Dan you are a legend. It has been a pleasure watching your videos evolve over the years. Hugely valuable learning resource from someone who cares a bit more about aesthetics. THANK YOU and Cheers!
I am really glad that we have someone like you in youtube, i am wondering if i can challenge you with some coding with matter.js, can this plugin add physics to text and letters ? would it be able to add physics to it ? for example a circle hitting the text and it would explode ? i would really really appreciate it if you can look into it :)
It would be really good to see you do a little project in Rust, teaching as you learn. Its learning curve is tough and current video intros are either too shallow or too dry. Your personality and style would be a great help!
Before you said ml5 has Function for flipping I was just thinking that you're gonna flip the snake move value to left for right label and right for left label.. but then I realized how lazy am i xd
it'd be interesting if you gave the AI classification the screen outputs as the pictures and then a recommended move as the classification, might come up with an AI that plays XD
Can someone please explain to me this concept : For example: say we have a PVector with a location embedded into, let's call it position, this position (vector) belongs to a different class within a another class say we have a class object, denoted by Class class b { PVector position; Class a = new Class(position); And, Class a = new Class(new PVector(position.x,position.y); } How come, when I run the first one, the orginal object (b) get's affected by what happens to class (a) but when I run the second one this case doesn't happen I thought left = rights and right doesn't matter Sorry if I am dumb.
Amazing... I have learned a lot thanks to you and your tutorials. Dan, could you please do a video like "11.3 Neuroevolution Flappy Bird with TensorFlow.js" but with the "Snake game". I've tried for myself but I can't get the snake to learn.
Question: When I add my own model, it reads my input just fine, but the snake doesn't move. I've changed literally nothing but the model link. Am I doing something wrong?
i hav no idea wut running a photo through the toy neural network would be like. but i know the computer vision series is pretty close to a doodle. i know again probably speaking very soon on this topic
Dan, thanks for the amaazzzing video! When I tried to train the model on Teachable Machine using "Posture", it seems p5.js is not supporting. Is it true?
Great video, thank you a lot. I just have 1 question/problem. I get: "High memory usage in GPU: 1186.54 MB, most likely due to a memory leak " warning. How can I fix this?
I am not able to download my model for TensorFlowLite as Floating? I click on export model and then click the download button, it changes to "converting your model" and then back to "download your model", but it does nothing and my model doesn't download? Is it just me or is this happening to you or anyone else too? I need to download the model for TensorflowLite
for what i understand in the code, the classifyVideo() calls gotResults(), witch in turn calls classifyVideo() , making a infinite recursion. If this script was in python, it would break because it would research the recursion depth limit. Why in javascript works? i mean every ocurrance of a function ocupates a space in the memory and the code is adding more and more of that, never ending one ocurrance, sometime this would full the stack.
Is there a way to get the Teachable Machine to convert poses into keystrokes and mouse clicks? It would be awesome to control my favorite PC video games with gestures.
Great idea! You could send network messages from p5 to some piece of hardware or software that emulates key presses etc! Maybe it would be easier from python tensorflow directly?
I ran into error which shows "InvalidStateError: Failed to execute 'drawImage' on 'CanvasRenderingContext2D': The image argument is a canvas element with a width or height of 0." Anyone knows why?
Great, 1)How can I run the model locally without internet connection? 2)How can I add my exported model to my react native or xamarin mobile app ? Thanks
@thecodeingtrain how can i use ipcam in place of webcam of pc?? because i have no webcam, so i want to use my mobile cam as webcam, so what argument can i pass in createCapture() function so i cam use url in that function and can access my mobile cam like webcam?? please answer
@@TheCodingTrain but there is no any option for place url as a parameter in createCapture() function?? Because of this is easy and best way to utilize that.
@@TECHNOSTARTERSS it gave me 30fps 720p with webm codec, so have even tried encoding a camera to streaming video over IP? If not try doing with a raspberry Pi for starters. There are many products today which streams at 1080p like baby monitors, security systems and worst of all check Facebook live specs! I worked on networking and I'm not that aware of JavaScript if you are aware of the solution let me if not let the expert do the talking isn't that why I comment here!
Dan you are a legend.
It has been a pleasure watching your videos evolve over the years.
Hugely valuable learning resource from someone who cares a bit more about aesthetics.
THANK YOU and Cheers!
ML wouldn't have been the same without these Coding Train videos on YT :)
Agreed
Me: Finally gets it to compile.
