Skip to content
项目
群组
代码片段
帮助
当前项目
正在载入...
登录 / 注册
切换导航面板
M
mtbookv2
概览
Overview
Details
Activity
Cycle Analytics
版本库
Repository
Files
Commits
Branches
Tags
Contributors
Graph
Compare
Charts
问题
0
Issues
0
列表
Board
标记
里程碑
合并请求
0
Merge Requests
0
CI / CD
CI / CD
流水线
作业
日程表
图表
维基
Wiki
代码片段
Snippets
成员
Collapse sidebar
Close sidebar
活动
图像
聊天
创建新问题
作业
提交
Issue Boards
Open sidebar
NiuTrans
mtbookv2
Commits
d20babe2
Commit
d20babe2
authored
Jan 25, 2021
by
单韦乔
Browse files
Options
Browse Files
Download
Email Patches
Plain Diff
13章图以及15章文字
parent
78e95635
隐藏空白字符变更
内嵌
并排
正在显示
6 个修改的文件
包含
29 行增加
和
28 行删除
+29
-28
Chapter13/Figures/figure-computation-of-dropout.tex
+6
-5
Chapter13/Figures/figure-difference-between-word-level-and-sequence-level-in-knowledge-distillation.tex
+10
-10
Chapter13/Figures/figure-ensemble-knowledge-distillation.tex
+9
-9
Chapter13/Figures/figure-label-smoothing.tex
+2
-2
Chapter15/Figures/figure-attention-distribution-based-on-gaussian-distribution.tex
+1
-1
Chapter15/chapter15.tex
+1
-1
没有找到文件。
Chapter13/Figures/figure-computation-of-dropout.tex
查看文件 @
d20babe2
...
...
@@ -15,7 +15,7 @@
\node
[neuronnode] (neuron
_
y') at (2.4 *
\nodespace
,-1.5 *
\neuronsep
)
{
\scriptsize
{$
x
_{
i
}^{
l
+
1
}$}}
;
\node
[anchor=north] (standard) at ([yshift=-4em]neuron
_
z.south)
{
\scriptsize
{
标准网络
}}
;
\node
[] (standard) at ([xshift=-1em]neuron
_
z.west)
{
\scriptsize
{$
\mathb
f
{
w
}_{
i
}^{
l
}$}}
;
\node
[] (standard) at ([xshift=-1em]neuron
_
z.west)
{
\scriptsize
{$
\mathb
i
{
w
}_{
i
}^{
l
}$}}
;
\node
[] (standard) at ([xshift=0.6em,yshift=0.3em]neuron
_
z.east)
{
\scriptsize
{$
f
$}}
;
\draw
[->,line width=0.3mm] (neuron
_
b.east) -- (neuron
_
z.130);
...
...
@@ -41,7 +41,7 @@
\node
[neuronnode] (drop
_
neuron
_
r1) at (4.4*
\nodespace
,-2.5*
\neuronsep
)
{
\scriptsize
{$
r
_{
1
}^{
l
}$}}
;
\node
[anchor=north] (standard) at ([xshift=2em,yshift=-4em]drop
_
neuron
_
z.south)
{
\scriptsize
{
应用Dropout后的网络
}}
;
\node
[] (standard) at ([xshift=-1em]drop
_
neuron
_
z.west)
{
\scriptsize
{$
\mathb
f
{
w
}_{
i
}^{
l
}$}}
;
\node
[] (standard) at ([xshift=-1em]drop
_
neuron
_
z.west)
{
\scriptsize
{$
\mathb
i
{
w
}_{
i
}^{
l
}$}}
;
\node
[] (standard) at ([xshift=0.6em,yshift=0.3em]drop
_
neuron
_
z.east)
{
\scriptsize
{$
f
$}}
;
%structure
\draw
[->,line width=0.3mm] (drop
_
neuron
_
b.east) -- (drop
_
neuron
_
z.130);
...
...
