7.6: TensorFlow.js Color Classifier: Training Data Tensors (inputs)

In this video, I convert the data from the JSON file to tensors focusing on the inputs first.

[bijilkp]

Iam a big fan of you sir .in this time am understand that am verry late 2 find u .you have an excellent skill and a working Brain.

[giorgosd]

Why is normalization to range [0, 1] needed?
Every data in the dataset is already on the same scale of [0, 255]. Or is it required because output nodes will be [0, 1] anyway? Or does it make initial random weights more accurate?
And also the weights of the input layer could just do this normalization by multiplying inputs by 1/255

[egor.okhterov]

Hmm... I understood everything in this video.

[TheoParis]

the two dislikes are the people who hate AI

[BlessyGasagara]

Anybody know where we can find the JSON data set ?

[David]

what is that little squirrel on your atom?

[matze]

why exist this video on youtube but not on your channel??

[stefomc]

With what programming language to start? I mean one of easiest. I know HTML

[pablocalofatti]

hi guy's, i have this error
Uncaught (in promise) TypeError: data.entries is not iterable
at setup (sketch.js:14)
at m.<anonymous> (p5.min.js:3)
at _runIfPreloadsAreDone (p5.min.js:3)
at m._decrementPreload (p5.min.js:3)
at p5.min.js:3

and here its my code,
//globals
let data;


//preload the json with the dataset that i have all ready made in firebase with my other page
function preload(){
data = loadJSON('cdc.json');
}

function setup() {


let colors = [];
for (let record of data.entries) {
let col = [record.r/255, record.g/255, record.b/255];
colors.push(col);

}
console.log(colors);
}
im stuck here, the inspector says that the error its in for (let record of data.entries) {

[stefomc]

But i dont like Scratch because thats not coding thats puzzle