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
机器未来
Paddle
提交
d4fcd2a5
P
Paddle
项目概览
机器未来
/
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
d4fcd2a5
编写于
12月 01, 2017
作者:
D
dangqingqing
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Fix the doc of LSTM operator.
上级
44e39144
变更
1
隐藏空白更改
内联
并排
Showing
1 changed file
with
13 addition
and
13 deletion
+13
-13
paddle/operators/lstm_op.cc
paddle/operators/lstm_op.cc
+13
-13
未找到文件。
paddle/operators/lstm_op.cc
浏览文件 @
d4fcd2a5
...
...
@@ -181,7 +181,7 @@ class LSTMOpMaker : public framework::OpProtoAndCheckerMaker {
AddComment
(
R"DOC(
Long-Short Term Memory (LSTM) Operator.
The defalut implementation is diagonal/peephole connection
The defalut implementation is diagonal/peephole connection
(https://arxiv.org/pdf/1402.1128.pdf), the formula is as follows:
$$
...
...
@@ -198,27 +198,27 @@ c_t = f_t \odot c_{t-1} + i_t \odot \tilde{c_t} \\
h_t = o_t \odot act_h(c_t)
$$
where the W terms denote weight matrices (e.g.
\f$W_{xi}\f
$ is the matrix
of weights from the input gate to the input),
\f$W_{ic}, W_{fc}, W_{oc}\f
$
where the W terms denote weight matrices (e.g.
$W_{xi}
$ is the matrix
of weights from the input gate to the input),
$W_{ic}, W_{fc}, W_{oc}
$
are diagonal weight matrices for peephole connections. In our implementation,
we use vectors to reprenset these diagonal weight matrices. The b terms
denote bias vectors (
\f$b_i\f$ is the input gate bias vector), \f$\sigma\f
$
denote bias vectors (
$b_i$ is the input gate bias vector), $\sigma
$
is the non-line activations, such as logistic sigmoid function, and
\f$i, f, o\f$ and \f$c\f
$ are the input gate, forget gate, output gate,
$i, f, o$ and $c
$ are the input gate, forget gate, output gate,
and cell activation vectors, respectively, all of which have the same size as
the cell output activation vector
\f$h\f
$.
the cell output activation vector
$h
$.
The
\f$\odot\f$ is the element-wise product of the vectors. \f$act_g\f$ and \f$act_h\f
$
The
$\odot$ is the element-wise product of the vectors. $act_g$ and $act_h
$
are the cell input and cell output activation functions and `tanh` is usually
used for them.
\f$\tilde{c_t}\f
$ is also called candidate hidden state,
used for them.
$\tilde{c_t}
$ is also called candidate hidden state,
which is computed based on the current input and the previous hidden state.
Set `use_peepholes` False to disable peephole connection
(http://www.bioinf.jku.at/publications/older/2604.pdf). The formula
is omitted here
.
Set `use_peepholes` False to disable peephole connection
. The formula
is omitted here, please refer to the paper
http://www.bioinf.jku.at/publications/older/2604.pdf for details
.
Note that these
\f$W_{xi}x_{t}, W_{xf}x_{t}, W_{xc}x_{t}, W_{xo}x_{t}\f
$
operations on the input
\f$x_{t}\f
$ are NOT included in this operator.
Note that these
$W_{xi}x_{t}, W_{xf}x_{t}, W_{xc}x_{t}, W_{xo}x_{t}
$
operations on the input
$x_{t}
$ are NOT included in this operator.
Users can choose to use fully-connect operator before LSTM operator.
)DOC"
);
...
...
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