@@ -60,12 +60,12 @@
%equ
\node
[anchor=west,inner sep = 2pt] (line1) at (9*
\nodespace
,0)
{
未应用Dropout:
}
;
\node
[anchor=north west,inner sep = 2pt] (line2) at (line1.south west)
{$
z
_{
i
}^{
l
+
1
}
=
\mathb
f
{
w
}^{
l
}
\mathbf
{
x
}^{
l
}
+
b
^{
l
}$}
;
\node
[anchor=north west,inner sep = 2pt] (line2) at (line1.south west)
{$
z
_{
i
}^{
l
+
1
}
=
\mathb
i
{
w
}^{
l
}
\mathbi
{
x
}^{
l
}
+
b
^{
l
}$}
;
\node
[anchor=north west,inner sep = 2pt] (line3) at (line2.south west)
{$
x
_{
i
}^{
l
+
1
}
=
f
\left
(
z
_{
i
}^{
l
+
1
}
\right
)
$}
;
\node
[anchor=north west,inner sep = 2pt] (line4) at (line3.south west)
{
应用Dropout:
}
;
\node
[anchor=north west,inner sep = 2pt] (line5) at (line4.south west)
{$
r
_{
j
}^{
l
}
\sim
$
Bernoulli
$
(
1
-
p
)
$}
;
\node
[anchor=north west,inner sep = 2pt] (line6) at (line5.south west)
{$
\tilde
{
\mathb
f
{
x
}}
=
\mathbf
{
r
}
*
\mathbf
{
x
}$}
;
\node
[anchor=north west,inner sep = 2pt] (line7) at (line6.south west)
{$
z
_{
i
}^{
l
+
1
}
=
\mathb
f
{
w
}^{
l
}
\widetilde
{
\mathbf
{
x
}}^{
l
}
+
b
^{
l
}$}
;
\node
[anchor=north west,inner sep = 2pt] (line6) at (line5.south west)
{$
\tilde
{
\mathb
i
{
x
}}
=
\mathbi
{
r
}
*
\mathbi
{
x
}$}
;
\node
[anchor=north west,inner sep = 2pt] (line7) at (line6.south west)
{$
z
_{
i
}^{
l
+
1
}
=
\mathb
i
{
w
}^{
l
}
\widetilde
{
\mathbi
{
x
}}^{
l
}
+
b
^{
l
}$}
;
\node
[anchor=north west,inner sep = 2pt] (line8) at (line7.south west)
{$
x
_{
i
}^{
l
+
1
}
=
f
\left
(
z
_{
i
}^{
l
+
1
}
\right
)
$}
;
\end{tikzpicture}
\ No newline at end of file
Chapter13/Figures/figure-difference-between-word-level-and-sequence-level-in-knowledge-distillation.tex
查看文件 @
d20babe2
...
...
@@ -4,10 +4,10 @@
\begin{scope}
[]
% Column 1
\node
[prob,minimum size=0.1cm] (prob11) at (0,0)
{}
;
\node
[prob,minimum size=0.5cm,anchor=center] (prob21) at ([yshift=-0.5cm]prob11.center)
{$
.
7
$}
;
\node
[prob,minimum size=0.5cm,anchor=center] (prob21) at ([yshift=-0.5cm]prob11.center)
{$
0
.
7
$}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob31) at ([yshift=-0.5cm]prob21.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob41) at ([yshift=-0.5cm]prob31.center)
{}
;
\node
[prob,minimum size=0.3cm,anchor=center
] (prob51) at ([yshift=-0.5cm]prob41.center)
{$
.
2
$}
;
\node
[prob,minimum size=0.3cm,anchor=center
,font=
\tiny
] (prob51) at ([yshift=-0.5cm]prob41.center)
{$
0
.
2
$}
;
\begin{pgfonlayer}
{
background
}
\coordinate
(bottomleft) at ([shift=
{
(-0.25cm,-0.25cm)
}
]prob51.center);
\coordinate
(topright) at ([shift=
{
(0.25cm,0.25cm)
}
]prob11.center);
...
...
@@ -18,8 +18,8 @@
% Column 2
\node
[prob,minimum size=0.1cm,anchor=center] (prob12) at ([xshift=1cm]prob11.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob22) at ([yshift=-0.5cm]prob12.center)
{}
;
\node
[prob,minimum size=0.4cm,anchor=center
] (prob32) at ([yshift=-0.5cm]prob22.center)
{
$
.
