ChatGPT Consistent Character GPT - The Ultimate Character Consistency Hacks
Вставка
- Опубліковано 15 чер 2024
- In this tutorial, I'll share my top-secret hacks for achieving flawless character consistency using images generated by the Character Consistency GPT on ChatGPT's GPT store. You'll learn how to finesse character details, modify hairstyles, and seamlessly integrate multiple characters into a single, cohesive scene using Canva Magic Studio, Photoshop AI Generative Fill and my secret tool. Our goal is to create the best consistent results for your storybooks or films, making sure each character remains true to your vision across various scenes.
🌟 Join the waitlist for my upcoming AI storybook creation course and receive a 90% discount: forms.gle/tyKekRQAsNsCi5xL8
📝 Notes for this video: mia-meow.notion.site/Consiste...
◽️◽️◽️◽️◽️◽️
🔔 Subscribe for more tips like this🔔
/ @iammiameow
💌 Join my email list: miamew-newsletter.beehiiv.com...
◽️◽️◽️◽️◽️◽️
👇 GREAT TOOLS I USE
Best UA-cam Analytics Tool ➡️ VidIQ vidiq.com/digimeow
Easy yet Powerful Video Editing Tool for Free ➡️ Capcut capcutaffiliateprogram.pxf.io...
The Ultimate Image Control ➡️ Stylar AI www.stylar.ai/?via=mia
Create AI voice ➡️ Elevenlabs - elevenlabs.io/?from=partnerhodge151
Join the best newsletter platform ➡️ Beehiiv www.beehiiv.com/?via=mia-meow
Create Your KDP Book Content ➡️ BookBolt - bookbolt.io/5488.html
Create Your Own AI Talking Head ➡️ HeyGen app.heygen.com/guest/template...
*These links are affiliate links. If you make a purchase, I may earn a small commission at no extra cost to you. Thank you for your support!
◽️◽️◽️◽️◽️◽️
⏳Time Travel
0:33 The Consistent Character GPT
3:06 Bring together
5:42 My secret hack
7:31 Take it to the next level
◽️◽️◽️◽️◽️◽️
🐱About Meow🐱
My name is Mia and I share my journey with digital products, Kindle Direct Publishing (KDP), self-publishing together with the help of AI tools. With over a decade in digital marketing, and a vision of a world where entrepreneurship is accessible to all, I am on a mission of making it easy for everyone to start their online businesses with minimal resources. If you're interested in tips and tricks in working from home, making money online with AI, or building a one-person online business, subscribe and join me on my journey!
The information on this UA-cam channel and the resources available are for educational and informational purposes only.
◽️◽️◽️◽️◽️◽️
If you like my videos, you can also buy me a coffee☕️ as a small treat for me❤️
www.buymeacoffee.com/miameow
#dalle3 #photoshop #generativefill #canvamagic - Навчання та стиль
![[The NO Prompt Method] MULTIPLE Consistent Characters with Custom GPT & DALL-E](/img/n.gif)
Do you find these tips helpful? What are the other unanswered questions? Let me know your thoughts about these tricks or some better ways of achieving the same result 😀
I found this video very helpful… I’m just about to watch it again.
Thank you! Glad it was helpful 😃
I followed your tips and created my first AI generate video in my channel~ Thank you!
Happy to hear that! Congrats on your video/channel! 😄
Hi Mia,😊
You got another subscriber! Amazing work, very helpful video. Thanks Mia.
Thank you, glad it helped! 😄
I wish we could combine multiple consistent characters together with a consistent background in order to create scenes for stories. Just one consistent character isn’t enough for creating full scenes. We need control over multiple consistent elements.
Good point. I am thinking about how to solve the issue of consistent background as well. Maybe there will be new tech next week for the rescue 😬
@@iammiameow Hopefully, consistent backgrounds will become possible soon. Also, I hope we’ll get the ability to pose multiple of our consistent characters together in an image, as scenes often require characters to be interacting with each other. Just having a standalone pose of one consistent character usually isn’t enough.
