// Cloned by Samaksh Chandra on 26 Jul 2022 from World "Character recognition neural network" by "Coding Train" project
// Please leave this clone trail here.
//For everywhere I have edited and inserted my code, I have used the text 'my code here'
//and to indicate the end to it, I have used the text 'my code ends'
// Port of Character recognition neural network from here:
// https://github.com/CodingTrain/Toy-Neural-Network-JS/tree/master/examples/mnist
// with many modifications
// --- defined by MNIST - do not change these ---------------------------------------
const PIXELS = 28; // images in data set are tiny
const PIXELSSQUARED = PIXELS * PIXELS;
// number of training and test exemplars in the data set:
const NOTRAIN = 60000;
const NOTEST = 10000;
//--- can modify all these --------------------------------------------------
// no of nodes in network
const noinput = PIXELSSQUARED;
const nohidden = 64;
const nooutput = 10;
const learningrate = 0.1; // default 0.1
// should we train every timestep or not
let do_training = true;
//my code here
// how many to train and test per timestep
//const TRAINPERSTEP = 30;
//here I have changed the value of TRAINPERSTEP to 25
const TRAINPERSTEP = 25;
//const TESTPERSTEP = 5;
//Similarly I have changed the value of TESTPERSTEP to 10;
const TESTPERSTEP = 10;
//my code ends
// multiply it by this to magnify for display
const ZOOMFACTOR = 7;
const ZOOMPIXELS = ZOOMFACTOR * PIXELS;
// 3 rows of
// large image + 50 gap + small image
// 50 gap between rows
const canvaswidth = ( PIXELS + ZOOMPIXELS ) + 50;
const canvasheight = ( ZOOMPIXELS * 3 ) + 100;
const DOODLE_THICK = 18; // thickness of doodle lines
const DOODLE_BLUR = 3; // blur factor applied to doodles
let mnist;
// all data is loaded into this
// mnist.train_images
// mnist.train_labels
// mnist.test_images
// mnist.test_labels
let nn;
let trainrun = 1;
let train_index = 0;
let testrun = 1;
let test_index = 0;
let total_tests = 0;
let total_correct = 0;
// images in LHS:
let doodle, demo;
let doodle_exists = false;
let demo_exists = false;
let mousedrag = false; // are we in the middle of a mouse drag drawing?
// save inputs to global var to inspect
// type these names in console
var train_inputs, test_inputs, demo_inputs, doodle_inputs;
// Matrix.randomize() is changed to point to this. Must be defined by user of Matrix.
function randomWeight()
{
return ( AB.randomFloatAtoB ( -0.5, 0.5 ) );
// Coding Train default is -1 to 1
}
// make run header bigger
AB.headerCSS ( { "max-height": "95vh" } );
//--- start of AB.msgs structure: ---------------------------------------------------------
// We output a serious of AB.msgs to put data at various places in the run header
var thehtml;
// 1 Doodle header
thehtml = "<hr> <h1> 1. Doodle </h1> Top row: Doodle (left) and shrunk (right). <br> " +
" Draw your doodle in top LHS. <button onclick='wipeDoodle();' class='normbutton' >Clear doodle</button> <br> ";
AB.msg ( thehtml, 1 );
// 2 Doodle variable data (guess)
// 3 Training header
thehtml = "<hr> <h1> 2. Training </h1> Middle row: Training image magnified (left) and original (right). <br> " +
" <button onclick='do_training = false;' class='normbutton' >Stop training</button> <br> ";
AB.msg ( thehtml, 3 );
// 4 variable training data
// 5 Testing header
thehtml = "<h3> Hidden tests </h3> " ;
AB.msg ( thehtml, 5 );
// 6 variable testing data
// 7 Demo header
thehtml = "<hr> <h1> 3. Demo </h1> Bottom row: Test image magnified (left) and original (right). <br>" +
" The network is <i>not</i> trained on any of these images. <br> " +
" <button onclick='makeDemo();' class='normbutton' >Demo test image</button> <br> ";
AB.msg ( thehtml, 7 );
// 8 Demo variable data (random demo ID)
// 9 Demo variable data (changing guess)