Me: 07:10
Awesome...please keep going on this subject! It is really fascinating and open so many possibilities!
Best teacher ever!
Probably the best series on the channel! 😍
This is the coolest thing I have ever seen on you tube to date!!! love it!!!!
great video! As always! Simply love watching your creative ways of coding something new
You are the hero we need, but don't deserve, when it comes to programming
I love your teaching.
I am really glad that we have someone like you in youtube, i am wondering if i can challenge you with some coding with matter.js, can this plugin add physics to text and letters ? would it be able to add physics to it ? for example a circle hitting the text and it would explode ? i would really really appreciate it if you can look into it :)
It would be really good to see you do a little project in Rust, teaching as you learn. Its learning curve is tough and current video intros are either too shallow or too dry. Your personality and style would be a great help!
You are tiger 🐅 of programming.... your all videos are great 👍
Before you said ml5 has Function for flipping I was just thinking that you're gonna flip the snake move value to left for right label and right for left label.. but then I realized how lazy am i xd
Teachable Machine: "My left is your right. Isn't that confusing?"
Oh boy that was just the greatest thing one could ever teach on youtube. Thanks
And I thought this was about an AI that learns how to play Snake.. :P Good Video anyways!
CZcams user Code Bullet has a video about an AI that learns to play snake.
it'd be interesting if you gave the AI classification the screen outputs as the pictures and then a recommended move as the classification, might come up with an AI that plays XD
@@johnydl gg
Codebullet is not a programming channel
More like heres something I have done using these techs and thats it
KuroKaktus Codebullet is not a programming channel
More like heres something I have done using these techs and thats it
Bestest teacher ever
brilliant and creative demo. very impressive for me.
Thank you Mister,
You are an amazing one who each time gave me tons of ideas, thank you for that and the rest.
Best regards
Why would you flip the whole video? Just flip the right and left output...
Exactly.
Why even do that just train model accordingly
Those were Quickest 7:58 minutes of my life. I was like video is over?!?!?!? ALREADY?!?!?
Cool but why did you flip the video instead of changing the output.
Thats really taking performance in the long run
Can someone please explain to me this concept :
For example:
say we have a PVector with a location embedded into, let's call it position, this position (vector) belongs to a different class within a another class
say we have a class object, denoted by Class
class b {
PVector position;
Class a = new Class(position);
And,
Class a = new Class(new PVector(position.x,position.y);
}
How come, when I run the first one, the orginal object (b) get's affected by what happens to class (a) but when I run the second one this case doesn't happen
I thought left = rights and right doesn't matter
Sorry if I am dumb.
I came to see a snake AI. I saw something that is even more fun and exciting...😄
Thanks for the great lessons!
thanks you are amazing
Thank you so much for so many great videos
Amazing! Keep up the good work!
Amazing... I have learned a lot thanks to you and your tutorials.
Dan, could you please do a video like "11.3 Neuroevolution Flappy Bird with TensorFlow.js" but with the "Snake game". I've tried for myself but I can't get the snake to learn.
Been waiting for this, Dan.
the video editing seems to be better and great video btw
Wow Amazing...
And I am here again !!!
Have a niiiiice dayyyy!!!
what microphone do you use for your videos?
Thanks for the sneak peek!
Love you always
wow that was really fun to watch
Question: When I add my own model, it reads my input just fine, but the snake doesn't move. I've changed literally nothing but the model link. Am I doing something wrong?
same!!!
Fantastic! Tanks!
6:35 not using a switch statement...
😱
@@Robin-fp5xr agreed...
I personally find if...else statements more readable...
It is the same way if you use switch
Stunning!!!!
Now I'm a big fan of you
Amazing! Don't stop ))
i hav no idea wut running a photo through the toy neural network would be like. but i know the computer vision series is pretty close to a doodle. i know again probably speaking very soon on this topic
Hey can you tell me what software you use to record your tutorials and for the green screen
very cool thank's
This is awesome.
Dan, thanks for the amaazzzing video! When I tried to train the model on Teachable Machine using "Posture", it seems p5.js is not supporting. Is it true?
Great video, thank you a lot. I just have 1 question/problem. I get: "High memory usage in GPU: 1186.54 MB, most likely due to a memory leak " warning. How can I fix this?
👑
Thank youuuu!
You're a wizard
This is so cool
Brilliant!!
Instead of flipping make the left hand as right and vice versa
You have plus 1 in the exam 😆
How can we access predictions generated in p5.js from java or processing? Thanks in advance.
