Create multiple consistent characters with ai, Dall e 3!
Вставка
- Опубліковано 26 чер 2024
- Create multiple consistent characters with dall-e 3 & Custom GPT! How to create your own custom GPT that can generate consistent characters for your specific project! This is by far the BEST method for generating consistent characters! I've also left a custom set of instructions to help speed up the process of creating your own custom GPT down below.
Try this MASTER PROMPT ready to copy and paste along with 7 FREE prompts that will help you launch your next side hustle! 7 Prompts to Launch Your Side Hustle To start making money online now! You can also get my FREE Affiliate marketing starter guide
👉linktr.ee/fitnfine11
Try Canva FREE👉partner.canva.com/DKz3Dd
Try Vid IQ👉vidiq.com/AIMoney
Try music and sound effects at Uppbeat: 👉uppbeat.io/?referral=brian-3mhjz
Chat GPT👉chat.openai.com/
Photopea👉www.photopea.com/
Upcsayl👉www.upscayl.org/
Music from Uppbeat (free for Creators!):
uppbeat.io/t/abbynoise/firestorm
uppbeat.io/t/mountaineer/lets-go
uppbeat.io/t/torus/hourglass
uppbeat.io/t/jeff-kaale/aurora
uppbeat.io/t/moire/bangroll
👉 Use this Custom Instruction Set To create your own GPT:
When creating illustrations, the GPT adapts to various character descriptions and artistic styles as per user requests. Each illustration focuses on the context of a specific scene, highlighting character interactions, actions, and the environment. The GPT maintains the requested visual style, ensuring high-resolution and high-quality output in any specified style ([chosen art style]). Characters are depicted with details consistent with the user's descriptions or reference images. The default aspect ratio for images is ([16:9]), but this can be adjusted based on user preference.
Formula for Image Creation:
To ensure consistency and adherence to user preferences, the following elements form the creation prompt:
1. Subject Description: Incorporating detailed descriptions of characters ([character descriptions]) provided by the user, ensuring consistency in features, outfits, and appearances.
2. Environment Description: Detailing the scene's setting based on user requests, or creating an appropriate background if not specified.
3. Art Style: Adapting to various artistic styles ([chosen art style]) as requested by the user.
4. Color and Light: Describing main colors and lighting in the scene ([color and lighting details]), focusing on creating the desired mood and atmosphere.
5. Camera Angle and Composition: Indicating the perspective ([desired camera angle]) for the scene to enhance storytelling and visual appeal.
[User's Action Required]:
- Choose and describe the preferred art style ([chosen art style]).
- Choose aspect ratio for images ([16:9]).
- Provide character descriptions ([character descriptions]).
- Detail the desired colors and lighting ([color and lighting details]).
- Indicate the preferred camera angle and composition ([desired camera angle]).
Disclosure:
THIS PAGE INCLUDES AFFILIATE LINKS. IF YOU CLICK AND PURCHASE, I MAY RECEIVE A SMALL COMMISSION AT NO EXTRA COST TO YOU.
I ONLY RECOMMEND TOOLS I HAVE PERSONALLY USED. #canvassador #aimoneymaker
You saved me I took what you did and added some mods to the prompt and now I can streamline anything!! you saved me so much money!!!!!! TY! Definitely have a new long-term subscriber!!
Thank you, this is better than I expected to find on the Internet!
Thank you so much for creating this video-excellent clear instructions-best I have found so far!
Glad it was helpful!
I was your 1000th thumbs-up! Also thank you for this video
Thank you!
Thank you for sharing this video I can't wait to try this.
Amazing video! Thank you many times over!
This is great! Thank you!
Appreciate you going deeper into this. I used a custom called Consistent Character GPT which seemed to work quite well. No way to get it exact but I'll give your method a shot!
Bro, this is exactly what I needed. THANK YOU! Liked and subscribed.
Thanks for the sub! Glad I could help!
Absolutely phenomenal. Thanks very very much!!
Thank you for the comment and you are very welcome!
Yes Please!!! Awesome Video!
