3net.js

3net.js

A simple library for implementing 3 layer neural networks

Initialization

var three_net = require('3net.js');  // Install with 'npm install 3net.js'
var inputLayer = 400;
var hiddenLayer = 25;
var outputLayer = 10;
var net = three_net.createNet(inputLayer, hiddenLayer, outputLayer);  

Training

options = {
    "iters": 500,           // Max number of iterations for gradient descent. Default is 500.
    "learning_rate": 0.5,   // Learning rate for gradient descent. Default is 0.5.
    "regularization": 0.1,  // Regularization parameter to prevent overfitting. Default is 0.1.
    "error_bound": 0.001    // If the change in cost is less than this value during gradient descent, it finishes. Default is 0.001.
};

// Data and label must be an array matching the dimensions of the input layer and output layer. 
// If options is not specified, the default values will be used. Returns true if training was successful, else returns false.
var success = net.train(data, label, options);
if (success) console.log("training complete");  

Predicting

net.predict(data);  // Returns an array with the output layer activations

Importing and exporting

var savedNet = net.exportNet();                 // Exports as JSON
var copiedNet = three_net.importNet(savedNet);  // Imports as JSON