Skip to content

A dynamic multilayer neural network implementation with RProp for MNIST classification.

Notifications You must be signed in to change notification settings

Noctino52/RPropNetBuilder

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

RPropNetBuilder

RPropNetBuilder is an implementation of a dynamic multilayer neural network. It simulates both forward and backward propagation, allowing for adjustments to network parameters such as the number of internal layers and nodes.

Functionalities:

  • Forward and backward propagation simulation
  • Adjustable internal layers and nodes
  • Utilizes the RProp weight update algorithm
  • Supports activation functions (Identity, Sigmoid) and error functions (Sum of Squares, Cross Entropy)

What is it?

It's a Python-based program for creating and training neural networks, specifically designed to solve a 10-class classification problem using the MNIST dataset.

Technologies used:

  • Python
  • Neural network framework (implemented from scratch)

Stage of the project:

The project is in a complete state, with all functionalities implemented and tested using the MNIST dataset. Further development may include extending capabilities or adapting it for different datasets or problems.

The project aims to demonstrate proficiency in neural network implementation and training. It's designed to solve a specific classification problem using established algorithms.

About

A dynamic multilayer neural network implementation with RProp for MNIST classification.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages