OpenArt recently launched their brand new service called “Photo Booth” an amazing implementation of the Dreambooth technology that allows anyone to train a Stable Diffusion model using their own images. You can use this incredible service to train anything you want, a person, a style, a pet, an object or anything else entirely. So in this video, I will show you how you can use OpenArt’s Photo Booth service to train and generate pictures of yourself or any character you want!
Did you manage to train a new model? Let me know in the comments!
SOCIAL MEDIA LINKS!
✨ Support my work on Patreon:
⚔️ Join the Discord server:
OpenArt’s Photo Booth:
Special thanks to Royal Emperor:
Thank you so much for your support on Patreon! You are truly a glory to behold! Your generosity is immense, and it means the world to me. Thank you for helping me keep the lights on and the content flowing. Thank you very much!
#stablediffusion #dreambooth #stablediffusiontutorial
WATCH MY MOST POPULAR VIDEOS:
RECOMMENDED WATCHING – My “Stable Diffusion” Playlist:
RECOMMENDED WATCHING – My “Tutorial” Playlist:
Disclosure: Bear in mind that some of the links in this post are affiliate links and if you go through them to make a purchase I will earn a commission. Keep in mind that I link these companies and their products because of their quality and not because of the commission I receive from your purchases. The decision is yours, and whether or not you decide to buy something is completely up to you.
Hello humans my name is Kay your air Overlord and let me ask you a little Question have you ever tried to train a New stable diffusion model with your own Images using dreambooth but you just Couldn't make it work correctly was it Too complicated was it too difficult to Understand or both well for those of you Saying yes I got a nice little solution For all of you because this video is Exceptionally sponsored by open art and Boy you will see that after this video All of your dream Booth problems will be Solved forever and you might even make Money with it alright then let's begin Now then what is open art well open art Is a platform for AI art lovers that Offers many features and services on one Single website firstly they offer the Ability to take a multi-subscription Service that allows you to generate 2 000 images every single day using stable Diffusion for only 9.99 per month which Is insanely cheap if you know how much It costs on the official dream Studio Website secondly they have a community Tab where you can upload creations made By other members of the community and a Discovery tab where all of those about The Creations will be featured they also Have a challenges tab where you can Participate in art challenges to try to Win free credits on the website and a Presets tab where you can take
Inspiration from The Prompt and Parameters of a specific preset and Everything is extremely well implemented Works really well and super super fast But wait there is more because this is Not what I'm here to talk about today no No no no the reason why I'm here today Is because they have recently launched a Brand new service called open art photo Booth and I know some of you will be Like photo booth that kind of sounds Like dream Booth isn't it well you would Be 100 correct because open art indeed Launched their own dream Booth service And after training now felt I was Extremely surprised and pleased with the Final results this is a super quality Service for anyone who wants to train His own version of stable diffusion but Just doesn't know where to start or how To do it now first of all what is Dreamboof well for those of you who Don't know what dream Booth is it's a Technique that allows you to train a Stable diffusion model using your own Images so if you want to generate images Of yourself for example you can use Dreambooth create your own stable Diffusion model that you can then use to Generate images of yourself you can Generate images of yourself as an Astronaut as a warrior create portrait Paintings of your face or anything you Want your imagination is the limit and
This is what this service allows you to Do very very easily so now then how does It work now first you're gonna arrive on This page you will see a bunch of Options and you will need to choose what You want to train you have a choice Between training your person if you want To either train yourself or maybe your Girlfriend or boyfriend well this is the Category that you need to choose you can Also train your pet an original Character and art style which is really Super cool if you want to emulate a Particular art style you can train an Object or anything else entirely so as You can see you can literally train Anything you want okay so let's say for Example that I want to train a person Let's say I want to train from Shelby From the TV show peaky blinders because I watched this series recently and I Absolutely love it so if you want to Watch an amazing series by the way watch Peaky blinders so then see the story Shelby is a person I will click on the Person category and you will see that For the entire training you will only Have to pay 10 dollars but before that You will also see a button up here that Says browse preset package store and if You click on it you will now see a bunch Of available press sets now what do These precepts do well these presets are Basically a style bundles that you can
