nndiagram: Producing 'LaTeX' Code for Drawing Neural Network Diagrams

In nndiagram: Generator of 'LaTeX' Code for Drawing Neural Network Diagrams with 'TikZ'

Producing 'LaTeX' Code for Drawing Neural Network Diagrams

Description

The nndiagram command is used to produce 'LaTeX' code for drawing well-formatted neural network diagrams. To make the code work in a 'LaTeX' editor, users need to install and import two 'TeX' packages, i.e., TikZ and ifthen in the setting of 'TeX' file. Syntax of importing these packages is included in the output of function.

Usage

nndiagram(
  input,
  hidden,
  keep = NULL,
  omit = NULL,
  title = NULL,
  color = "black",
  alpha = 1,
  layer.sep = 2.5,
  layer.label = NULL,
  input.label = NULL,
  output.label = NULL,
  suppress = FALSE
)

Arguments

input	a positive integer that specifies the number of input neurons.
hidden	a positive integer vector that specifies the number of neurons on each hidden layer. For example, c(4,4,4) specifies that there are 3 hidden layers and 4 neurons on each hidden layer. Non-positive and non-integer numbers are not allowed.
keep	an optional character vector that specifies the connections of neurons to be kept. For example, c("1->4","5->8") specifies the connections from neuron 1 to 4 and from neuron 5 to 8. Neurons are counted from top to bottom and from left to right. As special cases, "->4" specifies all connections with neuron 4 as destination, and "5->" specifies all connections with neuron 5 as source. No space in the string is allowed. Default is NULL.
omit	an optional character vector that specifies the connections of neurons to be omitted. The specification of connections is the same as that in 'keep'. However, users are not allowed to assign values to 'keep' and 'omit' simultaneously. Default is NULL.
title	an optional character that specifies the main title of diagram. Default is NULL.
color	an optional character that specifies the color of lines. Default is "black".
alpha	an optional numeric value between 0 and 1 that specifies the opacity of lines. 1 indicates lines to be opaque, and 0 indicates lines to be transparent. Default is 1.
layer.sep	an optional positive numeric value that specifies the distance between layers of a neural network. Default is 2.5.
layer.label	an optional character vector that specifies label for each layer, including input, hidden and output layers.
input.label	an optional character vector that specifies label for each input neuron.
output.label	an optional character that specifies label for output neuron.
suppress	an optional logical value that specifies whether nndiagram should suppress the output of 'LaTeX' code to be directly printed on console. Default is FALSE.

Value

nndiagram uses cat() to print out 'LaTeX' code on console, if not suppressed. Also, nndiagram saves the same output as a character vector invisibly, so users could use cat() to print it out later at their demand, as shown in Examples. The nndiagram 'LaTeX' output can be directly copied and pasted to produce neural network diagram in any 'LaTeX' editor.

Note

This package is an ongoing project, and more functions will be added in the future, such as those to produce pdf version of diagrams or convert handdrawing neural network diagrams to computerized ones. Collaborations are sincerely welcome. Comments and suggestions are always highly appreciated.

Author(s)

Chencheng Fang, Bonn Graduate School of Economics, University of Bonn. Email: [e-mail]

See Also

nndiagram_oversize; nndiagram_nodeCoverup; activation_curve.

Examples

# A neural network with 3 neurons on input layer, 4 neurons on each of 3 hidden layers,
# and 1 neuron on output layer. No connection is omitted and all other arguments are default.
nndiagram(input=3, hidden=c(4,4,4))

# Same as the first example but connections from neuron 1 to 4 and from neuron 5 to 8 are omitted.
nndiagram(input=3, hidden=c(4,4,4), omit=c("1->4","5->8"))

# Same as the first example but connections with neuron 4 as destination are omitted.
nndiagram(input=3, hidden=c(4,4,4), omit=c("->4"))

# Same as the first example but connections with neuron 5 as source are omitted.
nndiagram(input=3, hidden=c(4,4,4), omit=c("5->"))

# Suppress the output of 'LaTeX' code to be directly printed on the console and save the output
# to an object, which can be printed later at demand.
nnd <- nndiagram(input=3, hidden=c(4,4,4), suppress=TRUE)
cat(paste(nnd,"\n"))

nndiagram documentation built on April 18, 2023, 5:11 p.m.