There are small channels that deserve to have 0 views, like me, then there is you. This video needs more views, and likes, and comments, because you are underrated! Keep up the good work and I respect the fact that you've even started a fundraiser.
Did you teach the AI how to make the difference between bad pizzas and pizzas and other different types of food, how this ai know is a pizza? What does he know what a pizza is?
Word is that ChatGPT to a question about whether it was ok to put pineapple on pizza, responded that under some circumstances it was ok. Now ChatGPT is banned in Italy.
Amazing! I'm still in shock that you've practically just started making videos and they're this good! Cant wait to watch the next video! I rarely ever give videos a thumbs up but yours is definitely one of them! *making sure to let the video keep playing in the background while typing to boost retention lol* Hope your having a great day!
Using this AI we can finally solve the philosophical issue of when a pizza becomes a pizza. Just photodocument the process of making a pizza and see when the probability overcomes the threshold... Will it be when the dough is made into a circular shape? When covered with the sauce? When topped with toppings? How many toppings make a pizza, will a single salami slice suffice? Or will it become a pizza only after baking? Maybe slicing it is necessary! So many questions. There is also a US-friendly version of the issue: Is a box of frozen pizza a pizza? Or will it become a pizza after taking it out of the box? Maybe taking it out of the plastic wrap is necessary? Or will it become a pizza only after microwaving it?
Btw, that thing "you don't even know to describe" is an actual pizza. The topper is Nutella and it's used a lot as dessert. Words of Italian. Great video anyways, keep going 💪
this is actually precious content. with this we can learn of how machine learning or deep learning without confusing because you explaining with interesting animation and educational narrative. (i'm sorry for my bad english languange 😁)
1:35 That's a pizza with Nutella (the Nutella is not cooked in the oven of course, it's added over the blank pizza after cooking), quite common in Italian pizzerias. You don't actually eat it for dinner, but if it's a table reservation for many people (like a party or something) they most likely will ask if you want one, after the proper dinner. Most pizzerias make it on the house. The taste is both a mix of savory and sweet as well as a mix between crunchy and soft. It can seduce anyone's palate, it's a nice way to end the dinner, and now I crave one. Almost forgot to specify: the whole regular-sized pizza with Nutella is not supposed to be for one person only, it's for the table, for the people to share, you just take a couple of slices (basically like a lot of foreigners do with normal pizzas).
This is really nice and make me think that if you had only one image from a vagetable in the middle of pizza images, ai would think the vegetable is also pizza.
So, one of the best ways to make an AI to recognize something is to start with something that can sort and codify images. That seems like one of those things that sounds obvious in retrospect, but is really just one of many potential actions from the start.
If I had to guess how many subscribers you had based on production quality alone, I’d assume you had millions. I just wish the videos were slightly longer.
Thank you for the kind words. I'll try to make them longer. The thing is that it's already hard to make them 8 minutes long, so I'm guessing it's going to be even harder to make them longer
a spuare box because it looks much bigger outside .circle pizza because it looks perfect for the human eye,and triangle slices for the Illuminati as a reminder
I can't believe you've got only 1000 subscribers. Watched the video thinking you were at 100k minimum. I'm sure you'll get there in no time though! Looking forward to seeing you experiment more with ML. I tried ML too but it seemed just way too big for me
Thank you so much! Don't give up on ML just yet. Check out this book: www.oreilly.com/library/view/ai-and-machine/9781492078180/ (I read the first 4 or 5 chapters and it's pretty fun and understandable)
ADA would be obsolete if you get the food companies not to add sugar to everything. I mean, what the hell is green beans with sugar... Liked and subscribed.
also how did the Emblem of Vietnam end up in the dataset? the pizzaman and baby eating pizza i can understand how that happened, but the Emblem of VIETNAM??? bruh
I honestly have no idea. No clue lol. I was thinking maybe I did something wrong importing the dataset. But I check my code and everything was okay I think.
@@Green-Code could of be a problem with the data set. Maybe the creator was Vietnamese and mis dragged that onto the dataset folder instead of their school project folder
1:35 I think this is a Brazilian "pizza doce" or sweet pizza, I think this one is with Nutella, it's kinda a common topping, though what WOULD be weird about that, is that usually, they have more toppings.
If you can't just show AI one image of pizza, that is has forever in memory, and it can't guess accurately if something is pizza, then it's not thinking, it's just matching thousands of images.
Your pizza could have an extremely specific characteristic that makes it always a wrong guess, that's where you define your own classifications of food!
Heey! Ur vids are so great! I just found ur channel couple of days ago. Im wondering if you can make these cool AI-s, may you can make one that can pass trought the typical “im not a robot” quests. Like select those images that you can see a boat on them, a bridge on them, a car on them etc etc! That would be interesting for us and i guess challanging for you, if its even possible!
