I tried to build a ML Text to Image App with Stable Diffusion in 15 Minutes

Подписаться 266 тыс.

Просмотров 110 тыс.

50% 1

What's happening guys, welcome to the sixth episode of CodeThat!? I think
So there's been a lotta talk about text to image generation using machine learning...well I thought I would give it a crack and try to whip up a Python App that could do it.
And along comes Stable Diffusion, an open source SOTA model for text to image generation from Stability AI, well the rest is history. In this ep we focus on using the Stable Diff model through Hugging Face Diffusers and build up an app using Tkinter so that you can use it on your desktop with nothing more than a 4GB GPU.
Get the Code: github.com/nicknochnack/Stabl...
Oh, and don't forget to connect with me!
LinkedIn: bit.ly/324Epgo
Facebook: bit.ly/3mB1sZD
GitHub: bit.ly/3mDJllD
Patreon: bit.ly/2OCn3UW
Join the Discussion on Discord: bit.ly/3dQiZsV
Happy coding!
Nick
P.s. Let me know how you go and drop a comment if you need a hand!
#machinelearning #codingchallenge #tkintertutorial #python

Наука

Опубликовано:

19 сен 2022

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 329

@pon1 Год назад

I've wanted to make a Stable Diffusion UI ever since the model became public, but didn't know where to start as I'm not used to Python, but now you've given me a great start! Thanks!

@NicholasRenotte Год назад

Code it here team: github.com/nicknochnack/StableDiffusionApp

@NicholasRenotte Год назад

@@ShineSincere check the github repo, I updated it to include the file. You just need to get your token from Hugging Face hub

@rohitchan007 Год назад

Hey Nick. You're my favourite RU-vidr of all time. Thank you so much for sharing your knowledge with us. I have learnt so much from you.

@NicholasRenotte Год назад

Oh that means a ton to hear @Rohit! I'm so glad you're enjoying it!

@Quantumvp Год назад

@@NicholasRenotte Hi Nick I have tried to complete the project but when i get to the end and ready to deploy I get this raise EnvironmentError( OSError: There was a specific connection error when trying to load CompVis/stable-diffusion-v1-4: (Request ID: eYm7eENfp-7CbfSsEEf53)..

@rakeshkumarrout2629 Год назад

hey nick thank you for bringing content like these .This is really so helpful .How to learn everything about diffusion models.How it works.tech behind it and how it works.make a small diffusion model from scartch.

@patricklynch6871 Год назад

Awesome work as always, Nich!

@nazarzaki44 Год назад

I enjoy your videos Nick! Thank you so much for the hard work and sharing

@atunraseayomide2504 Год назад

You have to be my best youtuber yet. You push me to do more daily, you are an inspiration Nicholas !!

@NicholasRenotte Год назад

That means a ton @Atunrase!! Keep at it!!

@atunraseayomide2504 Год назад

@@NicholasRenotte 😁💯💯

@anirbanpatra3017 Год назад

Is there a way to deploy this project??

@Don69Muk 10 месяцев назад

I like how you deliver the content. Good job dude

@einsteinsboi Год назад

Woohooo! You did it!!! AMAZING!! Super impressive work Nick!

@NicholasRenotte Год назад

WE MADE IT MARY!!!! 😂😂

@harshpanchal7412 Год назад

Such a Great idea and execution sir, big help for the project 🙌✨

@gabrielvalentim197 Год назад

Hey Nick. Thank you very much for your content. I really like your videos and they helps me understand more and more about ML!

@NicholasRenotte Год назад

I'm so glad you're enjoying them @Gabriel!

@egemenaksu Год назад

hey Nick! I believe that the code link in the description is a link to another project of yours. Absolutely amazing video as always, keep up the good work. Cheers!

@NicholasRenotte Год назад

Updated, thanks a mil @Egemen!

@NoahElRhandour Год назад

wow thats an awesome idea for a video !!! PS: you are a really great youtuber i enjoy your content a lot

@NicholasRenotte Год назад

Thank you so much @Noah!!

@rayaneghilene5152 Год назад

This is very entertaining and exciting content, thank you so much for sharing such great work with us

@fustigate8933 Год назад

Thanks for looking into this particular field!

@NicholasRenotte Год назад

Anytime @Fustigate!! glad you enjoyed it!

@hanemellethy9036 Год назад

That's awesome. I like the video and learning more from you. Thanks for that wonderful video. I share your videos with my students 😀 keep it up. Thanks Nicholas

@NicholasRenotte Год назад

Thanks so much @Hanem!!