const greenspan = "<span style='font-weight:bold; font-size:x-large; color:darkgreen'> " ;
//--- end of AB.msgs structure: ---------------------------------------------------------
//my code here
//We need to choose an appropriate activation function for our neural network to optimally decide on the weights
//for that, I am using the swish activation function, which is given by the formula below
//y = x.sigmoid(x)
//after having defined the sigmoid function, now writing the swish function becomes easy
function my_swish_function(x){
return x*(1/(1+Math.exp(-x)))
//where (1/(1+Math.exp(-x))) is the sigmoid function
}
//then again here I implement the softmax function
//because we have a multi-class classification problem
//and we use sigmoid for binary class classification
function softmax_for_nn(doodle_train_list){
var iterable = 0;
var add = 0;
//since softmax calculates three probablities, we have to intialize them here for updating them later on
var probabilities = [[0],[0],[0]];
while(iterable<=doodle_train_list.length - 1){
add += Math.exp(doodle_train_list[iterable])
//we have used the iterable in doodle_train_list to calculate the exponent of each element in the training list of doodles
iterable++;
}
//having defined the exponents, now we have to divide it by addition of all such doodles
//because softmax = e**y/sum(e**y) for all i values
var local_iter = 0;
while(local_iter<=doodle_train_list.length - 1){
probabilities[local_iter] = Math.exp(doodle_train_list[local_iter])/add;
local_iter++;
//updated probabilities
}
//now we have to send it back to the calling function
return probabilities;
}
//my code ends
function setup()
{
createCanvas ( canvaswidth, canvasheight );
doodle = createGraphics ( ZOOMPIXELS, ZOOMPIXELS ); // doodle on larger canvas
doodle.pixelDensity(1);
// JS load other JS
// maybe have a loading screen while loading the JS and the data set
AB.loadingScreen();
$.getScript ( "/uploads/codingtrain/matrix.js", function()
{
$.getScript ( "/uploads/codingtrain/nn.js", function()
{
$.getScript ( "/uploads/codingtrain/mnist.js", function()
{
console.log ("All JS loaded");
nn = new NeuralNetwork( noinput, nohidden, nooutput );
nn.setLearningRate ( learningrate );
loadData();
});
});
});
}
//my code here
//this is a function to perform matrix multiplication operations
class Define_nn_data{
constructor(define_x_input, define_y_output){
this.define_x_input = define_x_input;
this.define_y_output = define_y_output;
this.storage_list = [];
iterable = 0;
//first we will intiialize the matrix
while(iterable<this.define_x_input){
local_iter = 0;
storage_list[iterable] = [];
while(local_iter<=this.define_y_output - 1){
storage_list[iterable][local_iter] = [];
local_iter++;
}
iterable++;
}
//this array list will hold all the inputs that we will pass to our neural network model
}
// initialize_my_matrix(define_x_input,define_y_output){
// iterable = 0;
// //first we will intiialize the matrix
// while(iterable<define_x_input){
// local_iter = 0;
// storage_list[iterable] = [];
// while(local_iter<=define_y_output - 1){
// storage_list[iterable][local_iter] = [];
// local_iter++;
// }
// iterable++;
// }
// return storage_list;
// }
//here we define the multiplication of an array with a scalar value
perform_scalar_multiply(value){
iterable = 0;
while(iterable<this.define_x_input){
local_iter = 0;
while(local_iter<this.define_y_output){
this.storage_list[iterable][local_iter] *= value;
local_iter++;
}
iterable++;
}
}
output_from_matrix(my_mat){
//this is a method to extract any element from the matrix
var extract_content = new Define_nn_data(my_mat.length, 1);
iterable = 0;
while(iterable<my_mat.length){
extract_content.storage_list[iterable][0] = my_mat[iterable];
iterable++;
}
return extract_content;
}
perform_subtract(my_matrix_1, my_matrix_2){
