如何根据参数绘制神经网络？

Question

Izaak Neutelings 的这个页面提供了一些绘制神经网络的代码：https://tikz.net/neural_networks/运行该页面顶部的简单代码可得到以下结果：

\documentclass[border=3pt,tikz]{standalone}
\usepackage{tikz}
\tikzstyle{mynode}=[thick,draw=blue,fill=blue!20,circle,minimum size=22]
\begin{document}
\begin{tikzpicture}[x=2.2cm,y=1.4cm]
  \foreach \N [count=\lay,remember={\N as \Nprev (initially 0);}]
               in {4,5,5,5,3}{ % loop over layers
    \foreach \i [evaluate={\y=\N/2-\i; \x=\lay; \prev=int(\lay-1);}]
                 in {1,...,\N}{ % loop over nodes
      \node[mynode] (N\lay-\i) at (\x,\y) {};
      \ifnum\Nprev>0 % connect to previous layer
        \foreach \j in {1,...,\Nprev}{ % loop over nodes in previous layer
          \draw[thick] (N\prev-\j) -- (N\lay-\i);
        }
      \fi
    }
  }
\end{tikzpicture}
\end{document}

这很好，并且可以在tikzpicture环境中进行修改。但是如何通过参数（例如网络拓扑）来控制结果？

Answer 1

这是实现此目的的一种方法。基本思路是：

• 移动tikz代码到newcommand NN
• 用合适的参数替换常量
• 包括一些样式，固定箭头形状

这是众多可能的解决方案之一，包括以下@Skillmon 的评论：

\documentclass[border=3pt,tikz]{standalone}
\usepackage{tikz}
\usetikzlibrary{arrows.meta}% <<< for arrow shapes

\newcommand\NN[6]{% x, y, size, layers, nodeColor, lineColor
 % see webpage \tikzstyle{mynode}=[thin,draw=#5,fill=white,circle,minimum size=#3]%
 % see comment \tikzset{mynode/.style={thin,draw=#5,fill=white,circle,minimum size=#3}}%
 \begin{tikzpicture}[x=#1cm,y=#2cm,
                     mynode/.style={thin,draw=#5,fill=white,circle,minimum size=#3}]%
   \foreach \N [count=\lay,remember={\N as \Nprev (initially 0);}]%
               in {#4}{ % loop over layers%
    \foreach \i [evaluate={\y=\N/2-\i; \x=\lay; \prev=int(\lay-1);}]%
                 in {1,...,\N}{ % loop over nodes%
      \node[mynode] (N\lay-\i) at (\x,\y) {};%
      \ifnum\Nprev>0 % connect to previous layer%
        \foreach \j in {1,...,\Nprev}{ % loop over nodes in previous layer
          \draw[thin,-{Latex[]}, draw=#6] (N\prev-\j) --(N\lay-\i);%
        }%
      \fi%
    }%
   }%
 \end{tikzpicture}%
}

\begin{document}
    %   x, y,size, layers,  nodeColor, lineColor
    \NN{2}{2}{10}{4,2,5,3,2}{black}{black!50}
    
    \NN{2}{2}{20}{5,4,3,2}{blue}{red!50}
    
\end{document}

由此产生了两个用于演示的网络：