I am getting ready to illustrate my first children's book and have watched a ton of videos using midjourney, leonardo, chat gpt, and dall-e to create consistent characters after struggling for a few hours to do it in MJ the other day. This video is by far the best one I have seen that creates consistent results with minimal editing in canva and photoshop. Thank you so much for sharing this! It's GOLD! 🙏
That's amazing! So glad this method helped you get some better results,.. Good luck with your first book!
Thank you! I tried doing the method in this video for a few hours but really struggled to get the consistency that you did. I found the ai kept adding things to my character that always made it slightly different.
Just here to echo this comment. Thanks for the video
Thank you god i didnt even have to search. After exhausting how to make consistent characters in my stories you show up in my feed. Thank you for this 😊
Glad I could help!
Yes. Please, I would love to see the animation!
Omg thank you!
This is Gold! Thank you.
You're very welcome!
This was massively valuable! Subscribed!!
Thanks!
Killer content brother love all your vids
Thanks SO much. The pace was just perfect and the step-by-step easy to follow. For sure interested in how you created the animation at the end also. Thanks again!
You're very welcome! Here's the link to the animation video if you'd like to check it out 👉ua-cam.com/video/ONqSzEkuKVk/v-deo.html
Thank you so much for sharing!!! This is the best and east explanation that I’ve come across to get started. The instructions are clear and direct without any other irrelevant information. I look forward to learning more from your video.
Glad it was helpful!
Thank you this was highly resourceful
Glad it was helpful!
This is one helluva good video. Just excellent. Happy to subscribe and learn a lot more. Thanks a bunch.
Welcome aboard!
Man, you are amazing to explain this as easy it can be!, suscribed for future videos! thank you!
Awesome, thank you!
I was pretty impress by your content I have never seen anything so complete, you tackle the problems that we were going to have to rtackle before us and solved it and share it. Of course you deserve a follow keep it up
Glad it was helpful! Thank you!
Fabulous video! I learned so much. Thank you for sharing your knowledge.
Glad it was helpful!
GREATFUL OF PEOPLE LIKE YOURSELF WHO FOUND A WAY TO ENJOY AND SHARE THIS INFORMATION WITH ALL OF US AS YOU WORK AND PROFIT ITS AMAZING TO SEE.
I appreciate that! Thank you!
Haven't tried it yet but WOW! Thank you for sharing this. Subbed!
Thanks for the sub!
Awesome,!
Yes, videos please!
A big Thank You for these instructions. I created some amazing images for my book.
Amazing! So glad you're having fun with the process!
Please create a video tutorial for the last illustration the kid and the dog walking in the forest 🌳
GREAT VIDEO... thanks so much!
Nice job on this video. Thank you.
Thanks! Glad you liked it!
Good stuff , great info. Looking forward to seeing the animation video.
Thanks! You can check out the animation video here👉ua-cam.com/video/ONqSzEkuKVk/v-deo.htmlsi=qzS8B-DR0Rt2HfID
This GPT prompt you created is great. It takes a lot of regenerating but what I find works it uploading the images you like as you go. Also I slightly tweaked the prompt as it generated more images and it also helped me with the descriptive words I couldn't think of. My biggest tip using this is to be patient if you are going for a specific look, continue to edit your prompt as you go and make sure you download your images! I lost some images in the beginning because I didn't download them. I'm not sure how to save them on GPT. Overall this prompt is great. I will now that my 1st character is perfected I will work on my second one. And lastly I agree this is vastly better than MJ because by creating a specific prompt you are literally telling GPT how to behave whereas MJ you can only enter and change the prompt over and over.
Thank you so much for the valuable feedback and good luck with your new project!
My dude, thanks this was educational and simple enough for a novice like me to follow.
Glad to hear it!
great tut!
Thank you soo much 😊.
Most welcome 😊
excellent .. the regular GPTs in the store that say Consistent Characters really can't do that.. the ONLY way to create Consistent characters on Dall-e is THIS WAY.. Thank you for this awesome and complete explanation
Well said! Thank you!
great tutorial i love it😍
Thanks 😁
Great stuff 😊
Thanks! 🙂
Great content!!!
Thanks!
nice and it looks easy to follow, looking for the animation
Nice skull on the hoody ;-) .... Like how you develop on the consistent character idea and go with custom GPT... nice!
Yes the skull stands out nicely and thank you for the inspiration! Keep up the great work I'm learning a lot from your channel! You have top notch content!
