使用 latex 绘制神经网络之Resnet 结构
\begin{figure}[H]\begin{center}\begin{tikzpicture}[scale=0.75]\node[circle, fill=cyan!60,inner sep=4pt] (input) at (1.5, 4.5) {$y_{in}$};\node[rectangle, rounded corners, fill=green!60,inner sep=8pt]
·
\begin{figure}[H]
\begin{center}
\begin{tikzpicture}[scale=0.75]
\node[circle, fill=cyan!60,inner sep=4pt] (input) at (1.5, 4.5) {$y_{in}$};
\node[rectangle, rounded corners, fill=green!60,inner sep=8pt] (ResNet) at (1.5, 2.5) {ResNet block};
\node[rectangle, fill=green!0,inner sep=2pt] (resnet1) at (4, 2.5) {};
\node[rectangle, fill=green!0,inner sep=2pt] (resnet2) at (7, 2.5) {};
\draw[line width=1.3pt,->] (input) -- (ResNet);
\node[circle, fill=green!60,inner sep=3.5pt] (output) at (1.5, 0.5) {$y_{out}$};
\draw[line width=1.3pt,->] (ResNet) -- (output);
\node[circle, fill=cyan!60,inner sep=4pt] (input) at (9.5, 5.5) {$y_{in}$};
\node[rectangle, rounded corners, minimum width = 2.5cm, minimum height = 0.5cm, fill=green!60,inner sep=4pt] (layer-begin) at (9.5, 4) {hidden layer$_1$};
\node[rectangle, rounded corners, minimum width = 2.5cm, minimum height = 0.5cm, fill=green!60,inner sep=4pt] (layer-mid) at (9.5, 2.75) {$\cdots\cdots$};
\node[rectangle, rounded corners, minimum width = 2.5cm, minimum height = 0.5cm, fill=green!60,inner sep=4pt] (layer-end) at (9.5, 1.5) {hidden layer$_N$};
\node[circle, fill=green!60,inner sep=2pt] (otimes) at (9.5, 0.25) {$\oplus$};
\node[circle, fill=green!60,inner sep=3pt] (output) at (9.5, -1.25) {$y_{out}$};
\draw[line width=1.3pt,->] (input) -- (layer-begin);
\draw[line width=1.3pt,->] (layer-begin) -- (layer-mid);
\draw[line width=1.3pt,->] (layer-mid) -- (layer-end);
\draw[line width=1.3pt,->] (layer-end) -- (otimes);
\draw[line width=1.3pt,->] (otimes) -- (output);
\draw[line width=1.3pt,->] (input)..controls(13.5,5) and (13.5,1)..(otimes);
\path[-stealth, line width=1.5pt] (resnet1) edge [double] (resnet2);
\end{tikzpicture}
\end{center}
\caption{ResNet}
\label{fig01}
\end{figure}
更多推荐
已为社区贡献6条内容
所有评论(0)