PlotNeuralNet

Latex code for drawing neural networks for reports and presentation. Have a look into examples to see how they are made. Additionally, lets consolidate any improvements that you make and fix any bugs to help more people with this code.

Examples

Following are some network representations:

VGG16

Mask-RCNN

Single Shot MultiBox Detector (SSD)

Long Short-Term Memory (LSTM)

Getting Started

1. Install latex and other dependencies.

   • Ubuntu 16.04

     sudo apt-get install texlive-latex-extra

   • Ubuntu 18.04.2
     Base on this website, please install the following packages.

     sudo apt-get install texlive-latex-base
     sudo apt-get install texlive-fonts-recommended
     sudo apt-get install texlive-fonts-extra
     sudo apt-get install texlive-latex-extra

   • Windows

   1. Download and install MikTeX.
   2. Download and install bash runner on Windows, recommends Git bash or Cygwin

2. Install package

   git clone https://github.com/xrz000/PlotNeuralNet.git
   cd PlotNeuralNet
   pip install -e . # or python setup.py develop

3. Execute the example as followed.

   cd examples/Python
   bash ../../tikzmake.sh test_simple

TODO

• Python interface
• Add easy legend functionality
• Add more layer shapes like TruncatedPyramid, 2DSheet etc
• Add examples for RNN and likes.

Latex usage

See examples/LaTex directory for usage.

Python usage

First, create a new directory and a new Python file:

$ mkdir my_project
$ cd my_project
vim my_arch.py

Add the following code to your new file:

import sys
from plotnn import plotnn
import plotnn.tikzeng as tk


def main():
    namefile = str(sys.argv[0]).split('.')[0]
    arch = [
        tk.Conv2D("conv1", out_width=512, out_channel=64, location=(0, 0, 0), offset=(0, 0, 0),
                  height=64, depth=64, width=2, caption="Conv"),
        tk.Pool("pool1", location="conv1-east", offset=(1, 0, 0), height=32, depth=32, width=2, caption="MaxPool"),
        tk.Connection("conv1", "pool1"),

        tk.Conv2D("conv2", out_width=128, out_channel=64, location="pool1-east", offset=(2, 0, 0),
                  height=32, depth=32, width=2, caption="Conv"),
        tk.Connection("pool1", "conv2"),

        tk.Pool("pool2", location="conv2-east", offset=(1, 0, 0), height=16, depth=16, width=2, caption="MaxPool"),
        tk.Connection("conv2", "pool2"),

        tk.Softmax("soft1", out_channel=10, location="pool2-east", offset=(3, 0, 0), height=16, depth=16, width=2, caption="Softmax"),
        tk.Connection("pool2", "soft1"),

        tk.Sum("sum1", location="soft1-east", offset=(1.5, 0, 0), radius=2.5, opacity=0.6),
        tk.Connection("soft1", "sum1"),
    ]
    plotnn.generate([arch], namefile + '.tex')


def main():
    namefile = str(sys.argv[0]).split('.')[0]
    to_generate(arch, namefile + '.tex')

if __name__ == '__main__':
    main()

Now, run the program as follows:

bash ../tikzmake.sh my_arch