iterable = 0;
out_mat = new Define_nn_data(my_matrix_1.define_x_input,my_matrix_1.define_y_output);
while(iterable<out_mat.define_x_input){
local_iter = 0;
while(local_iter<out_mat[iterable].define_y_output){
out_mat.storage_list[iterable][local_iter] = my_matrix_1.storage_list[iterable][local_iter] - my_matrix_2.storage_list[iterable][local_iter];
local_iter++;
}
iterable++;
}
return out_mat;
}
//method to fill any data that is required for training or testing
fill_data(){
var out_mat = [];
iterable = 0;
while(iterable<this.define_x_input){
local_iter = 0;
while(local_iter<this.define_y_output){
out_mat.push(this.storage_list[iterable][local_iter])
local_iter++;
}
iterable++;
}
return out_mat
}
gen_values(){
iterable = 0;
while(iterable<this.define_x_input){
local_iter = 0;
while(local_iter<this.define_y_output){
this.storage_list[iterable][local_iter] = Math.random() * 3 - 2;
local_iter++;
}
iterable++;
}
}
perform_add(my_matrix_data){
iterable = 0;
var sum_matrix = new Define_nn_data(this.define_x_input,this.define_y_output)
if(my_matrix_data instanceof Define_nn_data){
while(iterable<this.define_x_input){
local_iter = 0;
while(local_iter<this.define_y_output){
this.storage_list[iterable][local_iter] += my_matrix_shape.storage_list[iterable][local_iter];
local_iter++;
}
iterable++;
}
}
else{
iterable = 0;
while(iterable<this.define_x_input){
local_iter = 0;
while(local_iter<this.define_y_output){
this.storage_list[iterable][local_iter] += my_matrix_data;
local_iter++;
}
iterable++;
}
}
}
switch_matrix_elements(my_matrix){
switched_matrix = new Define_nn_data(my_matrix.define_y_output,my_matrix.define_x_input);
iterable = 0;
while(iterable<my_matrix.define_x_input){
local_iter = 0;
while(local_iter<my_matrix.define_y_output){
switched_matrix.storage_list[local_iter][iterable] = my_matrix.storage_list[iterable][local_iter];
local_iter++;
}
iterable++;
}
return switched_matrix;
}
perform_multiply(my_matrix1, my_matrix2){
//here if the column shape of the second matrix is not equal to the column shape of first matrix
//then we cannot multiply
if(my_matrix2.define_y_output!= my_matrix1.define_x_input){
console.log('cannot perfrom matrix multiplication');
return 0;
}
else{
//do the matrix multiplication
var mat_mul = new Define_nn_data(my_matrix1.define_x_input, my_matrix2.define_y_output);
//since row of first should match the column of second matrix
iterable = 0;
while(iterable<mat_mul.define_x_input){
local_iter = 0;
while(local_iter<mat_mul.define_y_output){
var add = 0;
sub_local_iter = 0;
while(sub_local_iter<my_matrix1.define_y_output){
add += my_matrix1.storage_list[iterable][sub_local_iter] * my_matrix2.storage_list[sub_local_iter][local_iter];
sub_local_iter++;
}
mat_mul.storage_list[iterable][local_iter] = add;
local_iter++;
}
iterable++;
}
}
return mat_mul;
}
chart_out_function(f_value){
//here f_value is the function value
iterable = 0;
while(iterable<this.define_x_input){
local_iter = 0;
while(local_iter<this.define_y_output){
var data = this.storage_list[iterable][local_iter];
this.storage_list[iterable][local_iter] = f_value(data);
local_iter++;
}
iterable++;
}
}
}
//after this class I will define my neural network solution
class my_nn_solution{
//here i will first define the input layers, hidden layers, and the output layers
//we can do that easily with the help of a constructor to initialize values
constructor(layer_input,layer_hidden_layer_output){
this.layer_input = layer_input;
this.layer_hidden = layer_hidden;
this.layer_output = layer_output;
//after having defined these, our next step is to initialize the weights and biases of all the layers
//this is done using the same constructor
this.w_ih = new Define_nn_data(this.layer_hidden, this.layer_input);
//w_ih refers to the weight of the hidden layer relative to the input layer