However, I think backgrounds might be quite a bit harder than characters for the AI to accurately depict consistently. This is because backgrounds tend to have a lot more variety in different elements than characters do, which could make it especially difficult for the AI to understand the 3-dimensionality of backgrounds based on image training data alone. Maybe if there was a way for the AI to also be trained on 3D models, we can show the AI a 3D background that we want to depict consistently, and ask it to depict that 3D background as a 2D illustration. This way we can pose our characters in that background from different camera angles, and get accurate results.
Good thinking! Thanks for the insight. Yes, I think a 3D Model needs to be created for us to get consistent background result. Hopefully someone will resolve this issue soon.
@@iammiameow At this point, they’re trying to get AI to create a consistent character by using a single image as reference, instead of training on multiple images. I think this improvement will help with character consistency, but I don’t think it will be enough for total character accuracy. Since AI can’t view all angles of a character from a single image, there’s many things that it probably won’t guess correctly, because it cannot see those things. So, this is where I think 3D could be helpful, assuming the AI could be trained to create 3D models from images, as well as to analyze 3D models.
The workflow that I was thinking could work is this. We show the AI a 2D image of our character. The AI creates a 3D model of our character based on the image reference we provided. After that, we have the option to tell the AI how we want to edit the 3D character generated, in order to better fit our reference image. Once we edit the 3D model and approve the final result, then the AI can analyze that 3D model from all angles and use that information to generate consistent images of our character. In other words, it would learn from the 3D in order to create more consistent 2D images. This is just a thought, but I haven’t seen any AI that does this yet.
I think this would be especially helpful for creating consistent backgrounds, since backgrounds are even more complex than characters.
Meshy AI and Zoe depth on hugging face seem to be the solution to your questions 😀
Another excellent video, Mia. Indeed, you deserve a lot more subscribers ;)
Thank you! ☺️
True
I am currently on my new book's cover, that's why I looked up AI picture generating techniques. Your video was exactly what I was looking for and if this turns out to work well, I will mention you in the notes of thanks!
Wow, that’s awesome. Congrats on the new book. I’m happy that it helps 😄
I just subscribed because it bink this is really helpful and informative... ❤ Thank yiu for sharing.❤
Thanks for subbing! Glad it helps 😊
This channel should deserve more subscriber
Thank you！😊
Awesome and thank you!
We need more input about AI tools in drawings!
Thank you! 😀
I'm looking for resources for character harmony. You've had your match. Your methods are truly excellent. It was a very successful work in terms of character performance.
Thank you! I’m happy that it helps 😄
Thank you, this is great!
You’re welcome, glad it helps 😄
Nice tutorial video.....Well explained .....😃👌
Thank you, glad it helps 😄
Thank you so much for the effort you put into making these tutorials for us! Just a personal preference, I rather hearing your own voice as in the previous videos instead of the AI voice you've tested in this one 🙂
Thanks for the feedback! It does sound mechanical. But it was actually my own voice recording the whole time! It is funny how the AI is so good that real voice can sound fake now. 😂 I’ll remember to be more casual next time for sure! 😁
@@iammiameow Wait is it your real voice? That's crazy hahah don't worry about being more casual, just be yourself!
@@vencinar Thanks! I am flattered. I thought my accented, unstable voice is definitely not gonna be mistaken as AI voice but I made it lol
thanks for sharing! great video!
Thanks for watching! 😊
Future video idea: ":How to make image-to-image consistent characters!?!" Oh, and love your videos.
Thanks for the idea! Noted!
Great video as always! I hope you make a video about midjournay consistent charachter in ninji6
Thanks! I’ll look into it! 😃
You use low variations option. But they haven't rolled that out yet for v.6
😊
Thanks you so much, this is helpful
You’re welcome ☺️
Great update. ❤
Thank you ❤️
I tried the GPTs you recommended. It's amazing! By the way, how did you get the seed number?
Thanks for watching. You can simply ask the GPT: what’s the seed number of this image. 😀
I love your tutorial. Could you also show how to make them move other than using runway?
Thank you! What type of tools are you thinking about?