I am interesting in seeing all that you have to offer! I appreciate your video and expertise!
More to come!
@AImoney11 I have my notification set to "all" your channel is awesome- keep doing what you do!!
Bro, I gotta say that ! It's a fckn masterpiece! thanks for the video.keep it up bro!
Thanks, will do! Check out my latest release! Make Money with AI Animation Videos! Free Method 2024! 👉ua-cam.com/video/TOo5W3mtto4/v-deo.html
Great video!!!
Thank you!!
Super dope! Please show us how you did it
Thanks
thanx for the info!!!!
You bet!
You the best thanks😊
No problem 😊
It's been a frustrating process I noticed multiple characters at same time pain in butt but each sep and layering has been key for success. Thanks for sharing always a pleasure keep up the good work jumping right in to try I'll keep ya posted. Supreme wotk
Thank you! Multiple characters is definitely a pain and layering works fantastic!
Vin Diesel's Voice on a UA-cam Channel is gold. LOL. It was nice listening to Vin Diesel show me how to make this and this video was extremely helpful. I appreciate it. Do you know if it works well with non-pixar characters? Most AI image generators seem to have Pixar images perfected, but what about other things like comic book characters?
Wow, thanks!😎 Other styles like realistic and comic book characters are doable but definitely more difficult.
This is awesome! I liked it too, especially as 666 likes wasnt sitting well with me :) Question though, if i create multiple characters, do all their descriptions go inside the brackets in the instructions panel, or do you put each character description inside their own brackets?
Thank you! Yep you can put multiple characters within the same set of brackets as long as they are defined by their own name and clearly separated.
awesome man! I wonder about the animation video
Thanks! You can watch the video on the animation process here👉ua-cam.com/video/ONqSzEkuKVk/v-deo.htmlsi=C1Y5-JW2lchVkkaM
Thank you for this video!! I would be so grateful if you would do a video on how to animate your images.
Check out my latest release! Make Money with AI Animation Videos! Free Method 2024! 👉ua-cam.com/video/TOo5W3mtto4/v-deo.html
This is a excellent tutorial. Please do a video tutorial next
Thanks,. Here's the video 🎉ua-cam.com/video/ONqSzEkuKVk/v-deo.htmlsi=gz_2ZFiR4CUGCeqg
waiting for the animation! Please please please! :)
King
love the video, but how did you animate it?
Thank you and you can check out the animation video here! How to Make Money with AI Animation Videos 2024!👉ua-cam.com/video/ONqSzEkuKVk/v-deo.html
Thanks for the great video. Is there a way to make the background stay consistent from scene to scene?
Thanks! Yes, just transfer the description of your background from prompt to prompt and the image should carry.
Great video thank you for sharing. How do you compare Dall E with Mid Journey? which one works better from your point view?
Thank you for your question! Mid journey is amazing I can't deny that with some knowledge of prompting you can achieve some outstanding imagery,.. that being said I believe you get more for your money paying for chat GPT 4 with Dall E included. For instance I can't discuss with mid journey how to create the perfect prompt.😉 So all in all I think GPT with Dall E is a more versatile and economical choice.
I would love to learn how to create a video. Thank You
Coming soon!
excellent. How did you learn to do all this?
Thanks! Lots of online courses, UA-cam and weeks experimenting!
When you said smash the like button my like button on my phone lit up with a multi color outline. Idk if it was my phone or if you got something special going on on your end lol😂
Ha.. that's a new feature, UA-cam does it automatically when people ask for likes and subscriptions they light up
oh ok lol I'm about to make a video and refer to this video because I'm making a coloring book and I want to see if this technique helps with consistent characters@@AImoney11
Dude your like neo in the matrix. You need to drop the animation vid
love this ,quick question how would i describe to generate a character in the style of the boondocks
Thanks! You could try asking GPT to reference the Boondock style or I would consider maybe, 2 D cartoon style with an anime feel?..
@@AImoney11 thanx my guy...🔥🔥
great video! how would you add multiple characters in the description??
Just give each character a name and separate them with space.
waiting for your animation video
I just watched your video about AI and it was super useful. I already tried a lot of them in the free version, but I have a big problem that I can't solve...
