let targetLabel = 'C';

let letters = ['C', 'D', 'E'];

​

let state = 'collection';

​

function setup() {

  createCanvas(400, 400);

​

  const options = {

    inputs: ['x', 'y'],

    outputs: ['label'],

    task: 'classification',

    debug: 'true'

  };

​

  model = ml5.neuralNetwork(options);

}

​

function keyPressed() {

  // to train the model, I'm checking if the key pressed is 't'

  if (key === 't') {

    state = 'training';

    console.log('starting training 🏃‍♀️')

    // function in ml5 that will normalize the data (between 0 and 1)

    model.normalizeData();

    let options = {

      epochs: 100

    }

    model.train(options, whileTraining, finishTraining);

  } else {

    // check if the key pressed is one of the target letters in the letters array

    if (letters.includes(key.toUpperCase())) {

      targetLabel = key.toUpperCase();

    }

  }

}

​

function finishTraining() {

  console.log(`I'm done training 🏃‍♂️`);

  state = 'prediction';

}

​

function whileTraining(epoch, loss) {

  console.log(epoch);

}

​

function mousePressed() {

​

  let inputs = {

    x: mouseX,

    y: mouseY

  }

​

  if (state === 'collection') {

    let target = {

      label: targetLabel

    }

​

  // function in p5js that takes key pair/values matching the defined

  // inputs and outputs defined in the options object

  model.addData(inputs, target);

​

  stroke(0);

  noFill();

  circle(mouseX, mouseY, 24);

  fill(0);

  textAlign(CENTER, CENTER);

  text(targetLabel, mouseX, mouseY);

  } else if(state === 'prediction') {

    model.classify(inputs,gotResults);

  }

}

​

function gotResults(error,results) {

  if(error){

    console.error(error);

    return;

  }

  console.log(results);

}