4
$}
;
\node
[prob,minimum size=0.3cm,anchor=center
] (prob42) at ([yshift=-0.5cm]prob32.center)
{$
.
3
$}
;
\node
[prob,minimum size=0.4cm,anchor=center
,font=
\tiny
] (prob32) at ([yshift=-0.5cm]prob22.center)
{
0
$
.
4
$}
;
\node
[prob,minimum size=0.3cm,anchor=center
,font=
\tiny
] (prob42) at ([yshift=-0.5cm]prob32.center)
{$
0
.
3
$}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob52) at ([yshift=-0.5cm]prob42.center)
{}
;
\begin{pgfonlayer}
{
background
}
\coordinate
(bottomleft) at ([shift=
{
(-0.25cm,-0.25cm)
}
]prob52.center);
...
...
@@ -33,7 +33,7 @@
\node
[prob,minimum size=0.1cm,anchor=center] (prob23) at ([yshift=-0.5cm]prob13.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob33) at ([yshift=-0.5cm]prob23.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob43) at ([yshift=-0.5cm]prob33.center)
{}
;
\node
[prob,minimum size=0.4cm,anchor=center
] (prob53) at ([yshift=-0.5cm]prob43.center)
{$
.
6
$}
;
\node
[prob,minimum size=0.4cm,anchor=center
,font=
\tiny
] (prob53) at ([yshift=-0.5cm]prob43.center)
{$
0
.
6
$}
;
\begin{pgfonlayer}
{
background
}
\coordinate
(bottomleft) at ([shift=
{
(-0.25cm,-0.25cm)
}
]prob53.center);
\coordinate
(topright) at ([shift=
{
(0.25cm,0.25cm)
}
]prob13.center);
...
...
@@ -42,7 +42,7 @@
% \node [anchor=center] (word13) at ([yshift=0.7cm]prob13.center) {fine};
% Column 4
\node
[prob,minimum size=0.5cm,anchor=center] (prob14) at ([xshift=1cm]prob13.center)
{$
.
8
$}
;
\node
[prob,minimum size=0.5cm,anchor=center] (prob14) at ([xshift=1cm]prob13.center)
{$
0
.
8
$}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob24) at ([yshift=-0.5cm]prob14.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob34) at ([yshift=-0.5cm]prob24.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob44) at ([yshift=-0.5cm]prob34.center)
{}
;
...
...
@@ -92,7 +92,7 @@
\begin{scope}
[xshift=2.7in]
% Column 1
\node
[prob,minimum size=0.1cm] (prob11) at (0,0)
{}
;
\node
[prob,minimum size=0.5cm,anchor=center] (prob21) at ([yshift=-0.5cm]prob11.center)
{$
1
.
$}
;
\node
[prob,minimum size=0.5cm,anchor=center] (prob21) at ([yshift=-0.5cm]prob11.center)
{$
1
$}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob31) at ([yshift=-0.5cm]prob21.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob41) at ([yshift=-0.5cm]prob31.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob51) at ([yshift=-0.5cm]prob41.center)
{}
;
...
...
@@ -106,7 +106,7 @@
% Column 2
\node
[prob,minimum size=0.1cm,anchor=center] (prob12) at ([xshift=1cm]prob11.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob22) at ([yshift=-0.5cm]prob12.center)
{}
;
\node
[prob,minimum size=0.5cm,anchor=center] (prob32) at ([yshift=-0.5cm]prob22.center)
{$
1
.
$}
;
\node
[prob,minimum size=0.5cm,anchor=center] (prob32) at ([yshift=-0.5cm]prob22.center)
{$
1
$}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob42) at ([yshift=-0.5cm]prob32.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob52) at ([yshift=-0.5cm]prob42.center)
{}
;
\begin{pgfonlayer}
{
background
}
...
...