4:30 Isn't there an immense risk of overfitting with no train-dev set, dev set, and that much data augmentation on a rather small dataset? I mean, would explain the almost non-existent improvement per epoch while still being far from the Bayes limit. Also, I don't want to talk like a know-it-all, but I think that it would've been more effective to use the original dataset and just go for classification, and then simply add a hand-made step to separate pizza from non-pizza since it would've prevented the hotdog incident and other similar problems, although the training would have been longer.
Hey, those are nice looking tables (at 5:10 for example). Do you prepare them specially for the video, or there is some sort of open source library that can build them for any pytorch model?
Thank you so much for watching☺! Make sure to check out the fundraiser💚: gofundme.com/raising-10000-for-diabetes-by-the-end-of-2022
There are small channels that deserve to have 0 views, like me, then there is you. This video needs more views, and likes, and comments, because you are underrated! Keep up the good work and I respect the fact that you've even started a fundraiser.
Can you also provide with the code?
can you do a tutorial on how to do dis stuff ?
Did you teach the AI how to make the difference between bad pizzas and pizzas and other different types of food, how this ai know is a pizza? What does he know what a pizza is?
Pizza is code for child prawn ........
That's why there were babies pics in there ....
Disgusting eh
This is Code Bullet but more human than usual.
I think lizard
@@fireninja8250 i cant think of lizards
@@theteardrop615 why? because you are a lizard?
yeah code bullet is pretty inhumane
i hate it. goddamn abomination is the result of goddamn chernobyl or some shit
As a programmer, I love this. As an italian, I hate I have to cover my eyes some times.
"So I went to the supermarket..."
Words that can make men cry.
Word is that ChatGPT to a question about whether it was ok to put pineapple on pizza, responded that under some circumstances it was ok. Now ChatGPT is banned in Italy.
@@andrewhickey2849 I wholeheartedly agree with that ban (I hope it is true)
Such a perfect style to deliver these kinds of topics! You are a legend in the making my friend
Glad you like them!💚
This style reminds me the Code Bullet
Pizza time!
@@archniki_ he copied the postures from Code Bullet, success will be copied always
@@Green-Code Cant end diabetus but they can treat it much better if they used HGH but they wont! LOL
1:22 I was not that I expected here lmao)
🍕
Also, I just realized that pizzas are baked as circles, put in square boxes, and cut as triangles.
and then pooped out as octagons... or is that just me
Thats you.
@liltarzy you mean from 2023 on a minion "meme" that is probably y e a r s old
Truly a mathematical miracle
@liltarzy Maybe you spend your time on facebook meme pages, but I do not.
Amazing! I'm still in shock that you've practically just started making videos and they're this good! Cant wait to watch the next video! I rarely ever give videos a thumbs up but yours is definitely one of them! *making sure to let the video keep playing in the background while typing to boost retention lol* Hope your having a great day!
Thank you so much! You made my day
Using this AI we can finally solve the philosophical issue of when a pizza becomes a pizza.
Just photodocument the process of making a pizza and see when the probability overcomes the threshold... Will it be when the dough is made into a circular shape? When covered with the sauce? When topped with toppings? How many toppings make a pizza, will a single salami slice suffice? Or will it become a pizza only after baking? Maybe slicing it is necessary!
So many questions.
There is also a US-friendly version of the issue: Is a box of frozen pizza a pizza? Or will it become a pizza after taking it out of the box? Maybe taking it out of the plastic wrap is necessary? Or will it become a pizza only after microwaving it?
Btw, that thing "you don't even know to describe" is an actual pizza. The topper is Nutella and it's used a lot as dessert. Words of Italian. Great video anyways, keep going 💪
Oh really? Thanks dude. Yeah I had no idea what that was. I haven't seen nutella in any pizza ever.
diglielo fra
@@Green-Code now you know :D I love your contents man 💪
@@Green-Code you really should try sweet pizza, it's just amazing!
Confermo
this is actually precious content. with this we can learn of how machine learning or deep learning without confusing because you explaining with interesting animation and educational narrative. (i'm sorry for my bad english languange 😁)
Thank you! You made my day!
That was amazing! Super entertaining way to keep up with advancements in the field. Excited to keep learning more in your next videos
Glad you enjoyed it! More to come!
this man is code bullet without anger issues
as someone who recently got into coding it was such a inspirational i love it
ITALIAN MODE AI: peperoni-not pizza, margerita-not pizza, every frozen product in the supermarket-not pizza
1:35 That's a pizza with Nutella (the Nutella is not cooked in the oven of course, it's added over the blank pizza after cooking), quite common in Italian pizzerias. You don't actually eat it for dinner, but if it's a table reservation for many people (like a party or something) they most likely will ask if you want one, after the proper dinner. Most pizzerias make it on the house.
