Skip to content

A (toy) neural network library with auto-encoder demo

License

Notifications You must be signed in to change notification settings

virtualwiz/tinymlp

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

27 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

TinyMLP

Introduction

This project implemented the single hidden layer perceptron and its back propagation training algorithm. It can be used as a toy for tinkering with simple neural networks. This library can potentially be ported onto embedded processors, to help to solve simple pattern recognition problems.

Demonstration Program

The demonstration program constructed a 4-2-4 autoencoder, trained it with an example data set and finally displayed the testing results. You should see it automatically finds an encoding scheme of the four input patterns.

To try out the demonstration:

$ make
$ ./demo

As the program uses only the basic gradient descent algorithm, it may be trapped in a local optimal solution. If the error function still does not converge to the target value after a certain number of epochs, it will automatically start from another random point in the error space.

About

A (toy) neural network library with auto-encoder demo

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published