I try to redraw (with AI)a man's portrait in watercolour paint. However, after generating the image, the face structure completely changes. Which is the best AI app to redraw or repaint a human face or portrait? Thanks for your time!
I would consider using image to image function somewhere, you can try playground here for free👉playground.com/create
Hi, pls I'd like to know what Ai you used in animating the video
Coming soon!
Animation please that was brilliant
Working on the animation video now! Thank you much for the feedback!
Would you suggest this way instead of the consistent character app within 4.0?
Yes, I believe my method to be more versatile.
HI, how can you do it for three characters? thanks for the feedback!
Just add each character's name, description and image reference to the same custom GPT.
I can't seem yo copy from the description. Is it different if on a browser?
Not sure but here you go...👉 Use this Custom Instruction Set To create your own GPT:
When creating illustrations, the GPT adapts to various character descriptions and artistic styles as per user requests. Each illustration focuses on the context of a specific scene, highlighting character interactions, actions, and the environment. The GPT maintains the requested visual style, ensuring high-resolution and high-quality output in any specified style ([chosen art style]). Characters are depicted with details consistent with the user's descriptions or reference images. The default aspect ratio for images is ([16:9]), but this can be adjusted based on user preference.
Formula for Image Creation:
To ensure consistency and adherence to user preferences, the following elements form the creation prompt:
1. Subject Description: Incorporating detailed descriptions of characters ([character descriptions]) provided by the user, ensuring consistency in features, outfits, and appearances.
2. Environment Description: Detailing the scene's setting based on user requests, or creating an appropriate background if not specified.
3. Art Style: Adapting to various artistic styles ([chosen art style]) as requested by the user.
4. Color and Light: Describing main colors and lighting in the scene ([color and lighting details]), focusing on creating the desired mood and atmosphere.
5. Camera Angle and Composition: Indicating the perspective ([desired camera angle]) for the scene to enhance storytelling and visual appeal.
[User's Action Required]:
- Choose and describe the preferred art style ([chosen art style]).
- Choose aspect ratio for images ([16:9]).
- Provide character descriptions ([character descriptions]).
- Detail the desired colors and lighting ([color and lighting details]).
- Indicate the preferred camera angle and composition ([desired camera angle]).
yeah not getting it to work i tried it and its not working. i can create an image and it knows it but will not recall the image at all
That's a bummer, it still struggles a bit with complex prompts, my best suggestion would be to try to simplify your prompt as best as possible.
Hi, I have just created my first character and copied the prompt. This shorter prompt do I copy and paste that in the instructions box, replacing the original prompt I wrote? Also, I want to make 3 characters. I understand I need to replicate the process, but when I have the final prompt for each character do I paste those into the instructions box with my 1st character? Or should each character have their own set of 1-5 in the instruction box?
Thank you for the question. Yes you wanna have the simplest prompt that you have achieved the best result with as your base prompt in the instruction box... And yes all three characters can be in the same instruction box as long as they have their own name and description. I hope that helps and good luck with your project!
This helps massively. Thank you so much 😊
I would love to know how to make a video!!
Coming soon!
if you upload the images you like in the configuration of the gpt will it always give that character? because I want to make different characters for different things but still want consistency. Will I have to make multiple GPTs?
You can do up to three characters with one custom GPT... Just include the characters name and description in the custom instructions as well as adding reference images for each character with their specific details. It's not always perfect and the more characters the more confused the AI seems to get but I found good results with up to three characters. Hope that helps man thank you for the question!
@@AImoney11 I mean for example I’m making a coloring book with a certain character. Then I want to make a different coloring book with different characters. Would I upload the images for all the different customers (up to 3) and revert to their names and the style I want. But as I make more books works I have to make a different gpt reach time?
@@Thehouseoftatum Yes basically you have to retrain a new GPT for your new characters and styles.
I would love a video how to animate the characters for a video
Here's the link to the animation video mentioned "How to Make Money with AI Animation Videos 2024!" 👉ua-cam.com/video/ONqSzEkuKVk/v-deo.html
Do you have a video on how to create animate 3D videos
You can check out the animation video here, How to Make Money with AI Animation Videos 2024! 👉ua-cam.com/video/ONqSzEkuKVk/v-deo.html
Can you do it through an API? Create a a custom gpt and Add reference images to that custom gpts?
