DynNet and GINNet projects

Version française

Launch GINNet

- Overview - Features - Screenshots - Downloads - FAQ - Authors

GForge project Documentation Bug tracker Forums

Overview

DynNet

DynNet is a Java library which contains basic elements for building neural networks. Neural network classical models are already available (Multi-layer perceptron, Kohonen self-organizing maps, neural gas, growing neural gas, etc.).

GINNet

GINNet (Graphical Interface for Neural Networks & co) is a decision-making platform written in Java. It has been developped to favorize the developpement and use of neural networks.

This platform is composed of 3 parts:

Classes for data importation from various sources (text files, databases...), and statistics tools. This data can be exported to a corpus, which can be used directly by neural networks.
DynNet library
A graphical user interface that integrates these tools into an application that handle every aspect of decision-making process.

DynNet has been developped by the CORTEX team at LORIA, a computer science research center at Nancy, France