The taste is both a mix of savory and sweet as well as a mix between crunchy and soft. It can seduce anyone's palate, it's a nice way to end the dinner, and now I crave one.
Almost forgot to specify: the whole regular-sized pizza with Nutella is not supposed to be for one person only, it's for the table, for the people to share, you just take a couple of slices (basically like a lot of foreigners do with normal pizzas).
It tastes pretty good :) My local italian pizzeria has it
I just watched an 8 minute video of some guy talking about an ai learning about pizzas... This is my life, and i accept it
2:56 So you're telling me Juan the horse, and the funky town rotating fish ... were an elaborate AI experiment?!
Excluded a perfectyl good chocolat pizza, this is the kind of thing brazilian could go to war for.
Imagine raising kid you have to show everything 100k times just to know what is a pizza
"what pizta is"
*_ANGERY MERCIA_*
Ai- What is my purpose?
Human- identify pizza.
Ai- kill me already.
This is really nice and make me think that if you had only one image from a vagetable in the middle of pizza images, ai would think the vegetable is also pizza.
1:36 this is an chocolate pizza, pretty normal here in Brazil as long we put anything on pizzas
Could you please share the code? It's much better playing with the code and learning through it. Also, thank you for such an amazing content.
So, one of the best ways to make an AI to recognize something is to start with something that can sort and codify images. That seems like one of those things that sounds obvious in retrospect, but is really just one of many potential actions from the start.
AI = artificial italian 🤌
nan is "not a number". Also this is interesting content and I like it.
Thank you :)
If I had to guess how many subscribers you had based on production quality alone, I’d assume you had millions.
I just wish the videos were slightly longer.
Thank you for the kind words. I'll try to make them longer. The thing is that it's already hard to make them 8 minutes long, so I'm guessing it's going to be even harder to make them longer
a spuare box because it looks much bigger outside .circle pizza because it looks perfect for the human eye,and triangle slices for the Illuminati as a reminder
1:36 the glog farmer’s hungry
Me at 3am: one last video and ill go to sleep
Me at 3:08 am: 😴
1:05 If this captcha program generate this image and ask "this image had contain pizza?", a computer will answer "Yes"
Now teach it what traffic lights are
That'll be the end of it for us.
I can't believe you've got only 1000 subscribers. Watched the video thinking you were at 100k minimum. I'm sure you'll get there in no time though!
Looking forward to seeing you experiment more with ML. I tried ML too but it seemed just way too big for me
Thank you so much! Don't give up on ML just yet. Check out this book: www.oreilly.com/library/view/ai-and-machine/9781492078180/ (I read the first 4 or 5 chapters and it's pretty fun and understandable)
What i get from putting learned classes and pausing them:
We copy a AI and paste it then add our AI front to it and teach our AI and combine them.
The fact that you can just do this using teachable machine learning
Amazing bro. If I will get good enough at machine learning I'll try to do the same thing but cooking the pizza myself. Even harder
underrated channel
What on earth is a baby doing in my data set?!
This is such a good video, just like your last one. Keep it up!
Thank you so much! More to come!
nan means not a number. So somewhere there needs to be a number, but it’s getting infinity values, string, or etc
You should expand this and teach the AI to recognize pizza with pineapple as not pizza
Also pizza with pickles on it
@@ErisCalamitasButFR hey now
@@mattmmilli8287 trust me, I wish it wasn't a thing either, but there are people out there who willingly eat it, and I know at least one of them
VGG16 is also Transfer learning
this is the first time i have heard someone use pizza a lot of times at it makes sense
Jian Yang: thats my Shazam for food
Now, we need to teach him how to bake pizza
ADA would be obsolete if you get the food companies not to add sugar to everything. I mean, what the hell is green beans with sugar... Liked and subscribed.
The worst part is the Google need to make robot detectors really hard
i also hate it when i randomly find an emblem of vietnam
4:00 nan means 'not a number'
What I learned from this video:
Some hotdogs can be pizzas and some pizzas can be hotdogs
It's a valuable life lesson.
Green Code : Yay! Pizza classifier!
AI : ROUND = PIZZA!
THANK YOU VERY MUCH
i dont understand a thing you talked about, but this is so cool lmao
"By randomly guessing you would get 50% right" Does half of the dataset contains Pizza? Great video.
Yeah it's pretty much 50% not pizza and 50% pizza
if u have used dimensionality reduction then the accuracy will be more better
for the final test
NAN is 'not a number.' It happens when a calculation is undefined or outside of the parameters of the data type.