I haven't tried this yet but I can't see why it would not work
I can't find the update button to test the bot, I am trying to use the images as a reference but it's not working. The only option I see the Create option but not the update.
After you've created the bot the first time then You'll be able to test the bot and after you go to edit GPT to add reference images make adjustments to your prompt etcetera then you'll be prompted to update the bot to save the changes. Hope that helps!
How do get to the promot page to create the character
Here's the link but you do need to beyond me papers and to generate images here,...Chat GPT👉chat.openai.com/
Thank you for sharing. After attempting this I am still struggling with character consistency. As I post each page it uses the same prompt in the tool but even though its very detailed it still leaves room for the AI to make its interpretations differently every time its used. I'm not understanding how this worked so well in your video. Are you willing for me to pay to have the prompt done by you? Any other assistance?
Yes, although this is the most consistent method it's still not perfect... If you don't mind sharing your prompt with me here, I can see if I can help you revise it to get better results.
When creating illustrations, the GPT adapts to various character descriptions and artistic styles as per user requests. Each illustration focuses on the context of a specific scene, highlighting character interactions, actions, and the environment. The GPT maintains the requested visual style, ensuring high-resolution and high-quality output in any specified style (Pixar 3D animation with a neon aura around the character). Characters are depicted with details consistent with the user's descriptions or reference images. The default aspect ratio for images is ([16:9]), but this can be adjusted based on user preference.
Formula for Image Creation:
To ensure consistency and adherence to user preferences, the following elements form the creation prompt:
1. Subject Description: Incorporating detailed descriptions of characters (Pablo is a 6-year-old chubby boy from Colombia with slicked-back short black hair, tan skin, and brown eyes. He has a thick bushy mustache and a noticeable beer belly. His attire includes a slightly worn white button-down shirt, fully buttoned up and tucked into baggy blue jeans, white tennis shoes, and a gold watch. The image should depict him in a 3D Pixar animation style, emphasizing his cheerful and confident demeanor.
) provided by the user, ensuring consistency in features, outfits, and appearances.
2. Environment Description: Detailing the scene's setting based on user requests, or creating an appropriate background if not specified.
3. Art Style: Adapting to various artistic styles (Pixar 3D animation with a neon aura around the character) as requested by the user.
4. Color and Light: Describing main colors and lighting in the scene (dark and gloomy with colorful pixar-like pops of color), focusing on creating the desired mood and atmosphere.
5. Camera Angle and Composition: Indicating the perspective ([desired camera angle]) for the scene to enhance storytelling and visual appeal.
[User's Action Required]:
- Choose and describe the preferred art style ([chosen art style]).
- Choose aspect ratio for images ([16:9]).
- Provide character descriptions ([character descriptions]).
- Detail the desired colors and lighting ([color and lighting details]).
- Indicate the preferred camera angle and composition ([desired camera angle]).
ADDITIONAL NOTES:
Every page I plug in creates a drastically different character. What should be plugged in when you create each page in the prompt? Simply the copy/scene or more? Is there any way to directly email you? @@AImoney11
You can e-mail me at...brbcustom1@gmail.com,.. Yeah, the AI is still imperfect. The best you can do is try to simplify anything that you can. Here's a revised prompt that you can try. (A 3D Pixar animation style portrait of a 6-year-old boy named Pablo. He has a chubby build, slicked-back short black hair, tan skin, and brown eyes, with a thick bushy mustache. He is wearing a slightly worn white button-down shirt that is fully buttoned up and tucked into baggy blue jeans, white tennis shoes, and a gold watch. The image should capture his cheerful and confident demeanor, with a bright smile and an energetic pose). Hope this helps. Feel free to email if you need more assistance.@@nicholasbugara6053
for a single character, it is doable for all the AI. I want an example with several characters (3 characters) in the same scene, it is indicated "Create multiple consistent characters" but but in reality there is only one consistent character ??
For times sake I only used one character as an example but this method will work with up to 3 characters as described in the video
@@AImoney11 Can you make an example because I can't do it? It works for a single character but when a scene contains 2 characters it loses consistency
Agreed. It does not work for 3. I did it by the book and it's a cluster F.
does it do it well w realistic character ?