@varunvijaywargi5497 Год назад

Hey Nicholas, I have been away from AI since almost a year now but this video is sort of pushing me to start looking back into it. Awesome one mate !! Let us catch up for a coffee some day in city.

@NicholasRenotte Год назад

VARUN! Hell yes, whenever you're ready let's do it!

@jacob4316 Год назад

@@NicholasRenotte What city is that? I'd like to link up and discuss a business opportunity. I need a good dev!

@PrabhuKiranKonda Год назад

Woah!! That's incredible. Great as always.

@NicholasRenotte Год назад

Thanks a mil @Prabhu!

@aimanifest Год назад

I absolutely love your videos my friend. You’ll be absolutely BLOWN away at the implementation this AI can achieve. My page has over 10 animations created with SD, and if you’re not interested, just check out all the other pages that use SD to create animations. What a world we live in. Truly insane!

@NicholasRenotte Год назад

Sweet!!

@selcukozdemir8941 Год назад

That was really inspiring. I learned a lot from this video and I hope everyone else do too.

@trevorsatori Год назад

Love your content, keep it up brotha 💎

@NicholasRenotte Год назад

SATORI DIGITAL!!! Thanks a million for checking it out :)

@sapnilpatel1645 Год назад

The video is amazing. learned a lot. Thanks for making it.

@saviruedirisinghe3751 23 дня назад

Thank you so much for sharing your knowledge with us. I have learn so much from you.

@SimonLam2024 Год назад

I enjoy this very much. I didn't know we can Text To Image in this way.

@NicholasRenotte Год назад

🙏🙏🙏 pretty cool right?!

@mrziedaan Год назад

He Nicholas! Thanks for your great videos, it would be awesome if you made a video about making your own difussion model.

@chienlingliu Месяц назад

well done ! thank you for sharing this !

@jojo23srb Год назад

Hey Nick, awesome tutorial! Do you know if it's possible to do this same thing using Intel CPUs? I've tried specifying 'cpu' as my device but am getting some errors.

@alecubudulecu Год назад

This is awesome. But I agree with some of the comments that I wish you would slow down and just reach this stuff. It’s awesome and we can follow along. But while you are an awesome coder … your an even better educator

@ahmetcals3651 Год назад

Yeahh finally you did it! I'm curious about where you learn all bunch of stuff? What are your recourses? How long preparation it takes to make a tutorial video for example?

@NicholasRenotte Год назад

A lot of it is just random googling and reading! I normally give myself one to two days to learn the code and then a day to do the challenge and record.

@ahmetcals3651 Год назад

Keep going what you doin! Learning a lot from you :)

@yusufhenry9365 Год назад

Bravo Brave coach!! Some of us might be needing a slow step by step explanation, so as to be on the same page

@NicholasRenotte Год назад

🙏 🙏 🙏 will probably do a walkthrough on a live stream!

@yusufhenry9365 Год назад

@@NicholasRenotte that would be much appreciated by the community Thank you Sir!

@yusufhenry9365 Год назад

@@NicholasRenotte that would be much appreciated by the community Thank you Sir!

@youdontknow8729 Год назад

BRO I HAVE FOUND DIAMOND WHILE I WAS SEARCHING FOR GOLD ---- THANK YOU FOR THE GREAT CONTENT

@NicholasRenotte Год назад

Ohhhhhh man, thank you soooooo much!!

@AIFocus01 Год назад

Thanks for this great content.Which version of python was used for this project?

@juangabrielvillamizar9871 Год назад

You are great, God bless you and continue to give you wisdom

@-_BahauddinTaha Год назад

Hey, Nick, Great Video. Learned something new! I've a question, Is there any way to do it, if you don't have a GPU support on your pc???

@NicholasRenotte Год назад

You can definitely do it, it'll just take a lot longer to generate!

@StarDust1510 Год назад

This made my day!!! Thanks a lot!

@NicholasRenotte Год назад

Thanks a mil for watching it @StarDust, ya made my day tuning in!!

@michaelzumpano7318 Год назад

That was exhilarating!

@alicanakca8116 Год назад

Love these challenges!

@NicholasRenotte Год назад

🙏🙏🙏

@SAHIL.SINGHYT Год назад

good work king, love you

@aimanifest Год назад

Also, would be interested in a discord bit code that episode! lol

@viveknath6073 Год назад

Woh You are Amazing And your content is massive ❤ everyone is Impressed When you challenge yourself for building new things in less time with accuracy I want this type of confidence to improve myself and Do something new 😌thanks Nic for providing a lot of knowledge craziness about ML and AI ❤️😌 it will help us

@NicholasRenotte Год назад

I'm so glad you enjoyed it @Vivek!! Also love that you love my slight craziness LOL, gotta keep it fun!!