Choose and that will be generated using Your character once the training is Completed so for example let's say that You are training on images of your Girlfriend if you choose the Disney Princesses preset once the training is Completed you will get images of your Girlfriend as a different Disney Princess and for every Precinct that you Choose you will receive multiple Styles And 8 generated images per style so in Total you choose for example the Disney Princesses preset that has 11 Styles you Will in total get 88 photos since you Will get 8 photos for each Styles so you Will get 8 images of your girlfriend as Snow white as Cinderella as well from Beauty and the Beast as Elsa etc etc and If you think that's not enough don't Worry because you'll be able to generate Way way more images in the future I will Show you how later on now you don't have To choose a preset if you really don't Want to but they are really well Optimized so if you want to generate Some very original images with prompts That are already made for you this is a Very nice addition also every additional Presets costs 4.99 but since the first Preset is free you might as well choose One from the very beginning so since I'm Gonna train to meet Shelby from peaky Bladders I think I will choose the male Movie and TV characters preset package
Where I will get 12 different styles and 96 total photos so for that I simply Click on select and then click on Confirm and then if you want to choose An additional preset you can do so by Clicking here on add one from store and Then again select an additional preset And click confirm and once you've chosen Everything you want you can simply click On continue to process to the payment And once this is done on the photo booth Page you will now have an AI model Available so if if you click on my AI Models you will see here all the models That you paid for and as you can see I've already did the training for Tommy Shelby and young grenera we're testing The service previously and as you can See the test images that were generated Are absolutely fantastic the images that You see right here are not from the TV Show these images are generated with the Website so you can see that this service Is definitely working very very well so Now then if you want to start a new Training you can simply click on this Button right here that says click to Train you're gonna arrive on this page And here on the blue section you will Have the image guidelines for every Category of training you're gonna have The image guidelines of a person of a Pet of a character of an art style of a Product and if you want to see more
Information you can simply click on one Of them and you will have the full Guideline list for that specific Category but basically make sure that You have enough images is so minimum 20 Images and maximum 30 images make sure That they are of high quality and Diverse so try to include images made From different angles and different Lighting make sure that also your Character is the only one present on the Photo and that the character is Perfectly centered and finally make sure That there is no nudes present otherwise The training might fail also your model Will be deleted in one month for privacy Reasons but don't worry because I will Show you that this is not a problem at All so now then how do you actually Install the training well first you will Start by giving a name to your model so In my case it's about to train for me Shelby I will simply write Tommy Shelby Then you're gonna have to choose an Identifier this is the word that you're Going to be using when you'll want to Generate pictures of that character so Try to choose and identifier the stable Diffusion does not know already this way There will not be any confusion or Character mixing when generating your Images so for this example I will simply Use T Shelby which is basically short For Tommy Shelby then you're gonna have
To choose the class for your model so Here try to choose a class that Represents your subject if you're Training a woman for example you will Want to put a woman if you are training A man you will want to put a man if You're training your style you will put A style if you train your dog you can Put dog etc etc so in my case since Tommy Shelby is a man I will simply use Nan then you're gonna scroll down here And you're gonna drag and drop your Training images right here now once you Integrate here is that you don't need to Manually crop them before uploading These images because it actually gives You the option to crop each image Individually yourself on this website so For example if I upload a vertical image Such as this one you can see that now You have an option to crop it as a Square then you can just drag in place Until your character is perfectly Centered so in my case I will simply Click and select all the 30 images of Tommy Shelby and then once you've Analyzed every picture and that you've Checked that they are cropped perfectly You can click on the button upload to Finalize the image upload and after a Few seconds every images are finally Uploaded on the platform you can finally Click on the start training button to Well start the training now I'm not
