degann 1.0

Library for generating artificial neural networks for modeling the behavior of dynamic systems

DEGANN

DEGANN is a library generating neural networks for approximating solutions to differential equations. As a backend for working with neural networks, tensorflow is used, but with the ability to expand with your own tools.

Features

• Generation of neural networks by parameters.
• Construction of tables with the numerical solution of ordinary differential equations of the first order
• Construction of tables with numerical solution of systems of ordinary differential equations of the first order
• Choosing the Best Neural Network from Several for Fixed Training Parameters
• Iterating over training parameters with choosing the best neural network for each set
• Export Neural Networks as a function in C++
• Export Neural Networks as a Parameter Set
• Import Neural Networks from a Parameter Set
• Building a dataset with complete training results for approximating the solution of a differential equation for each neural network that participated in training
• Advanced search for optimal topology using a language that describes the topology of a neural network
• Random search for optimal neural network topology
• Method for simulating annealing of optimal neural network topology
• Expert system for automatic selection of optimal parameters for algorithms for searching neural network topologies

Install

Manual

Download the repository as a zip archive, unpack and run the command from the root of the repository

pip install -r requirements.txt

This will download and install all the dependencies for the project, then you can use the source code of the library (e.g. create project nearby to the degann folder)

Via pip

Use command

pip install degann

Now you can use the degann package