this.w_ho = new Define_nn_data(this.layer_output, this.layer_hidden);
//w_ho refers to the weight of the output layer relative to the hidden layer
this.w_ih.gen_values();
//initializing with generic values
this.w_ho.gen_values();
//initializing with generic values
//now its time to add the biases
this.b_layerhidden = new Define_nn_data(this.layer_hidden, 1);
this.b_layeroutput = new Define_nn_data(this.layer_output, 1);
//initializing with generic values
this.b_layerhidden.gen_values();
//intializing with generic values
this.b_layeroutput.gen_values();
}
my_training_function(data_at_input_layer, result_from_output_layer){
//this training function will take the input data and feed it to the model
//and then collect the output data for the training data
var data_input = Define_nn_data.output_from_matrix(data_at_input_layer);
//next its important to define the data at the hidden layer
//which is nothing but a dot product of the array list from the input layer and the data extracted from the activation function
var data_hidden_layer = Define_nn_data.perform_multiply(this.w_ih,data_input);
//we have to use the activation function now
my_swish_function.chart_out_function(swish);
//specifying the outputs here
var data_at_output_layer = new Define_nn_data(this.w_ho, data_hidden_layer);
data_at_output_layer.add(this.b_layeroutput);
//now we will deliver the output to the softmax function as it is the final layer
//swish was for input layer and hidden layer
data_at_output_layer.storage_list = softmax_for_nn(data_at_output_layer.storage_list);
//now get the data from the array list into the matrix
var final_training_data = Define_nn_data.output_from_matrix(result_from_output_layer);
//the next step will be now to determine the errors
var error_at_output_layer = Define_nn_data.perfrom_subtract(final_training_data, data_at_output_layer);
//since errors are a difference between observed output and actual output
//hence, our gradient can be defined here using errors and learning rate
//learning rate has been fixed to 0.001
var grad_lr = Define_nn_data.chart_out_function(my_swish_function);
//this is important to get the total error at output layer
grad_lr.perform_multiply(error_at_output_layer);
grad_lr.perform_scalar_multiply(0.001);
//now we will attempt to adjust the weights using back-propagation
var backprop_quotient = Define_nn_data.switch_matrix_elements(data_hidden_layer);
var weights_backprop = Define_nn_data.perform_scalar_multiply(grad_lr, backprop_quotient);
//now we will change the weights of the layers
this.w_ho.add(weights_backprop);
this.b_layeroutput.add(grad_lr);
//now we have adjusted the weights too
//backtracking the process, now we need to do the same thing for the hidden layer
var hidden_backprop_quotient = Define_nn_data.switch_matrix_elements(this.w_ho);
var hidden_weight_backprop = Define_nn_data.perfor_scalar_multiply(hidden_backprop_quotient, error_at_output_layer);
}
//we have defined our training function
//now we will define our test function
my_testing_function(test_data){
//again we have to define the data at input layer in the form of an array list for our neural network to correctly
//classify it
var data_from_input = Define_nn_data.output_from_matrix(data);
//now we will define what test data is fed to our hidden layer
var data_at_hidden = Define_nn_data.perform_scalar_multiply(this.w_ih, data_from_input);
//note that I have used data_from_input at hidden layer and not test_data as
//test_data is fed to the input layer
//however, our hidden layer receives data from the input_layer after passing through the activation function
//and this is what we will do next, pass it through the activation function
data_at_hidden.chart_out_function(my_swish_function);