@viveknath6073 Год назад

@@NicholasRenotte yehh bro please provide this type contant i love it brooooh

@ramrayavarapu7321 Год назад

Astonishing brother, learned many things, Could you please share how you got the image text pairs dataset Also please tell me which one is better STABLE DIFFUSION or GAN

@bartoszhojnacki2233 Год назад

Congrats on completing it! Another cool video. I think you are a bit harsh to yourself - i mean loading the model etc shouldn't "consume" the time :D

@NicholasRenotte Год назад

Hahahaha, true, I think I just get too hyped up in the moment and forget to pause and do stuff. I completely forgot to turn on the RGB lights on my desk as well normally they're the red and blue ones from prev code that challenges

@amineproductions3055 Год назад

i love bro i hope you still make vidois in future more an more thank you

@VenkyBeast Год назад

Amazing! I almost thought that you gonna lose this challenge. lol.

@NicholasRenotte Год назад

LOL for once I was like nottt todayyy!

@rickyroxas3130 Год назад

Can you make a Traffic Sign Detection video next time sir? I really appreciated all of your works. Thank you.

@NicholasRenotte Год назад

You got it!

@suyashwagade3738 Месяц назад

nice one!!

@webdesignsbytom 6 месяцев назад

amazing work.

@mife-3200 11 месяцев назад

Hi) Did you download the AI model beforehand or you just connected with API?

@Dujana-2018 Год назад

Hey Nick you are awesome man. I have learned so many crazy things from you. Can this will run on CPU?

@NicholasRenotte Год назад

Thanks a mil @Daud! Sure can it'll just take longer to generate.

@wgb-10 Год назад

Amazing stuff man. You did it within the time frame! Let's celebrate with an amazon gift card 😂

@NicholasRenotte Год назад

I was soooooo tempted to just give it away for fun @Wesley!! Might add it to the pot for the next one?!

@wgb-10 Год назад

@@NicholasRenotte Nah its fine 😂

@freehawk78 Год назад

Great video! Would it be possible to make the Stable Diffusion API work in a standard website let's say a WP or another CMS, giving the user the possibility to generate the images from that page?

@zohaibshaik 7 месяцев назад

Hey NIck. While checking your project i ran into few issues which python version have you used I'm currently using 3.6.0

@meanamin2463 Год назад

Hello there , amazing tutorial , but i have a question that i am stuck with , what is the Auth_Token in the api file , is it python auth token for the enviroment or what ? I have been away from Python for a while now , so i am confused i would appreciat your hlep , cheers

@user-lx4xe5nm8v Год назад

Hello Nicholas, please build an app that can generate a talk video from an image and text like in D-ID. Thank you very much. Your work is awesome, I can learn a lot.

@huzi46 Год назад

Congrats you did it man🎉🎉🎉🎉

@NicholasRenotte Год назад

WE did it!! Thanks a mil @huzi!!

@huzi46 Год назад

@@NicholasRenotte really love the work you put in with ur vids, keep it up man you’re doing an excellent job ❤️

@ndumisomhlongo1125 Год назад

you're my role model

@TheFahad318 Год назад

Inspirational 🔥

@jeaywilson2168 Год назад

Great video Nicholas. What specifications do you advice I look for in a laptop I can use for about two years as a beginner in DL? I'm looking to get a hp pavilion 16gb RAM 512 gb SSD (or 1 tb SSD) Ryzen 7 5600h RTX 3050ti Do I need to be concerned about screen specifications? Are these specs good enough?

@NicholasRenotte Год назад

Screen specs don't matter too much, but the GPU does if you want to get into deep learning! In a laptop you'll kind of be a little limited either way but a 3050ti is a great start!

@jeaywilson2168 Год назад

@@NicholasRenotte Thanks, I was also hoping that perhaps I could use cloud services like paperspace, that'd suffice right? I just have a lot of questions, and being a novice I'd appreciate if you could guide me on my journey. Can I follow you socials, like twitter?

@hayagreevsivakumar9188 Год назад

Hahahah best video Ive seen about python, fun anxious and exciting!!

@SergeiKasimov135 Год назад

Hey Nick! I watched your video and I really like it. However, I have a problem with cuda. The error is "out of memory". Can you help with it, please? I've already checked a lot of websites to solve it,but nothing helped.

