Lightweight Neural Network / News: Recent posts

This is a lightweight implementation of a neural network in C. It provides a multilayer perceptron network with backpropagation training. It is small, fast, easy to embed in an application and easy to extend.

Lwneuralnet doesn't do anything fancy: no millions of different types of neural networks, no graphical user interface, just the routines for computing and training a backpropagation neural network. It is primarily intended for use in applications that just happen to need a simple neural network and have no use for complicated neural network libraries.... read more

Lightweight Neural Network (lwneuralnet), Version 0.7

This is a lightweight implementation of a neural network in C. It provides a multilayer perceptron network with backpropagation training. It is small, fast, easy to embed in an application and easy to extend.

Lwneuralnet doesn't do anything fancy: no millions of different types of neural networks, no graphical user interface, just the routines for computing and training a backpropagation neural network. It is primarily intended for use in applications that just happen to need a simple neural network and have no use for complicated neural network libraries.... read more

Lwneuralnet is a lightweight implementation of a neural network in C. It provides a multilayer perceptron network with backpropagation training. It is small, fast, easy to embed in an application and easy to extend.

Lwneuralnet doesn't do anything fancy: no millions of different types of neural networks, no provisions for batch training, no graphical user interface, just the routines for computing and training a backpropagation neural network. It is primarily intended for use in applications that just happen to need a simple neural network and have no use for complicated neural network libraries.... read more

More documentation added, API stablized. Only some polishing left (and maybe a larger example application) before it can be called 1.1.

A websiite (http://lwneuralnet.sourceforge.net) has been created.

Peter