var data_at_output = Define_nn_data.perform_multiply(this.w_ho,data_at_hidden);
//this is to get the correct data from the hidden layer
data_at_output.add(this.b_layeroutput);
data_at_output.storage_list = softmax_for_nn(data_at_output.storage_list);
//this is the same what we did in the training function
return data_at_output.fill_data();
}
}
//now we will write several functions to perform various operations on our matrix which we just initialized
//now we will write a function for subtracting elements from the array
// function check_for_mat_mul(my_matrix_1, my_matrix_2){
// iterable = 0;
// if(my_matrix_2[0].length!=my_matrix_1.length){
// console.log('Mat mul not possible')
// var counter = 0;
// return counter;
// }
// else{
// counter = 1;
// return counter;
// }
// }
function perform_multiply(my_matrix_1, my_matrix_2){
var check = check_for_mat_mul(my_matrix_1, my_matrix_2);
if(check === 0){
console.log('Matrix multiplication not allowed')
return 0;
}
else{
mul_matrix = initialize_my_matrix(x,y);
iterable = 0;
var sum = 0;
while(iterable<mul_matrix.length){
local_iter = 0;
while(local_iter<mul_matrix[iterable].length){
for(var i = 0;i<col_y1;i++){
sum += my_matrix_1[iterable][i] * my_matrix_2[i][local_iter];
}
mul_matrix[iterable][local_iter] = sum;
local_iter++;
}
iterable++;
}
}
return mul_matrix;
//matrix multiplication code is complete now
}
//after intializing the matrix, this is a function to fill the matrix with values
// function fill_values(my_matrix,x,y){
// iterable = 0;
// local_const = 0;
// temp_matrix = initialize_my_matrix(x,y);
// while(iterable<my_matrix.length){
// temp_matrix[iterable][local_const] = my_matrix[iterable];
// iterable++;
// }
// }
//function to get the rows of a matrix
// function get_row(my_mat){
// return my_mat.length
// }
// //function to get the columns of a matrix
// function get_col(my_mat, col){
// var columns_in_mat = [];
// for(var i = 0;i<my_mat.length;i++){
// columns_in_mat.push(my_mat[i][col])
// }
// return columns_in_mat;
// }
function generalize_values(my_mat){
iterable = 0;
var row_my_mat = get_row(my_mat);
var col_my_mat = get_col(my_mat,0);
while(iterable<row_my_mat){
local_iter = 0;
while(local_iter<col_my_mat){
my_mat[iterable][local_iter] = Math.random()*2 - 1;
}
}
return my_mat;
}
//now I will define a neural network solution for the MNIST dataset
// class my_nn{
// constructor(i,h,o, lr){
// //x is my input nodes
// //y is my hidden nodes (hidden layer)
// //z is my output nodes
// this.i = x;
// this.h = y;
// this.z = z;
// //this is the weight of the network from the input layer to the hidden layer
// this.w_ih = initialize_my_matrix(this.h, this.i);
// //this is the weight of the network from the hidden layer to the output layer
// this.w_ho = initialize_my_matrix(this.o, this.h);
// this.w_ih = generalize_values(this.w_ih);
// this.w_ho = generalize_values(this.ho);
// this.b_hidden_layer = initialize_my_matrix(this.h, 1);
// this.b_output_layer = initialize_my_matrix(this.o, 1);
// this.b_hidden_layer = generalize_values(this.b_hidden_layer);
// this.b_output_layer = generalize_values(this.b_output_layer);
// }
// }
// load data set from local file (on this server)
function loadData(){
loadMNIST ( function(data)
{
mnist = data;
console.log ("All data loaded into mnist object:")
console.log(mnist);
AB.removeLoading(); // if no loading screen exists, this does nothing
});
}
function getImage ( img ) // make a P5 image object from a raw data array
{
let theimage = createImage (PIXELS, PIXELS); // make blank image, then populate it
theimage.loadPixels();
for (let i = 0; i < PIXELSSQUARED ; i++)
{
let bright = img[i];
let index = i * 4;
theimage.pixels[index + 0] = bright;
theimage.pixels[index + 1] = bright;
theimage.pixels[index + 2] = bright;
theimage.pixels[index + 3] = 255;
}