@henkhbit5748 Год назад

Awesome Nick👍👋👋 maybe u can generate that Mercedes win an F1 race😉

@NicholasRenotte Год назад

LOL you’re breaking my heart Henk 😂 maybe next season?! or the one after that?!

@gplgomes Год назад

Hei bro, the link was for the Titanic and not for the Stable.

@TheCopernicus1 Год назад

Awesome mate!

@NicholasRenotte Год назад

Cheers @Bass!

@out-of-sight Год назад

This is amazing. That's it. 👌

@CG23 Год назад

which extension need to install please tell full setup like how you create stable bud and how we publish file

@Genicist Год назад

congrats for completing this lol....but howdo you make those text animations also

@NicholasRenotte Год назад

LOL a ton of final cut pro and After Effects 😅

@ebenezeroseiappiah7107 Год назад

Good job!

@jgvisk1041 Год назад

Nick, your videos are amazing, but I was wondering if you would be willing to release the requirements.txt for the environments you run? I was able to follow the instructions of this challenge but then I face an error where torch says it is not compiled with cuda. I would like to make sure my versions are stable, and utilizing the requirements you have would help in troubleshooting.

@NicholasRenotte Год назад

Just added it to the github repo for you! Let me know how you go.

@jgvisk1041 Год назад

@@NicholasRenotte Wanted to reply back that the requirements did the trick! I had an issue with torchvision, and your requirements.txt helped me troubleshoot how to properly install it. Thank you. Please consider releasing the requirements when you do future projects, and possibly past ones if you don't mind. I haven't been able to complete the TFOD course due to continuous errors on dependencies, even beyond the error handling you included, and don't know where to properly troubleshoot. Having a perspective of the environments you run in and version of python helps a great deal.

@habs795 5 месяцев назад

Can you share the instructions please?

@Aidev7876 Год назад

I like your videos from the instructional aspect. However honestly I don't care of you google it or stack it or copy paste from notepad. Who cares? Just explain how it works and teach us. We are not trying to give you a grade based on speed or copy paste.

@NicholasRenotte Год назад

Thanks a mil @Angels Mobile! I add those constraints to make it challenging for myself as well, forces me to push myself to do better for you guys!

@Aidev7876 Год назад

@@NicholasRenotte that's fine man. I guess I'm older than the average in this group. So the other guys may like the challenging aspect you are doing. Although my comment got 10 likes so I'm not quite sure... In real life, nobody is going to check if you used Google and actually there is a value showing how to search for help real-time cause everyone will learn from that part too. But hey. As I said. Your content is great and you do have lots of positive energy. Up-to you if you want to keep the timer and no internet check. My 1 cent 😀

@Aidev7876 Год назад

@@domehouse79 ?

@ProfessorThock Год назад

He obviously had to learn all of this stuff and check the documentation beforehand anyway so it doesn’t really make much of a difference. He’s just going off memory from what he coded beforehand right? Pretty sure the 15 minute thing is just a fun/gimicky way of showing that it doesn’t have to take long to do something cool

@Pyth_onist 6 месяцев назад

@@NicholasRenotteplease help me out with the authtoken library it's not working on my python I've tried everything

@vigneshpadmanabhan Год назад

Amazing video as always… I’m not able to install some of those libraries while following the video.. where to find Authtoken?

@NicholasRenotte Год назад

1. Create a new python file called authtoken.py 2. To that add a variable called auth_token 3. Get your user token from hugging face and set the auth_token variable to that value Should look like this auth_token="ahjkhkjhkh

@habs795 5 месяцев назад

What are the basic software requirements for this project? Can you share the instructions?

@jeewan3108 Год назад

I am a lot confused with the virtual environment and global enviroments in pyhton, it is becoming so hard to import a specific verison of a module in a pyton project. Please make a video on working with virtual enviroments and how to keep them usefull everytime like you as a pro.

@shreyasnair994 Год назад

Hey Nicholas can you please post a video about using aws recognition for real time object detection and triggering a raspberry pi when certain objects are detected using aws lambda, I really would like to know if its even possible and if yes how it can be done.