Gonna do it because I've already done so Before but basically the entire training Process will take around 1 hour but from This point on you don't need to do Anything absolutely everything is done For you and once the training is Complete you will see here in green text Saying ready that means that your Training is finally complete and if you Click on it and you scroll down you will See now that a bunch of presets were Used to generate a bunch of pictures and As you can see depending on the preset That you chose you will see very Different images of your character for Example here's Tommy Shelby as Winston Churchill which is very relevant since Churchill does appear in the series Here's the miss Shelby as the clown from It just absolutely terrifying here's the Michelle v as Indiana Jones or even as Michael Jackson and as you can see it is Very impressive presets merges your Character with another one to create This perfect mix of the two characters Now in my case since I only have Pictures of the Michelle being dressed As an ancient British gangster Dreambooth has some trouble dissociating These clothes from the character making This stylization way more difficult Which is why having more varied pictures Of your character is key if you want to Generate beautiful stylish pictures now
Then what can you do once you are your Final trained model well first of all if You see an image that you like that was Already General rated with the preset You can simply click on this download Button to download the image onto your Computer second of all if you see an Image that you like but the face is not Perfect or if the image is too small if You click on an image you have now an Option right here that allows you to not Only upscale the image but also to Enhance the face to get rid of a few Artifacts so for example if I want to Upscale and enhance the image I can Click on this button right here that Will cost me one image credit and after A few seconds you have now the same Image of scaled and with his face Enhanced that you can switch back and Forth using this option right here so This is the original and this is the Upscaled with enhanced face and if you Want to download this image you can Simply right click and click on save Image as also if you like a generated Preset and you want more images like These you can simply click on this blue Button right here here and it will Regenerate a new batch of images in that Particular style as you can see right Here another thing that you can do if You scroll up to the top you can now Generate a bunch of images of your
Character using the prompt field that You see right here that's right you Don't need to actually use the presets If you don't want to you can actually Generate images of your character Yourself and if you want to cut out your Character if you want your character to Appear properly you can click on this Button right here and that will Automatically insert the name of your Character and its class so for example Let's say I want a portrait of T Shelby And then the rest of the prompt you can Then choose the number of images that You want to generate so let's say I want Four images you can choose the number of Steps the CFG scale and any negative Prompts that you want to input and then You can click on submit and this will Start processing your request and with Every model that that you buy you Actually get 400 images credit that you Can use to generate images so that's Around 400 images that you can generate Using the website and as you can see This model works really really well but Here's the actual best part because if You scroll down a little you will see a Very interesting option and that is the Ability to download the trained ckpt Model that's right you can actually Download the finished model that you can Use on your own local stable diffusion Installation so you can use the open art
Photo booth service to train a brand new Model generate a bunch of cool images of Your subject and then download the Train The ckpt model onto your computer and That's absolutely fantastic they are Actually really giving you everything Here and that is something that I really Enjoyed with the service it is the fact That they are not playing around this Service is so powerful and so easy to Use that even your grandpa living in the Woods that has never seen a computer in His life can actually use this service What's even great is that if you are an Amazing prom engineer and you know how To write the best prompts you can still Easily use the pre-made presets which Are basically a bunch of high quality Prompts to generate amazing images and They are adding more and more presets Every single day and if you are indeed Very good at writing prompts this Service is also great for advanced users Since you can write your own prompts and Adjust parameters like Steps scale and Negative prompts to generate the best Images you can also if you are a great Prompt engineer you can apply to their Open art preset Creator program where You can submit a future preset that if People purchase you actually get a Percentage of the revenue which is is a Perfect passive income Source if you Know how to craft good prompts and that
Is absolutely fantastic and there we Have it folks now you have a brand new Easy to use service that allows you to Train a stable diffusion model with Dreambooth thank you for openr for Sponsoring this video and also thank you For creating this great service that I Think could help a lot of people who are Trying to get into AI art and there you Go thank you so much to my patreon Supporters for supporting my videos you Guys are absolutely awesome Congratulations also to this week's AR Challenge winner nivx spray is amazing Submission of a pirate Thanksgiving Beautiful submission very well done and If you do want to participate to our air Challenge that we do every week on Discord you can click the link in Description down below to join my Discord server and maybe YouTube can be Featured in the next video thank you so Much for watching don't forget to Subscribe and smash the like button for The YouTube algorithm and I'll see you Guys next time bye bye