Thanks, very useful
Glad to hear that! Thanks for watching! 😊
Great video! I didn't think the PS version would make such an improvement, but it really blended the scene together perfectly. Do you know if there are any type of Copyright restrictions when using Generative Fill? I assume it's been trained on Adobe Stock which is a paid service as well. I've gotten accustomed to not worrying about it with other platforms. 😀
Thank you! I know that generative tool in Photoshop beta is for personal use only but you can use Photoshop 2024 for commercial use. Copyright is a more complicated issue just like any other AI created work. I think it is an evolving topic that I don’t think I have the knowledge to say yes or no on this one.
😊
좋아요. 쉽고 친절하시네요. 저도 도전할거에요.❤
감사합니다🙏
The Canva and Photoshop techniques are so user-friendly.
Thank you!!
thanks, it really helps
You’re welcome ☺️
Excellent work❤
Thank you ☺️
Amazing!
Thank you!
Thank you for your help!
You’re welcome!
@@iammiameow I wrote a children's book a few months ago. However, character consistency was an issue I kept running into when I was trying to create illustrations for the story. This video was a huge help!
I’m happy to hear that! 😄
unreal. this is cool
Thank you!
Hey Mia! That you for such an awesome video. I just want to know if there is any limit for image generation in Chat GPT plus subscription??? Just like midjourney!
Thanks for your kind words! I don’t have a definitive answer as it’s changing all the time but it seems like there’s a 20-50 images per day limit and 40 message per 3 hour window for plus user. The token for image and text are separate.
Nice video tutorial 💯
Thank you 🙏
Right
very cool video, thanks
You’re welcome 🙂
Thank you
You’re very welcome ☺️
Fantastic
Thanks!
Great video! Thank you! But how were you able to view the seed number of your images in Chat GPT 4?
Thank you! Just ask the GPT in the chat box: what is the seed number of this image? 🙂
Great 👍
Thank you 😄
I keep trying👍👍
Thanks!
very useful😀
Thank you 😄
In ChatGpt use generation ID and not seed, when you have made your character then ask for the Generation ID and you get a genID you can use forward, in your Prompt design, and end it with, use the Gen-id "put in the number you get from the first"
Yes, normally you can’t assign using seed but somehow that GPT seems to be able to do that. I mentioned the seed just because that’s what the GPT told me what it used to stabilize the consistency. If you are promoting yourself, GEN ID should be used and not seed. Thanks for mentioning!
Could you do a video where you go through this process of creating consistent characters and then using D-ID (I think that’s what it’s called? 🤔) to make them talk in an animated scene?
That's a great idea! Noted!
Lovely jubbly 👍🏻
Jest wow mam
Hi! Is it possible to incorporate these characters in video with myself in them, kind of like Blue's Clue to make the comparison? Thank you!
Hummm
If we add the detail in ChatGPT to include shadows cast beneath each character within the scene will it be able to do that? Or do we have to add that separately in another app?
And is there a way to scan an existing image or photo that we want ChatGPT to base a character on so it generates a characterization of that person or creature? Thank you.
Hmm, I think most of the images generated will have shadows. The part that I added shadows was because I was merging the characters with a background. It’s not that important if you are not mixing characters to background.
You can upload a reference image for ChatGPT to read and it can summarize what it reads as prompt to create an image but it may not look the same though, since the input of the image will be converted to natural language to feed to DALL-E as prompt, but you can try. Remember it has to be a paid version. 🙂
Great job! do you know if we can acces to the gpt via the openai api?
Thanks! It seems like you can from this article but I haven't tried it so I am not sure www.openaigeek.com/how-to-access-openai-gpt-store-features/
Is there any way to create layers for each part of the character? or maybe vectors of the characters?
Great question! I think you are probably looking for something like this: sketch.metademolab.com/ or you can use the Puppet Tool in After Effect.
Tip for creators: try to use 2 characters in every scene. 3+ characters and dall-e goes bonkers
nice
Thanks!
i really wish when people make videos they would give more detail as to how to get the chat gpt store. is this gpt pro? because its not in the unpaid version
Yeah, sorry. It’s the paid version. It was briefly mentioned in the video.
@@iammiameow ok Ty didn't hear it
Hello. To make a story video, both character and scene are required. Is there an AI that makes consistent scenes? There are always consistent character tools. I couldn't find any artificial intelligence to create a consistent scene or consistent background.
