shrink_rnn_memory_op.cc
paddle/operators/shrink_rnn_memory_op.cc
+16
-5
python/paddle/v2/fluid/tests/test_shrink_rnn_memory.py
python/paddle/v2/fluid/tests/test_shrink_rnn_memory.py
+4
-4
未找到文件。
paddle/operators/shrink_rnn_memory_op.cc
浏览文件 @
66ae0a8c
...
...
@@ -46,8 +46,19 @@ class ShrinkRNNMemoryOp : public ArrayOp {
auto
*
out_var
=
scope
.
FindVar
(
Output
(
"Out"
));
PADDLE_ENFORCE
(
out_var
!=
nullptr
,
"Output Out must be set"
);
auto
&
out_tensor
=
*
out_var
->
GetMutable
<
framework
::
LoDTensor
>
();
// should consider multiple levels
size_t
height
=
dst_num_rows
;
auto
lod_level
=
lod_rank_table
.
level
();
if
(
x_tensor
.
lod
().
size
()
>
lod_level
&&
x_tensor
.
lod
()[
lod_level
].
size
()
<
dst_num_rows
)
{
auto
lod_offset
=
framework
::
GetSubLoDAndAbsoluteOffset
(
x_tensor
.
lod
(),
0
,
dst_num_rows
+
1
,
lod_level
);
height
=
lod_offset
.
second
.
second
;
}
if
(
dst_num_rows
!=
0
)
{
out_tensor
.
ShareDataWith
(
x_tensor
.
Slice
(
0
,
dst_num_rows
));
out_tensor
.
ShareDataWith
(
x_tensor
.
Slice
(
0
,
height
));
}
}
};
...
...
@@ -64,11 +75,11 @@ class ShrinkRNNMemoryOpProtoMaker : public framework::OpProtoAndCheckerMaker {
AddOutput
(
"Out"
,
"(LoDTensor) The shrinked RNN step memory."
);
AddComment
(
R"DOC(
In dynamic RNN, we are able to handle sequences of different lengths.
Because of the multiple lengths, the size of each step input can be
In dynamic RNN, we are able to handle sequences of different lengths.
Because of the multiple lengths, the size of each step input can be
different, which may lead to a mismatching between the input of
the current step and the memory generated by the previous one. This
operator shrinks memory according to the size of the next step input,
the current step and the memory generated by the previous one. This
operator shrinks memory according to the size of the next step input,
to make sure that they can match each other.
)DOC"
);
}
...
...
python/paddle/v2/fluid/tests/test_shrink_rnn_memory.py
浏览文件 @
66ae0a8c
...
...
@@ -26,13 +26,13 @@ class TestShrinkRNNMemory(unittest.TestCase):
cpu
=
core
.
CPUPlace
()
tensor
=
core
.
LoDTensor
()
tensor
.
set_lod
([[
0
,
2
,
5
,
6
]])
tensor_np
=
numpy
.
random
.
random
(
size
=
(
3
,
100
)).
astype
(
'float32'
)
tensor_np
=
numpy
.
random
.
random
(
size
=
(
6
,
100
)).
astype
(
'float32'
)
tensor
.
set
(
tensor_np
,
cpu
)
exe
=
Executor
(
cpu
)
outs
=
exe
.
run
(
feed
=
{
'x'
:
tensor
},
fetch_list
=
[
mem1
,
mem2
,
mem3
])
self
.
assertTrue
(
numpy
.
allclose
(
tensor_np
[
0
:
3
],
outs
[
0
]))
self
.
assertTrue
(
numpy
.
allclose
(
tensor_np
[
0
:
2
],
outs
[
1
]))
self
.
assertTrue
(
numpy
.
allclose
(
tensor_np
[
0
:
1
],
outs
[
2
]))
self
.
assertTrue
(
numpy
.
allclose
(
tensor_np
[
0
:
6
],
outs
[
0
]))
self
.
assertTrue
(
numpy
.
allclose
(
tensor_np
[
0
:
5
],
outs
[
1
]))
self
.
assertTrue
(
numpy
.
allclose
(
tensor_np
[
0
:
2
],
outs
[
2
]))
mem3_mean
=
layers
.
mean
(
x
=
mem3
)
append_backward
(
loss
=
mem3_mean
)
...
...
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