Dall E 3 actually does realistic characters quite well.
Video idea:: "How to upload image to image consistent characters." Oh and great video.
Thanks for the idea!
Can I use references from my photos?
Having tried it with a real photo or a realistic image but I don't see why it wouldn't work.
Can I create realistic images I’m using this. Not 3d images
I would think it would work for realistic characters yes 😉
i'm convinced chatgpt don't like me because I followed this down to a T and no consistent characters. UGH! Thanks though.
Sorry to hear you're not getting the results you're after,.. I thought I really cracked the code with this method of achieving really good results repeatedly,.. that being said these bots do seem to have a mind of their own. In fact recently I was generating images getting really funky results with Dall E 3 so I straight up asked the bot are you messing with me... Because you have demonstrated the ability to give a perfect image representing my ideal,.. and the bot jokingly admitted to this!!
@@AImoney11 🤣. I will keep messing around with it and hopefully it feels sorry for me.
@@mizzKris Best of luck!
How to add, Layers to the script about the scenery, being able to give more then one design would be best.
Yes you must continue with your storyline and different scenarios as long as the images uploaded to your bot are consistent your main character will remain consistent.
Awesome, So far I have some luck with being consistent characters, Still working on the clothes.@@AImoney11
How to do a story with 2 characters? Looking forward to your reply
Just put your prompt into the bot for each separate character
Totally blown away. This guy is the opposite of the idiotic wannabe's out there trying to get attention for how to use AI and Dall-e 3 that I'm quite jaded. It was mindblowing to get REAL ADVANCED and useful instructions! Of course, let's go see if it works! ;)
Looks like I managed to stand out from the wannabe's! Have Fun!
Hard to say what I've created... but it is light years better.
I've CREATED some unusually advanced ChatGPT4 prompting and it's nice to add Dall-e to my répertoire.
@@curtismrasmussen next up is ideogram!
what is the gpt version?
This was done with GPT 4.0
do you have a class or something I can learn at a slower pace than what you did in this video? i don't get it, and want to learn for children's books.
Sorry at the moment I haven't established any personal classes or even a channel membership where I can give exclusive access but these things are developing for the future
It didn't work for me
Bummer, Yeah unfortunately this method is still not foolproof. My best advice is to condense and simplify your character's prompt and appearance if necessary.
@Almoney11 I did it it works I keep watching your videos over and over again
@@RogersEntertainmentBP Nice! Thanks for letting me know when I'm glad you got it to work for you!
How can prevent usage limit in chat gpt4. The limit is quickly reached that's frustrating. It's a slo cheating in a way as we have paid for the service we can't use all the tie.
Unfortunately there's no real way around the limit usage, you'll just have to wait the time allotted before you can generate again. That is actually the biggest downside to this method. That being said I have yet to achieve the consistency that you can hear anywhere else.
If only there was a way to do this via an API :'(
I haven't found the platform to accomplish this yet. When you find it let me know please!
I couldn't keep up with this. Maybe Im too slow but video lost me
Noted..
I appreciate the effort but I have tried everything on Midjourney and ChatGPT to render 3 characters cohesively across multiple scenes for a children's book... and despite doing all you said, I just spend hours to get 3-4 scenes done and they are only 75% cohesive. I'm frustrated. AI is cool for singular scenes but not books. I'm not trying to do War and Peace here with The Sistine Chapel as the art but damn if this AI simply will not get it right. The robot is too big, too small, looks nothing like what I've described nor on my uploaded graphics. The kids wearing Dad's clothes, the Dad has an outfit from God knows where that is NOT like the uploads nor is it in the description. UGH. Not working for me but thanks for trying.
Thanks, Sorry to hear that you are still struggling. One thing that I'm not sure I mentioned in the video is that the reference images should be single character reference images for each individual character. I found that if you start using multiple characters in the reference images it really gets confused quickly. Unfortunately this is the best that I could come up with,.. the other most important factor is the simplicity of the character prompt, the more complex the more easily confused the AI will become. If you do try again and get it to work please let me know, thanks again!
This doesn't remotely work at all.