This is a great idea. I think the closest I can think of is to turn your background into 3D. Something like meshy AI may solve the problem.
This is great but does it work the same for multiple human characters ie, more than 19?
Yeah it is harder if you have too many characters but yes you can create multiple human characters just opening different chat sessions with the consistent character gpt. The more you want in a the same scene the harder it is for sure.
Photoshop AI Generative 👌👌👌
Agree 😀
Is this feature only available for plus or team upgraded membership?
The GPT is only available for ChatGPT plus.
if you are to make a whole book to sell with this process, do you have to give credit to AI as part of the book? like: Story created by:AI and illustrated by: AI?
That's a great question, and the answer is still evolving as AI usage in creative fields is relatively new. Different platforms have varying guidelines; for example, Amazon KDP now asks about AI involvement during publishing but they haven't required you to put that in your book. While there might not be set rules about crediting AI in your book (yet), you should follow the specific guidelines of your chosen platform. 🙂
@@iammiameow awesome 👏 thanks for answering 👍
You’re welcome :)
Hi, I’m loving your video, but wondering if you can help, please? I can find the consistent characters prompt you’re mentioning, but very strangely I couldn’t get it to work. I now see in your video that there should be a prompt box/bubble with “CLICK HERE to start creating……” which doesn’t show on my paid version of ChatGPT. Thanks if you have any idea why I can’t see/use that? 😊
Thank you! Did you find the "Consistent Character GPT" in "Explore GPT"? I just tested mine and I still see that there's a "CLICK HERE to start creating..." prompt. I am not sure why it is not working for you 🤔 Maybe make sure you get to the right "Consistent Character GPT" and not just on the regular chat with ChatGPT? And yes, you have to have a paid version to access it and that's the one I have, too.
@@iammiameow thank you so much for taking the time to reply. After I message you here, I realised I could contact the creator of the prompt. They kindly replied and I'm trying it out now.
You’re welcome!
This is amazing! But I dont know how to use Photo shop 😢
You can also use other tools with inpainting function like Leonardo’s AI canvas editor to achieve similar results :)
@@iammiameow I have Leonaldo, but I am barely scratching the surface. Seen tutorials to learn a more how to use it effectively.
Leonardo is great! Lot’s you can do there for free :)
Chat gpt with dalle e is not free for commercial purpose I think right?
The images generated by ChatGPT’s DALL•E can be used for commercial purpose but yes, in order to access that you need the paid ChatGPT.
Does this only work with human character or does it work with animal as well?
If you go through the process of this GPT, it will create a human-like/personified animal. But you can also ask it to create animals without going through its set questions - it will then give you normal animals.
I love you 😍
lol thank you 😂
@@iammiameowWelcome to my heart 🫣🤗
why i click Explore GPTS dont show discover bar for search Consistent Character GPT
🤔 Not sure why that is. But if you go to the DALL·E section, you should be able to see it as well.
It requires an upgraded plan, is this correct?
Yes, this has to be in the paid version.
Also generates half body any turnaround?
Just prompt with "half-body portrait" I guess
lol you required chatgpt for consistence? its all depends on naming and seeds!
Thanks for the insight. You don’t “require” ChatGPT for consistency. Naming and seeds are for midjourney and Leonardo. This is just an alternative way if you want to achieve similar results with DALLE.
Can I generate these characters in leonardo ai for multiple images?
Yes, I think you can
ok thanks@@iammiameow
can i still use my laptop chrome windows8. I can't afford to buy a new laptop...😢😢😢
I guess as long as you have internet, lots of things can be done on the browsers :)
훌륭합니다... 한국분 이신것 같은데...?
내공이 느껴지네요... 명상하시는가 봅니다.?
Thank you for the kind words! 😊 I am not Korean, although lots of people think I am 😄. I’m glad you enjoyed the content. And yes, I do try to meditate sometimes!
Plz tell any copyright issues?
This is from OpenAI: Can I sell images I create with DALL·E?
Commercial use of DALL·E
Subject to the Content Policy and Terms, you own the images you create with DALL·E, including the right to reprint, sell, and merchandise - regardless of whether an image was generated through a free or paid credit.
