Lightweight Neural Network is a lightweight implementation in C of a backpropagation neural network. It is intended for use in programs that just happen to need a simple neural network and do not want to incorporate a needlessly complex neural network library. It features multilayer backpropagation neural network with settable momentum and learning rate, easy portability, and small size. Included is an example application that trains a neural network to recognize handwritten digits.
|Tags||Software Development Libraries Scientific/Engineering Mathematics Artificial Intelligence|
Release Notes: Support for trainable bias.
Release Notes: This release adds support for batch training.
Release Notes: This release has improved error handling for file I/O, and the option to save/load a binary representation of a neural network. Furthermore, it contains an example application that trains a neural network to recognize handwritten digits.
Release Notes: The return type of some functions were changed from network_t * to void. The name of the header file was changed to lwneuralnet.h and the library to liblwneuralnet.a. Some changes were made to make the API more consistent and additional documentation was included.
No changes have been submitted for this release.