//

// 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 counter = 0;

​

const options = {

  task:'regression',

  debug: true

}

​

function setup(){

  createCanvas(400, 400);

  nn = ml5.neuralNetwork(options);

​

​

  console.log(nn)

  createTrainingData();

  nn.normalizeData();

​

  const trainingOptions={

    batchSize: 24,

    epochs: 10

  }

  nn.train(trainingOptions,finishedTraining); // if you want to change the training options

  // nn.train(finishedTraining); // use the default training options

}

​

function finishedTraining(){

​

  if(counter < 400){

    nn.predict([counter], (err, results) => {

      if(err){

        console.log(err);

        return;

      }

      console.log(results[0]);

      let prediction = results[0]

      let x = counter;

      let y = prediction.value

      fill(255, 0, 0);

      rectMode(CENTER);

      rect(x, y, 10, 10);

​

      counter++;

      finishedTraining();

    })

  }

}

​

function createTrainingData(){

  for(let i = 0; i < width; i+=5){

    const iters = floor(random(5, 20))

    const spread = 20;

    for(let j = 0; j < iters; j++){

      let data = [i, height/2 - sin(i/40) * 50 + floor(random(-spread, spread))]

      fill(0, 0, 255);

      ellipse(data[0], data[1], 10, 10)

      nn.addData([data[0]], [data[1]])

    }

  }

}

​