I am not able to download my model for TensorFlowLite as Floating? I click on export model and then click the download button, it changes to "converting your model" and then back to "download your model", but it does nothing and my model doesn't download? Is it just me or is this happening to you or anyone else too? I need to download the model for TensorflowLite
Hahaha that was so nice and funny😂😍
Hi, I am getting many errors on the code please can you help me. I am using this project for school purposes.
Error: Error when checking : expected dense_Dense1_input to have 2 dimension(s), but got array with shape [1,224,224,3]
How about a version that trains model using image of snake game video state so it can play itself based on relative location to food?
I tried running the same but on Glitch. It doesn't seem to work?
Awesome 🙂
Awesome 😍
So usefull!,,
for what i understand in the code,
the classifyVideo() calls gotResults(), witch in turn calls classifyVideo() , making a infinite recursion.
If this script was in python, it would break because it would research the recursion depth limit. Why in javascript works? i mean every ocurrance of a function ocupates a space in the memory and the code is adding more and more of that, never ending one ocurrance, sometime this would full the stack.
It works because it only calls classifyVideo() again once the first one has finished! (and it takes some time for each classification.)
Is there a way to get the Teachable Machine to convert poses into keystrokes and mouse clicks? It would be awesome to control my favorite PC video games with gestures.
Great idea! You could send network messages from p5 to some piece of hardware or software that emulates key presses etc! Maybe it would be easier from python tensorflow directly?
Hello shifftman. I challenge you to replicate the marangoni effect in JavaScript
I ran into error which shows "InvalidStateError: Failed to execute 'drawImage' on 'CanvasRenderingContext2D': The image argument is a canvas element with a width or height of 0." Anyone knows why?
what programm is he using for coding?
Hello,chat can any one recommend me which language to start?with plzz
Here's my playlist for getting started! czcams.com/play/PLRqwX-V7Uu6Zy51Q-x9tMWIv9cueOFTFA.html
Python is quite nice to start but i am very biased since it's the main one i know, advices from other people would be nice
wow
you can make it by pose project
Great,
1)How can I run the model locally without internet connection?
2)How can I add my exported model to my react native or xamarin mobile app ?
Thanks
The discord is a great place for this discussion! thecodingtrain.com/discord
Please do a bubble shooter or match 3 game in a coding challenge
Share it in description
some tool for the reinforcement learning?
Since it trains locally should I have a GPU laptop
If not is there any backdoor to train this on Google Colab?
You don't need a fancy GPU for this! It works on most modern laptops.
Cool
Genius
@thecodeingtrain how can i use ipcam in place of webcam of pc?? because i have no webcam, so i want to use my mobile cam as webcam, so what argument can i pass in createCapture() function
so i cam use url in that function and can access my mobile cam like webcam?? please answer
Try going to your browser settings and changing the default camera! If your browser can see it then TM and p5 can as well.
@@TheCodingTrain but there is no any option for place url as a parameter in createCapture() function?? Because of this is easy and best way to utilize that.
Sadly it only works with my face, the model needs to be trained by other people who you want to share it
cool
Am I'm allowed to be here 0_0 lol sneak peek
You are!
I am interested to learn coding
that's really cool sir! but how to feed in an IP camera wirelessly
@@TECHNOSTARTERSS but the question still is will it be possible to do it? Even 5 fps would meet latency of a snake game wouldn't it!
@@TECHNOSTARTERSS it gave me 30fps 720p with webm codec, so have even tried encoding a camera to streaming video over IP? If not try doing with a raspberry Pi for starters. There are many products today which streams at 1080p like baby monitors, security systems and worst of all check Facebook live specs! I worked on networking and I'm not that aware of JavaScript if you are aware of the solution let me if not let the expert do the talking isn't that why I comment here!
@@TECHNOSTARTERSS so u r saying this is impossible!
@@TECHNOSTARTERSS cool! That's what i wanted also lets hear what the expert has to say!
Thank you 🙂I am 4years late 😞
What is name of the app that he wrote this code ?
it's the p5 web editor! editor.p5js.org/
You in the beginning forgetting that you are already flipped correct yourself to showing the wrong side
Can u give the code of ur snake game please
You can find everything here! thecodingtrain.com/TeachableMachine/2-teachable-game.html
Awesome..
Ok i am really early
arm language at my computer. i hope none of the neighbors are watching