@@ -121,7 +121,7 @@
\node
[prob,minimum size=0.1cm,anchor=center] (prob23) at ([yshift=-0.5cm]prob13.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob33) at ([yshift=-0.5cm]prob23.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob43) at ([yshift=-0.5cm]prob33.center)
{}
;
\node
[prob,minimum size=0.5cm,anchor=center] (prob53) at ([yshift=-0.5cm]prob43.center)
{$
1
.
$}
;
\node
[prob,minimum size=0.5cm,anchor=center] (prob53) at ([yshift=-0.5cm]prob43.center)
{$
1
$}
;
\begin{pgfonlayer}
{
background
}
\coordinate
(bottomleft) at ([shift=
{
(-0.25cm,-0.25cm)
}
]prob53.center);
\coordinate
(topright) at ([shift=
{
(0.25cm,0.25cm)
}
]prob13.center);
...
...
@@ -130,7 +130,7 @@
\node
[anchor=center] (word13) at ([yshift=0.68cm]prob13.center)
{
good
}
;
% Column 4
\node
[prob,minimum size=0.5cm,anchor=center] (prob14) at ([xshift=1cm]prob13.center)
{$
1
.
$}
;
\node
[prob,minimum size=0.5cm,anchor=center] (prob14) at ([xshift=1cm]prob13.center)
{$
1
$}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob24) at ([yshift=-0.5cm]prob14.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob34) at ([yshift=-0.5cm]prob24.center)
{}
;
\node
[prob,minimum size=0.1cm,anchor=center] (prob44) at ([yshift=-0.5cm]prob34.center)
{}
;
...
...
Chapter13/Figures/figure-ensemble-knowledge-distillation.tex
查看文件 @
d20babe2
...
...
@@ -20,21 +20,21 @@
\foreach
\curr
/
\prev
in
{
1/0,2/1,3/2
}
{
% models
\node
[modelnode,fill=yellow!20
]
(stu
\curr
1) at ([yshift=-3em]stu
\prev
1.south)
{
\rotatebox
{
90
}{
学生模型
$
1
$}
}
;
\node
[modelnode,fill=yellow!20
]
(stu
\curr
2) at ([yshift=-3em]stu
\prev
2.south)
{
\rotatebox
{
90
}{
学生模型
$
2
$}
}
;
\node
[modelnode,fill=yellow!20
]
(stu
\curr
3) at ([yshift=-3em]stu
\prev
3.south)
{
\rotatebox
{
90
}{
学生模型
$
3
$}
}
;
\node
[modelnode,fill=yellow!20
]
(stu
\curr
4) at ([yshift=-3em]stu
\prev
4.south)
{
\rotatebox
{
90
}{
学生模型
$
4
$}
}
;
\node
[modelnode,fill=yellow!20
]
(stu
\curr
5) at ([yshift=-3em]stu
\prev
5.south)
{
\rotatebox
{
90
}{
学生模型
$
5
$}
}
;
\node
[modelnode
]
(tea
\curr
1) at ([yshift=-3em]tea
\prev
1.south)
{
\rotatebox
{
90
}{
\color
{
red!60
}
教师模型
$
1
$}
}
;
\node
[modelnode
]
(tea
\curr
2) at ([yshift=-3em]tea
\prev
2.south)
{
\rotatebox
{
90
}{
\color
{
blue!60
}
教师模型
$
2
$}
}
;
\node
[modelnode,fill=yellow!20
,align=center]
(stu
\curr
1) at ([yshift=-3em]stu
\prev
1.south)
{
学
\\
[-0.5ex]
生
\\
[-0.5ex]
模
\\
[-0.5ex]
型
\\
[-0.5ex]
$
1
$
}
;
\node
[modelnode,fill=yellow!20
,align=center]
(stu
\curr
2) at ([yshift=-3em]stu
\prev
2.south)
{
学
\\
[-0.5ex]
生
\\