@primz01 Год назад

@Nicholas Renotte So the error is in 3.4 Build train and test Partition after I wrote res = preprocessing_twin(*example) I got a type error= preprocessing_twin takes 2 positional arguments but 3 were giving

@primz01 Год назад

So pls how do I fix the error thanks

@joaobentes8391 Год назад

hey nick great vido again ur a machine that inspired me to start programing. So i have a problem with other project of urs , the cartpole balacing one: in this line n_state, reward, done, info = env.step(action) it gives me a error when i run it it says : "ValueError: too many values to unpack (expected 4)" then i put down a truncated var like i saw in opaigym rep but after that it gives me another error while training the model it says: "ValueError: Error when checking input: expected flatten_1_input to have shape (1, 4) but got array with shape (1, 2)" and im cannot understand why could u pls explain if u know whats happening? thanks a keep up the good work!!!! : ) peace

@NicholasRenotte Год назад

Double check your neural net, this would indicate that your output layer is expecting something like [[1,0,0,1]] but is getting [[1,0]]

@hope-kz6tb Год назад

Thank you soooooooooooo much

@yolow8126 Год назад

bro your big brain I have a question how do you know to work with which liberay in .py

@NicholasRenotte Год назад

A lot of it is research and experimentation!

@gplgomes Год назад

I liked this. Is it possible to use my picture and mix with Ai scenes using this approach? Signing a big fan.

@NicholasRenotte Год назад

Apparently it is: huggingface.co/spaces/fffiloni/stable-diffusion-inpainting i haven't tried it yet though @Geraldo!

@ihebbibani7122 Год назад

Hi Nicholas , what changes should we make to make it run on CPU ? Thank you

@adamalazzawi Год назад

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu") I guess

@Slayer-33 Год назад

Teach me your ways baby Jebus!

@shaikhsara3952 8 месяцев назад

Which python Version does it need and is it mandatory a GPU computer to run this or installing torch without GPU option will be okay also it can't import many classes what shall I do please help me out?

@REY-em7fv 7 месяцев назад

hey man! such a great work. I wanted to make similar type of thing but only for car specific thing. Can you please help me where I can find model for cars.

@SunnyYadav-yv9ie Год назад

Hey nick can you create a video on how to integrate with replicate stable diffusion using node js

@ahmetcals351 Год назад

Hey Nick, the get the code link is seems wrong. It's mapping to Titanic code. Just letting you know :)

@NicholasRenotte Год назад

Thanks a mil @Ahmet, yeah that's my bad included the right link in the comments :)

@f16gaming-pubgmobile95 Год назад

Can you please tell about the dataset you used for this

@l3lxf Год назад

Which app do you use for recording videos ?

@NicholasRenotte Год назад

OBS :)

@franciskhan9875 Год назад

Thanks bro

@shayanshaikh534 Год назад

Is there a way I can change the batch size? I have 4GB ram but 512MB is the additional requirement as I have 1650Ti Any help please

@DhruvSharma-mh5vf Год назад

Hey Nick i just completed coding it but when I run a pytorch_bin file is being downloaded....is it necessary to download it as you didn't downloaded it

@aryananchan5764 Год назад

did anyone else find difficulties to run this code?? I can't generate any images its showing me error:'sample'

@Cinemaaereo Год назад

Hi! I found your video searching for a way to add a simple Login page to my Local Stable Diffusion PC. It's impossible to find something like this! Can you do some tip about it? Thanks. Sorry for my english.

@bilalmoten8855 Год назад

Can you share some way of making this into an api and running on heroku rather than my own pc?

@minam7733 Год назад

Can I use API JSON which I fetch or extract it from a video instead of text?

@NicholasRenotte Год назад

Would likely be a different use case if trying to pass in video or a frame as an input!

@maheshrudrapati4026 Год назад

Hey Nick, Your Get the Code is referring to the Titanic project, not the Text to image one. Can you please check and update it or can you share the direct link of this project excited to get my hands wet in this modle. Thanks for the great tutorial.

@NicholasRenotte Год назад

Cheers Mahesh, updated!

@habs795 5 месяцев назад

Can you plz share the instructions about this project?

@nazarzaki44 Год назад

What if I would like to get more images and not only one?

@fahadabdullah510 Год назад

Can you create anything like Virtual Dress Mode for users on a Website/App using GANs or at least give a little bit of info regarding on how to build such a neural network it will be very much helpful.

@NicholasRenotte Год назад

You got it!!

@fahadabdullah510 Год назад

@@NicholasRenotte Thank you very much and can you make a detailed video on it because i think it is a very important topic and there will be very much to learn from it.

@satishkumar-it2hl Месяц назад

Hey Nick how can I run this code in the CPU without GPU, when I am trying to run code in the CPU getting a lot of errors can you create a video on how we can run it on the CPU