Autoencoder

Variational autoencoder architecture. Made with https://github.com/battlesnake/neural.

Edit and compile if you like:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
% Variational autoencoder architecture.
% Made with https://github.com/battlesnake/neural.
\documentclass[tikz]{standalone}
\usepackage{neuralnetwork}
\newcommand{\xin}[2]{$x_#2$}
\newcommand{\xout}[2]{$\hat x_#2$}
\begin{document}
\begin{neuralnetwork}[height=8]
  \tikzstyle{input neuron}=[neuron, fill=orange!70];
  \tikzstyle{output neuron}=[neuron, fill=blue!60!black, text=white];
  \inputlayer[count=8, bias=false, title=Input Layer, text=\xin]
  \hiddenlayer[count=5, bias=false]
  \linklayers
  \hiddenlayer[count=3, bias=false, title=Latent\\Representation]
  \linklayers
  \hiddenlayer[count=5, bias=false]
  \linklayers
  \outputlayer[count=8, title=Output Layer, text=\xout]
  \linklayers
\end{neuralnetwork}
\end{document}
 
 
הההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההה
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX