//

// This software is released under the MIT License.

// https://opensource.org/licenses/MIT

​

/* ===

ml5 Example

Image classification using MobileNet and p5.js

This example uses a callback pattern to create the classifier

=== */

let nn;

let resultsText;

function setup() {

  // Options for Neural Network

  createCanvas(400, 400);

  const options = {

    inputs: ['x', 'y'],

    outputs: ['label'],

    task: 'classification',

    debug: true

  };

  // Create Neural Network

  nn = ml5.neuralNetwork(options);

​

  // trainModel();

  resultsText = select('#results');

​

  // Train the model

  let trainBtn = select('#trainBtn');

  // trainBtn.position(10, 50);

  trainBtn.mousePressed(function () {

    trainModel();

  });

​

  // Save and download the model

  // let saveBtn = createButton('Save Model');

  // saveBtn.position(10, 90);

  // saveBtn.mousePressed(function () {

  //   nn.save();

  // });

​

  // // Load the model from local files

  // let loadLocalBtn = createButton('Load the model from local files');

  // loadLocalBtn.position(10, 110);

  // loadLocalBtn.mousePressed(function () {

  //   nn.load('model/model.json', function () {

  //     console.log('Model Loaded!');

  //   });

  // });

​

  // Load Data

  let loadBtn = select('#load');

  loadBtn.changed(function () {

    nn.loadData(loadBtn.elt.files, function () {

      console.log('Data Loaded!');

    });

  });

}

​

function draw(){

  background(240);

}

​

function trainModel() {

  const trainingOptions = {

    epochs: 32,

    batchSize: 12

  }

  // Train

  nn.normalizeData();

  nn.train(trainingOptions, finishedTraining);

}

​

// Training callback

function finishedTraining() {

  classify();

}

​

function classify() {

  nn.classify({x:mouseX, y:mouseY}, gotResults);

}

​

function gotResults(error, results) {

  if (error) console.log(error);

  if (results) {

    // console.log(results)

    resultsText.html(`${results[0].label}`)

    classify();

  }

}