Here is the link: help.openai.com/en/articles/6425277-can-i-sell-images-i-create-with-dall-e
To day , Consistent Character GPT Skip questions for selection
What style or theme should the image be in? You can choose from:
1. Pixar Animation
2. Photography
3. 2D flat illustration
4. Anime pastel
Has anyone encountered a problem like this ?
I just tested it and it seems like it’s acting normally. Can you elaborate the problem?
@@iammiameow I capture this problem in my share drive
drive.google.com/drive/folders/1A7_WBEo1dos5vSiRhnsIhf-_z6TE8jOA?usp=sharing
How do you find the character’s seed number?
Hmm...I just asked ChatGPT for it
I asked Chat for it and it told me it couldn’t tell me that information. I think it was just being difficult, 😊
The creator may have changed the setting 😅
Can we use it's in a mobile phone
If you mean the consistent character GPT, yes, you can if you have the paid version.
@@iammiameow Yes Thanks meow for reply.
You’re welcome :)
why the chart gpt nor working
In what way? The creator sometimes update the GPT and it may be a bit different :(
@@iammiameow it's says it can't generate anything at this moment
Everytime I use it, it makes the head as big as a balloon and the body like a pencil.
That’s a funny style 😂 maybe try a few different times or double check on the prompts?
Why cant we create multiple gtp bots to handle each character?
I don’t quite understand the reason you want to do that? Could you please explain a bit more?
@@iammiameow i want to have more than one consistant character without confusing it. I thought i could have one gtp bot per character . Or is that not good
Do you mean to create individual characters and combine them together? I think the confusion is mostly happening at the DALL•E level and not so much at the GPT so separating the GPT for each character may not make too much impact. When generating images with multiple characters or too many specifications, DALL•E gets confused at this point. If I understand you correctly, you may want to try the consistent character GPT I mentioned in my other video and combine the characters together using other tools.
You are also made with AI 😮
Possibly 😬
I can’t figure this out
Which part is the part you have question with?
Chatgpt 4 i think not free?
Yeah, it has to be a paid version. 🙁
The character GPT doesn't really work well, at least for me, it was all over the board, similar, but not usable in any commercial way.
Sorry to hear that ☹️ It seems like the creator of that gpt has edited a few times and I haven’t tried the latest version. A lot of people seemed to have created some consistent results before though. Maybe midjourney’s consistent character function can yield better results now 🤔
Consistent CGPT stuck after i provide details about a character who lives in a place nortern southern dynasties :D
way to confuse DALL-E lol
I think they are not giving the seed number anymore
I’ve heard that’s the case, but I think it probably wouldn’t affect the results?
@@iammiameow For now, I think still working pretty well! By the well, thanks a lot for sharing this tutorials! Is really nice of you!
You’re welcome, glad it helped 😃
Is it just me or does it seem like whilst ai is geoundbreaking and marvelous, its also very disappointing t the same time. The fact it cant draw the same character twice in a diff. Pose seems mindboggling.
I get what you mean! I think this just shows how incredible our cognitive abilities are and how much more we have to learn about ourselves. Also, many other tools have improved on their character consistency ability recently, like Midjourney and Leonardo. You can check them out!
This process is done with chat GPT 4, it is not free 😐
Yeah, it is not free, unfortunately.
I have a character generated in Leonardo but cannot get variations at all in Leonardo. Can you copy that character into chatgpt instead of typing a description.
The closest I can think of for you to do is to upload the image to the paid ChatGPT and ask it (DALL•E) to create something similar. It’s almost the same as typing the description. But definitely won’t be the same though.
AI involved so many ways in design designers feel so useless😣
I think designers are in a better place to master AI art than us normal people. The eyes for aesthetic still require experience and training to put things together in a harmonious way. Just need to focus on different skills but the essence is still the same in my opinion 🙂
@@iammiameow i'm sure industry will ask for candidate who know how to train AI rather then candidate who know how to design manually 🤦🏻♂️ but hear and there there is always be a humans who control AI in his way🫡
True!
consistent character gpt is not free, it is a paid version
Thanks for mentioning. That’s right.
Ai is not a skill. Ai users use copyright of artists.
Thank you
You’re welcome ☺️