[-0.5ex]
模
\\
[-0.5ex]
型
\\
[-0.5ex]
$
2
$
}
;
\node
[modelnode,fill=yellow!20
,align=center]
(stu
\curr
3) at ([yshift=-3em]stu
\prev
3.south)
{
学
\\
[-0.5ex]
生
\\
[-0.5ex]
模
\\
[-0.5ex]
型
\\
[-0.5ex]
$
3
$
}
;
\node
[modelnode,fill=yellow!20
,align=center]
(stu
\curr
4) at ([yshift=-3em]stu
\prev
4.south)
{
学
\\
[-0.5ex]
生
\\
[-0.5ex]
模
\\
[-0.5ex]
型
\\
[-0.5ex]
$
4
$
}
;
\node
[modelnode,fill=yellow!20
,align=center]
(stu
\curr
5) at ([yshift=-3em]stu
\prev
5.south)
{
学
\\
[-0.5ex]
生
\\
[-0.5ex]
模
\\
[-0.5ex]
型
\\
[-0.5ex]
$
5
$
}
;
\node
[modelnode
,align=center,text=red!60]
(tea
\curr
1) at ([yshift=-3em]tea
\prev
1.south)
{
教
\\
[-0.5ex]
师
\\
[-0.5ex]
模
\\
[-0.5ex]
型
\\
[-0.5ex]
$
1
$
}
;
\node
[modelnode
,align=center,text=blue!60]
(tea
\curr
2) at ([yshift=-3em]tea
\prev
2.south)
{
教
\\
[-0.5ex]
师
\\
[-0.5ex]
模
\\
[-0.5ex]
型
\\
[-0.5ex]
$
2
$
}
;
% 集成 labels
\draw
[->,very thick]
([xshift=2pt]stu
\curr
5.east) to node [auto]
{
\small
集成
}
([xshift=-2pt]tea
\curr
1.west);
}
% iteration labels
\node
[font=\small,anchor=east,purple!80
]
(iterate1) at ([xshift=-1em]stu21.west)
{
\rotatebox
{
90
}{
轮数
$
1
$}
}
;
\node
[font=\small,anchor=east,purple!80
]
(iterate2) at ([xshift=-1em]stu31.west)
{
\rotatebox
{
90
}{
轮数
$
2
$}
}
;
\node
[font=\small,anchor=east,purple!80
,align=center]
(iterate1) at ([xshift=-1em]stu21.west)
{
轮
\\
数
\\
$
1
$
}
;
\node
[font=\small,anchor=east,purple!80
,align=center]
(iterate2) at ([xshift=-1em]stu31.west)
{
轮
\\
数
\\
$
2
$
}
;
% distillation labels
\node
[font=\small,anchor=south west]
(distill1) at ([yshift=1.2em]iterate1.north west)
{
知识蒸馏
}
;
...
...
Chapter13/Figures/figure-label-smoothing.tex
查看文件 @
d20babe2
...
...
@@ -55,8 +55,8 @@
\node
[anchor=south,font=
\scriptsize
] (w8) at (label7.north)
{$
0
.
1
$}
;
\node
[font=\scriptsize]
(line1) at ([xshift=13em,yshift=-1.5em]model
_
label7.east)
{$
Loss
=-
0
.
3
\log
p
_{
3
}
-
\sum
_{
i
=
1
}^{
7
}
0
.
1
\log
p
_{
i
}$}
;
\node
[font=\scriptsize]
(line2) at ([xshift=9.5em,yshift=3em]model
_
label7.east)
{$
Loss
=-
\log
p
_{
3
}$}
;
\node
[font=\scriptsize]
(line1) at ([xshift=13em,yshift=-1.5em]model
_
label7.east)
{$
\textrm
{
Loss
}
=-
0
.
3
\log
p
_{
3
}
-
\sum
_{
i
=
1
}^{
7
}
0
.
1
\log
p
_{
i
}$}
;
\node
[font=\scriptsize]
(line2) at ([xshift=9.5em,yshift=3em]model
_
label7.east)
{$
\textrm
{
Loss
}
=-
\log
p
_{
3
}$}
;
\begin{pgfonlayer}
{
background
}
\node
[rectangle,inner sep=0.5em,rounded corners=1pt,very thick,dotted,draw=red] [fit =(model
_
w3) (model
_
label1) (model
_
label7) (one
_
hot
_
w3)] (box1)
{}
;
...