theimage.updatePixels();
return theimage;
}
function getInputs ( img ) // convert img array into normalised input array
{
let inputs = [];
for (let i = 0; i < PIXELSSQUARED ; i++)
{
let bright = img[i];
inputs[i] = bright / 255; // normalise to 0 to 1
}
return ( inputs );
}
function trainit (show) // train the network with a single exemplar, from global var "train_index", show visual on or off
{
let img = mnist.train_images[train_index];
let label = mnist.train_labels[train_index];
//my code here
//I have written the function which will be prepared for training
//and after that I have written the epochs function
// optional - show visual of the image
if (show)
{
var theimage = getImage ( img ); // get image from data array
image ( theimage, 0, ZOOMPIXELS+50, ZOOMPIXELS, ZOOMPIXELS ); // magnified
image ( theimage, ZOOMPIXELS+50, ZOOMPIXELS+50, PIXELS, PIXELS ); // original
}
// set up the inputs
let inputs = getInputs ( img ); // get inputs from data array
// set up the outputs
let targets = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0];
targets[label] = 1; // change one output location to 1, the rest stay at 0
// console.log(train_index);
// console.log(inputs);
// console.log(targets);
train_inputs = inputs; // can inspect in console
//my_nn_solution.my_training_function ( inputs, targets );
//my code here
var i = 0;
for(i = 0;i<train_inputs.length;i++){
my_nn_solution.my_training_function(inputs,targets)
}
//my code ends
thehtml = " trainrun: " + trainrun + "<br> no: " + train_index ;
AB.msg ( thehtml, 4 );
train_index++;
if ( train_index == NOTRAIN )
{
train_index = 0;
console.log( "finished trainrun: " + trainrun );
trainrun++;
}
}
function testit() // test the network with a single exemplar, from global var "test_index"
{
let img = mnist.test_images[test_index];
let label = mnist.test_labels[test_index];
// set up the inputs
let inputs = getInputs ( img );
test_inputs = inputs; // can inspect in console
var i = 0;
for(i = 0;i<train_inputs.length;i++){
my_nn_solution.my_testing_function(inputs)
}
let prediction = nn.predict(inputs); // array of outputs
let guess = findMax(prediction); // the top output
total_tests++;
if (guess == label) total_correct++;
let percent = (total_correct / total_tests) * 100 ;
thehtml = " testrun: " + testrun + "<br> no: " + total_tests + " <br> " +
" correct: " + total_correct + "<br>" +
" score: " + greenspan + percent.toFixed(2) + "</span>";
AB.msg ( thehtml, 6 );
test_index++;
if ( test_index == NOTEST )
{
console.log( "finished testrun: " + testrun + " score: " + percent.toFixed(2) );
testrun++;
test_index = 0;
total_tests = 0;
total_correct = 0;
}
}
//--- find no.1 (and maybe no.2) output nodes ---------------------------------------
// (restriction) assumes array values start at 0 (which is true for output nodes)
function find12 (a) // return array showing indexes of no.1 and no.2 values in array
{
let no1 = 0;
let no2 = 0;
let no1value = 0;
let no2value = 0;
for (let i = 0; i < a.length; i++)
{
if (a[i] > no1value) // new no1
{
// old no1 becomes no2
no2 = no1;
no2value = no1value;
// now put in the new no1
no1 = i;
no1value = a[i];
}
else if (a[i] > no2value) // new no2
{
no2 = i;
no2value = a[i];
}
}
var b = [ no1, no2 ];
return b;
}
// just get the maximum - separate function for speed - done many times
// find our guess - the max of the output nodes array
function findMax (a)
{
let no1 = 0;
let no1value = 0;
for (let i = 0; i < a.length; i++)
{
if (a[i] > no1value)
{
no1 = i;
no1value = a[i];
}
}
return no1;
}
// --- the draw function -------------------------------------------------------------
// every step:
function draw()
{
// check if libraries and data loaded yet:
if ( typeof mnist == 'undefined' ) return;
// how can we get white doodle on black background on yellow canvas?
