🌙 🖧 A generic classification neural network framework written in Lua
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
src
.gitignore
README.md

README.md

Luanne

a series of small apps that accommodate a very simple image classification machine learning library.

This does not have practical use outside of the research project, but if you'd like to play with/modify it, feel free!

Dependencies:

  • luafilesystem sudo luarocks install luafilesystem

Limitations

  • Currently can only convert BITMAP images to vector files

Running the Project

Preparing your image folder to be vectorised

Run this command in your image folder, make sure to resize the images to 30x30 and replacing *.jpg with whatever filetype the images are currently

mogrify -format bmp -resize 30x30! *.jpg

Initialise project

navigate to the "src" folder and run:

lua init.lua

There will be on-screen instructions to follow, name your project something simple (with no spaces or dashes etc)

Learning

Once you've initialised your new project, run:

luajit main.lua learn {name of your project}

to learn from your given image datasets.

NOTE: To safely exit the learning loop (as to not cancel while saving synapses), change the value in "status" to 0, and the process will be cancelled on the next synapse save.

Testing

To test a single BITMAP file (which must be 30x30):

luajit main.lua do {name of your project} {bitmap filepath}