Coding Challenge #128: Sketch-RNN Snowflakes with ml5.js

preview_player
Показать описание


References:

Videos:

Related Coding Challenges:

Timestamps:
0:00 Introduction
0:55 SketchRNN
2:49 ml sketchRNN module
4:02 Let's Code!
4:29 Create model
6:31 Stroke object
9:14 strokePath
13:08 Ask for the next point
14:02 Pen up or down
16:45 Check for pen ending
18:25 Snowflakes!
19;46 Scaling the snowflakes

Editing by Mathieu Blanchette
Animations by Jason Heglund
Music from Epidemic Sound

#sketchrnn #machinelearning #quickdraw #ml5js #p5js
Рекомендации по теме
Комментарии
Автор

Excuse me, butI don't think those are snowflakes....
#snowfakes

nikitabeknel
Автор

I think you must reset pen to down before starting a new sketch. Awesome video!

ciscou
Автор

awesome, I am Arabic and I am followed you thanks for you a lot

hussainsalih
Автор

Wait, wait, wait... Hold up a moment... You're just glossing over the fact that you're doing RNN in real time in javascript?

You're just like "Oh yeah, we got this super stable Java-based Processing right here, and then we got the notoriously light-weight JavaScript version, meant to be little more than something to make the web slightly more fancy than CSS... We slapped a NEURAL NETWORK INTERFACE onto the JavaScript library. No big deal."

morphman
Автор

Interesting question about legality here... If anyone were to decide to print these snowflakes onto a shirt, or something like that, who would own the copyright?

The Processing Foundation, for having the code generating the snowflake? The programmer who created the code calling ML5.js? Google for making Sketch-RNN? The collective people who used Google Quickdraw to draw the training data? The computer that drew it? Or would it be public domain?

I know this won't be answered here, it would take an actual court case to answer the question, and even then the answer would only apply to that particular court's jurisdiction. It's just one of those quasi-philosophical questions I get sometimes when I watch videos like this.

morphman
Автор

i opened the video because I was interested about this next snowflake tutorial (just yesterday watched the first one) and 90% of the video was cat.... I'm so happy <3

catlord
Автор

where can i get your merch ?
im loving that pullover of yours

ivansakal
Автор

Yo, i just figured out ml5 are directing the users to your channel in the documentation. Are they sponsor you? Not that it's a bad thing just wondering

kshayk
Автор

That bell XD XD I need one for those kind of moments

ianborukho
Автор

What would be your strategy for putting these into objects in order to animate them (move them around the canvas different ways)? I was thinking save the line data to an array of arrays as it comes in and then use that as part of the constructor... thoughts??

angelezamarron
Автор

Daniel, maybe one day you could do a 10 min video on how to use ML5 combined with P3, it is hard when trying to establish such. (basic steps, as which files in which libraries, adjust code to read correctly etc)

d_v_d
Автор

I love this video and I love the end result!!

angelezamarron
Автор

I absolutely love your videos they are amazing and I learn so much from every video. Keep up the great 👍 work

dominicrann
Автор

are the majority of your videos in javascript? is that your language of choice?

wesmoto
Автор

i don't get one important thing - is it the case here that all drawn cats/snowflakes are not just plain (but exact) redrawings of one random sample from the google doodle dataset, but there is some sort of creativity added to each newly drawn cat/snowflake so they will always be unique?

PierreLaBaguette
Автор

where do you get your ideas for the coding challenges? I mean, is there a page where you find them or you just get the idea somehow. I am starting to code and i eould like to try easy challenges by myself but I dont find any.

DaminGamerMC
Автор

How about training the model with results of the first video?

rafalfaro
Автор

Omg. The object model is used inside the function which serves as a callback that is a part of the initialisation of the object model. #spaghetti!!!!

grainfrizz
Автор

So why does his code run at a really smooth consistent pace. While mine is looping as fast as my pc can loop it? :(

coltpeterson
Автор

Probably did you want to say: "Guacamole".

arturoaragon