// background('#ffffcc'); doodle.background('black');
background ('black');
if ( do_training )
{
// do some training per step
for (let i = 0; i < TRAINPERSTEP; i++)
{
if (i === 0) trainit(true); // show only one per step - still flashes by
else trainit(false);
}
// do some testing per step
for (let i = 0; i < TESTPERSTEP; i++)
testit();
}
// keep drawing demo and doodle images
// and keep guessing - we will update our guess as time goes on
if ( demo_exists )
{
drawDemo();
guessDemo();
}
if ( doodle_exists )
{
drawDoodle();
guessDoodle();
}
// detect doodle drawing
// (restriction) the following assumes doodle starts at 0,0
if ( mouseIsPressed ) // gets called when we click buttons, as well as if in doodle corner
{
// console.log ( mouseX + " " + mouseY + " " + pmouseX + " " + pmouseY );
var MAX = ZOOMPIXELS + 20; // can draw up to this pixels in corner
if ( (mouseX < MAX) && (mouseY < MAX) && (pmouseX < MAX) && (pmouseY < MAX) )
{
mousedrag = true; // start a mouse drag
doodle_exists = true;
doodle.stroke('white');
doodle.strokeWeight( DOODLE_THICK );
doodle.line(mouseX, mouseY, pmouseX, pmouseY);
}
}
else
{
// are we exiting a drawing
if ( mousedrag )
{
mousedrag = false;
// console.log ("Exiting draw. Now blurring.");
doodle.filter (BLUR, DOODLE_BLUR); // just blur once
// console.log (doodle);
}
}
}
//--- demo -------------------------------------------------------------
// demo some test image and predict it
// get it from test set so have not used it in training
function makeDemo()
{
demo_exists = true;
var i = AB.randomIntAtoB ( 0, NOTEST - 1 );
demo = mnist.test_images[i];
var label = mnist.test_labels[i];
thehtml = "Test image no: " + i + "<br>" +
"Classification: " + label + "<br>" ;
AB.msg ( thehtml, 8 );
// type "demo" in console to see raw data
}
function drawDemo()
{
var theimage = getImage ( demo );
// console.log (theimage);
image ( theimage, 0, canvasheight - ZOOMPIXELS, ZOOMPIXELS, ZOOMPIXELS ); // magnified
image ( theimage, ZOOMPIXELS+50, canvasheight - ZOOMPIXELS, PIXELS, PIXELS ); // original
}
function guessDemo()
{
let inputs = getInputs ( demo );
demo_inputs = inputs; // can inspect in console
let prediction = nn.predict(inputs); // array of outputs
let guess = findMax(prediction); // the top output
thehtml = " We classify it as: " + greenspan + guess + "</span>" ;
AB.msg ( thehtml, 9 );
}
//--- doodle -------------------------------------------------------------
function drawDoodle()
{
// doodle is createGraphics not createImage
let theimage = doodle.get();
// console.log (theimage);
image ( theimage, 0, 0, ZOOMPIXELS, ZOOMPIXELS ); // original
image ( theimage, ZOOMPIXELS+50, 0, PIXELS, PIXELS ); // shrunk
}
function guessDoodle()
{
// doodle is createGraphics not createImage
let img = doodle.get();
img.resize ( PIXELS, PIXELS );
img.loadPixels();
// set up inputs
let inputs = [];
for (let i = 0; i < PIXELSSQUARED ; i++)
{
inputs[i] = img.pixels[i * 4] / 255;
}
doodle_inputs = inputs; // can inspect in console
// feed forward to make prediction
let prediction = nn.predict(inputs); // array of outputs
let b = find12(prediction); // get no.1 and no.2 guesses
thehtml = " We classify it as: " + greenspan + b[0] + "</span> <br>" +
" No.2 guess is: " + greenspan + b[1] + "</span>";
AB.msg ( thehtml, 2 );
}
function wipeDoodle()
{
doodle_exists = false;
doodle.background('black');
}
// --- debugging --------------------------------------------------
// in console
// showInputs(demo_inputs);
// showInputs(doodle_inputs);
function showInputs ( inputs )
// display inputs row by row, corresponding to square of pixels
{
var str = "";
for (let i = 0; i < inputs.length; i++)
{
if ( i % PIXELS === 0 ) str = str + "\n"; // new line for each row of pixels
var value = inputs[i];
str = str + " " + value.toFixed(2) ;
}
console.log (str);
}