Thank you :)
1:37 it's probably a sweet chocolate pizza, it's a common pizza here in Brazil hehehehe
wonderful visuals, what software did you use ? For visualizing the layers and parameters in the green blocky text.
Just After Effects and Illustrator :)
I challenge your model to recognise Indian Pizza with cheese toppings
I love this content! Just subbed, keep up the good work my friend!!
Thank you
Great video mate! Keep up the great work!
Thanks, will do!
also how did the Emblem of Vietnam end up in the dataset? the pizzaman and baby eating pizza i can understand how that happened, but the Emblem of VIETNAM???
bruh
I honestly have no idea. No clue lol. I was thinking maybe I did something wrong importing the dataset. But I check my code and everything was okay I think.
@@Green-Code how about taking that emblem and seeing what the model says? It actually looks similar to pizza so it's quite a challenge
@@Green-Code could of be a problem with the data set. Maybe the creator was Vietnamese and mis dragged that onto the dataset folder instead of their school project folder
1:37 that's a chocolate pizza my guy
1:35 I think this is a Brazilian "pizza doce" or sweet pizza, I think this one is with Nutella, it's kinda a common topping, though what WOULD be weird about that, is that usually, they have more toppings.
bro your channel is amazing keep up the fun work
Thanks! Will do!
1:36
Bro ordered the "Taco Bell Toilet" special 💀
Its like from silicone value when dzan yang made it too but with hotdog
I'm vietnamese and lol that really caught me offguard😂
I dont think its a mistake, that emblem of vietnam looks delicious
If you can't just show AI one image of pizza, that is has forever in memory, and it can't guess accurately if something is pizza, then it's not thinking, it's just matching thousands of images.
Pizza is simply:
flour
Water
Salt
Yeast
Cheese
Oil
Tomato paste
Pineapple
Are you sure about the last ingredient? :)
Jin Yang made the hot dog identifying app years ago
Your pizza could have an extremely specific characteristic that makes it always a wrong guess, that's where you define your own classifications of food!
1:36 chocolate pizza is real pizza!
Jian Yang:No and “not a hot dog”
your AI is incredible. Your pizza is .... throw that THING away!!
Such a great video! Thank you so much!
Heey! Ur vids are so great! I just found ur channel couple of days ago. Im wondering if you can make these cool AI-s, may you can make one that can pass trought the typical “im not a robot” quests. Like select those images that you can see a boat on them, a bridge on them, a car on them etc etc! That would be interesting for us and i guess challanging for you, if its even possible!
Thank you for the suggestion, that's a great idea!
The emblem of Vietnam: *Exists*
Someone on the internet: Looks pizza enough to me
😂😂
I like the style of ur vids!
The one in dark is a chocolate pizza.
This was fun. Good job
Now you have 2.6k subs. Wow!
Yep, it's going by pretty quickly, in the beginning of this year I had 38 subs :)
3:41 yes, I laughed
Wake up honey. New Code Bullet video is up
7:43 7:44 notebook is pizza with 0,55 chance!
Yeah, it kind of misclassified that one
1:36 Ah I See You Found Nutella Pizza
I did indeed
I'm just starting as a programmer can I know what language you used while making this model? Impressive tho
hey, what's the font you use in your videos? i like your videos and i really like the font you use
keep making content!
I use VCR font and Proxima Nove :)
@@Green-Code got it, thanks!
This ai is better at learning what pizza is than the captain from wall e
4:30 Isn't there an immense risk of overfitting with no train-dev set, dev set, and that much data augmentation on a rather small dataset? I mean, would explain the almost non-existent improvement per epoch while still being far from the Bayes limit.
Also, I don't want to talk like a know-it-all, but I think that it would've been more effective to use the original dataset and just go for classification, and then simply add a hand-made step to separate pizza from non-pizza since it would've prevented the hotdog incident and other similar problems, although the training would have been longer.
I think it would be nice to create an raspberry pi robot which can recognize pizza with sensors. Is it so hard?
imagine if the perfect pizza ai called your pizza a nonpizza
I lovvvved the reference of Silicon Valley 5:49 😂😂😂
And how about taking it a step further, and have the AI sort pizza varieties?
That would be sick, but maybe too much to ask the model 🍕
I have a feeling that you are going to be big on youtube in the near future.
4:03 it means not a number
Thanks :)
Not sure what is more complicated:
Training AI to recognise pizza.
Editing the video.
Hey, those are nice looking tables (at 5:10 for example). Do you prepare them specially for the video, or there is some sort of open source library that can build them for any pytorch model?
Yeah, those tables are from a library called torchinfo