Simple-Swift-AI

A simple to use Feed Forward Neural Network built in Swift 3.0

Usage

• A data set with multiple inputs and outputs can be used to train the network.
• The default number of hidden layers is one, more can be added as per need.
• The neuron output values range from 0 to 1.0. Modify your data set accordingly.

Features

• Feed-forward neural network with input, hidden and output layers
• Multiple hidden layers can be added with little extra code
• Can come in use for categorization and function value prediction
• Optional debug screen included for viewing the training and test outputs on simulator

---

An example of XOR gate training is included.

---

FFNN Training screen

---

FFNN Test screen

---

By Arjun Gupta

Can be reached at : arjungupta497 (at) gmail.com

Twitter : https://www.twitter.com/arjunguptaxor