Create multiple consistent characters with dall-e 3 & Custom GPT

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
hello money makers and welcome back to the channel today I am super excited to share with you this method for generating multiple consistent characters that you can use in your story books animation projects comic books or whatever amazing projects you have in mind for example I've created these animations with this method and as you can see all my characters for all the different scenarios and scenes are very consistent these are by far the best results that I've achieved with any art generator to date and here's a couple more examples of comic book pages that I've created with this method and again you can see even with these extremely complex characters their style and look remains consistent so let me walk you through the process of building your own custom GPT so you can achieve results like this for your project and to help speed things up for you I've left a Basse prompt in the description that you can adapt to your specific use case all I ask is that you smash that like button for me so this video will get pushed to more people that might be struggling to create consistent characters for their projects all right without further Ado let's head to chat GPT do note you will need to upgrade a gpts plus plan for $20 a month to be able to create your own custom gpts and generate images using Dolly so once you're here just go to the explore tab then create a GPT here you would normally start by chatting back and forth with the bot to establish the parameters for the instructions but with this method you can skip the back and forth of creating a GPT manually and just jump straight to the configure page then copy The Prompt I left in the description of the video and paste it here in the instructions box then you want to name your Bot I'll name this bot storybook illustrator and for the description I'll just put generates consistent characters then you're going to need to manually fill in all your specific information in the parentheses of the bass prompt I provided now in this case my story only has one main character but as you've seen you can do up to three characters without issues and I actually really like this 3D Pixar style but I want to add a unique Twist of a neon Aura surrounding my character so in the first set of parentheses where it's asking for the style I'll put Pixar 3D animation with a neon Aura around the character and I want these images to be square so I'll change the aspect ratio from 16x9 to 1x1 and now if you already have a really good description for your character you can put that in the next set of parentheses here then again in the parentheses next to art style you want to put your new style Pixar 3D animation with a neon Aura around the character then here you can go into color and lighting if you like as well as adding any camera angles you might want to incorporate I'll just leave these for now and before we get too far I want to show you how to come up with a really good Basse prompt for your character as this is crucial for this method to work you can start by prompting GPT with as many details as you can think of for your main character including their name age hair color eye color clothing style skin color Etc so I prompted Marcus is a 13-year-old African-American boy with a 3D Pixar animated style appearance he has trendy spiky black hair with subtle purple highlights and brown skin his eyes are striking green he's dressed in an oversized black hoodie with a large white skull graphic baggy black jeans and all-white high-top sneakers Marcus has a neon aura that surrounds him this Aura should give a vibrant almost futuristic Edge to his appearance now this is actually my final revised prompt but you would just start with the basic description of your character being as specific as possible about as many details as you can think of then once you get an image that you like you're going to want to click your image and open the info tab which will give you the prompt that GPT created and used for this image then you want to copy this prompt with the copy Tab and pop it right back here into the GPT then prompt can you remove all the expletives from this description and just give me the specific details of the character without the explanations this will give you a nice condensed version of your character prompt then I would again try generating an image with this Bas prompt to see if you like the output and if you need to add any details or take any details away now is the time so after you've tweaked this prompt and produced a character that you're happy with then you're ready to save this image and use this prompt as your base prompt and if you have more than one character then just repeat this process for your remaining characters do note I wouldn't try to exceed three main characters as the AI will begin to get confused at this point okay once you have all your main characters base prompts here and you have at least one reference image for each character uploaded then you can go ahead and check these three boxes web browsing Dolly 3 image generation and code interpreter then once you click the update Tab and save you're ready to test the bot so let's view GPT and give it a shot this time I'll prompt the bot to generate my first scene and the bot is primed now with your images so you can just use your character's name and a description of the scene for example Marcus standing in front of a graffiti wall under the street lights and your Generations from now on we'll stay relatively consistent just by using the character's name but to enhance coherency you should continue to save the best and most similar images to the bot you can do this this by going to the drop- down arrow here and selecting edit GPT then uploading your new reference images once you've made your changes once again you need to save this and I'm publishing this to only me because this bot isn't built for the GPT store then yeah you can simply continue on this way generating the scenes for your project so let's try the next scene our prompt Marcus is spray painting a beautiful mural on the wall under the street lights and look at these results almost perfect adherence all right let's move on to the third scene and this scene brings in another character and if you're familiar with using AI art generators of any kind you'll know that as soon as you add a second or third element you start to lose consistency with your characters but watch this bam now we have our original character Marcus standing next to his older Mentor all right I'll demonstrate one more Scene this time with a lot more going on and see if we can still maintain Marcus The Prompt for this scene is Marcus is reprimanded by his parents and principle for his graffiti and there you go Marcus still looks like Marcus and we have his parents and the principal incorporated into the image these are Next Level results and I have yet to achieve this kind of consistency with my characters between different scenarios and scenes so I hope this video has demonstrated how effective creating your own custom GPT can be for generating multiple consistent characters for your projects and just think if your custom GPT is useful enough you could potentially make some money from open AI as well but that's a topic for another video and one more thing about these images if you're not familiar they are low resolution directly from dolly so if you plan on using them for commercial purposes I would highly recommend upscaling them you can accomplish this for free with upscale AI image upscaler this site gives excellent results I use the general photo fast ra setting and you also have the option here to batch upscale which is nice the only drawback is that if you plan on building your book or project Within canva These images will sometimes be larger than 25 megabytes and canva won't allow you to import them so to get around this you can go to a free Photoshop type tool called photo P once you have your image imported here you can change the size by selecting image from the two bar at the top then selecting image size and reducing the size of your image by at least half for instance these images from do e are 248 pixels wide after upscaling so I'll change this back down to 1,24 pixels click okay and save this image by going to file and exporting as a PNG now you'll have no trouble importing these into Cano and building your project and if you're curious as to how I created these animations for free then let me know in the comments and if you guys show enough interest I'll build a dedicated video on the topic also if you have any questions about the process covered in this video please don't hesitate to drop them in the comments and I will get back to you shortly I am confident this method will help you generate the most consistent characters for your projects if you're interested you can check out this video here to learn the entire process of creating a children's book and getting it listed for sale on Amazon that's it for now stay safe and be blessed [Music]
Info
Channel: AI Money Maker
Views: 19,270
Rating: undefined out of 5
Keywords: AI Art, easy side hustle, AI side hustle, make money online, Make money with AI, AI revolution, Print on Demand AI, POD Etsy, online business, ai tools, howtoai, digital products, kdp, ai marketing, one person business, free ai tools, custom gpt, custom gpts, custom gpt tutorial, character consistency, dall e 3, dalle 3 tutorial, ai consistent characters, dall e 3 consistent characters, consistent characters, openai, chatgpt, aimoneymaker
Id: sr8meU_Mc3g
Channel Id: undefined
Length: 8min 1sec (481 seconds)
Published: Sun Jan 21 2024
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.