...
Chapter15/Figures/figure-attention-distribution-based-on-gaussian-distribution.tex
查看文件 @
d20babe2
...
...
@@ -94,7 +94,7 @@
\node
[anchor=south,colnode,minimum height=4.2em,minimum width=1em] (b2) at ([xshift=1.67em,yshift=0em]b1.south)
{}
;
\node
[anchor=south,colnode,minimum height=3.7em,minimum width=1em] (b3) at ([xshift=1.67em,yshift=0em]b2.south)
{}
;
\node
[anchor=south,colnode,minimum height=4.2em,minimum width=1em] (b4) at ([xshift=1.67em,yshift=0em]b3.south)
{}
;
\node
[anchor=south,colnode,minimum height=
0.8
em,minimum width=1em] (b5) at ([xshift=1.67em,yshift=0em]b4.south)
{}
;
\node
[anchor=south,colnode,minimum height=
1.4
em,minimum width=1em] (b5) at ([xshift=1.67em,yshift=0em]b4.south)
{}
;
\node
[anchor=south,colnode,minimum height=0.15em,minimum width=1em] (b6) at ([xshift=1.67em,yshift=0em]b5.south)
{}
;
{
\scriptsize
...
...
Chapter15/chapter15.tex
查看文件 @
d20babe2
...
...
@@ -253,7 +253,7 @@ C(\mathbi{x}_j \mathbi{W}_K,\omega) &=& (\mathbi{x}_{j-\omega},\ldots,\mathbi{x}
\vspace
{
0.5em
}
\item
{
\small\bfnew
{
使用轻量卷积和动态卷积神经网络
}}
\upcite
{
Wu2019PayLA,DBLP:conf/interspeech/GulatiQCPZYHWZW20
}
。比如,分别在编码器和解码器利用轻量卷积或动态卷积神经网络(见
{
\chapternine
}
)替换Transformer的自注意力机制,同时保留解码器的编码-解码注意力机制,一定程度上加强了模型对局部信息的建模能力,同时提高了计算效率。
\vspace
{
0.5em
}
\item
{
\small\bfnew
{
使用1维卷积注意力网络
}}
(图
\ref
{
fig:15-5
}
(b))。可以使用一维的卷积自注意力网络(1D-CSAN)将关注的范围限制在相近的元素窗口中。其形式上十分简单,只需预先设定好局部建模的窗口大小
$
D
$
,并在进行注意力权重计算和对Value值进行加权求和时,将其限制在设定好的窗口范围内。
\item
{
\small\bfnew
{
使用1维卷积注意力网络
}}
(图
\ref
{
fig:15-5
}
(b))
{
\red
(没引用?)
}
。可以使用一维的卷积自注意力网络(1D-CSAN)将关注的范围限制在相近的元素窗口中。其形式上十分简单,只需预先设定好局部建模的窗口大小
$
D
$
,并在进行注意力权重计算和对Value值进行加权求和时,将其限制在设定好的窗口范围内。
\vspace
{
0.5em
}
\item
{
\small\bfnew
{
使用2维卷积注意力网络
}}
(图
\ref
{
fig:15-5
}
(c))。在一维卷积注意力网络的基础上,对多个注意力头之间的信息进行交互建模,打破了注意力头之间的界限。 1D-CDAN
{
\red
(S?)
}
的关注区域为
$
1
\times
D
$
,当将其扩展为二维矩形
$
D
\times
N
$
,长和宽分别为局部窗口的大小和参与建模的自注意力头的个数。这样,模型可以计算某个头中的第
$
i
$
个元素和另一个头中的第
$
j
$
个元素之间的相关性系数,实现了对不同子空间之间关系的建模,所得到的注意力分布
{
\red
(哪里得到了?)
}
表示了头之间的依赖关系。
...
...
编写
预览
Markdown
格式
0%
重试
或
添加新文件
添加附件
取消
您添加了
0
人
到此讨论。请谨慎行事。
请先完成此评论的编辑!
取消
请
注册
或者
登录
后发表评论