Use this TIP to create consistent images in ChatGPT, DALLE
Vložit
- čas přidán 19. 12. 2023
- #chatgpt #dalle
Did you know that in ChatGPT, DALL-E, you can create consistent images with the same style?
In the past, if you generated a great-looking image, it was difficult to duplicate a similar image but with a different subject. You'd often get a different color scheme, different lighting, and even different people subjects or people.
But with this tip in DALL-E and ChatGPT you can create the same style of image for many different keywords that will maintain coherency and consistency.
PROMPTS
🎯 Mega Prompts Database: metricsmule.com/buy-chatgpt-p...
🎁 Hundreds of prompts: metricsmule.com/chatgpt-promp...
🤖 DALL-E Prompts: metricsmule.com/ai/dalle-prom...
Let's Collaborate:
📧 info@metricsmule.com
Follow me on Twitter/𝕏
𝕏 / metricsmule
🍿 WATCH THESE NEXT
▶️ Storyteller Prompt: • This Prompt INSTANTLY ...
▶️ The Prompt Perfecter: • This Prompt Makes Your...
▶️ Similar Images Prompt: • Create Similar Images ...
This channel is so underrated and I come here often to beef up my prompt art. Great content
I appreciate that!
Unfortunately as soon as we change for a new scene, the gen ID doesn't seem to work. The result is completely new style or new character...
Thanks man, consistent content too, respect from the uk
Thank you!
Thank you! Exactly the thing I was looking for!
Glad I could help!
Great video, I was looking for coherent image generation in DallE for sooooo long, many thanks 👍👍👍
Glad it was helpful!
Perfect bro exactly what i was looking for.
Glad to hear it!
Tried it by setting the style for a table with traditional roast beef meal and applying it to a table with traditional Chinese meal. The style transferred perfectly and there will be future occasions when I will need this technique and thank you for it.
That's great, and thank you!
OMG! I didn't even know this is possible! Thank you for this amazing trick!
You're welcome!!
This channel rocks and kicks ass
Thanks so much
I love this. The Cowboys and Eagles helmets are a great image.
Love the videos - you must spend a lot of time doing the editing.
Thank you. Yes, it takes some time but I enjoy it.
Not working anymore. It creates after 2 or 3 tries complete different styles.
I like this video, but by the way how do you make the picture speak? Is there a free website or app to do that?
Thank you. Good question. That's one of my favorite features in Canva. I actually made a video on exactly how to do it. You only get 1 credit/day (for free) but that's been more then enough for me because that will produce four videos a day until the credit resets. Here's the video: czcams.com/video/5fW-aQ7oKgU/video.htmlsi=rBwH4wa7__6r2emJ I'm not sure if you need to have a Canva Pro subscription but they're offering a month free right now, so here's a link for that too if you want to try it for yourself: partner.canva.com/LXyOaa
GUys can anyone help me pls? Everytime i use chat gpt to use similar style etc, for some reason the ratio 16:9 ratio always creates a very nice scene, but the main character always look towards me like, the viewers perspective, even if i provide instructions and say for the main character to do something, it never does it... anyone can help? I don't have this issues if i do like 9:16 ratio tho... thanks
Thanks boss
You're welcome
Can you make unlimited prompt generater prompt for pod typography tshirt design working in ideogram please reply 😊
vid starts here 1:55
In case if I want to elaborate a scene with more elements, will this Gen_ids work?
just aheads up, the gen_ids is just a placebo. Ask for the gen_id, then tell it that its actually wrong, it will give you another, and another. The LLM is simply giving you the next most probable word, you were expecting an ID, well here it is. A simple prompt with "keep the same style" should be more than enough.
Which software do you use to edit your videos?
Final Cut Pro
@@metricsmulethx ❤
Are you needing some recommendations? If so, I'll be happy to share a few things that help me out a ton.
is using gen id still working?
Yes, I just used it. Is it working for you?
Oouch! That simple?? I don't use Dalle, I'm more a Leonardo kinda guy, so I recently did a 20 page comic book depicting an adventure for my 2 grandsons, and the issue of producing consistent characters nearly sent me around the twist. Got any ideas for leonardo??? Anyway have a great christmas or what ever is your thing. Cheers
Leonardo AI is my favorite app as well. And that's awesome that you were able to create that comic book. Have you tried using the Image Guidance tool? It does a pretty good job generating consistent images. As far as comic books, my favorite combination of model and preset is 3D Animation Style with 3D Render preset. And then the other thing you can try is searching fine-tuned models, but search in the "Community Models" for the word "comic" There are actually 15 fine-tuned models specific to creating comic book style images. Let me know how any of those options work out.
@@metricsmule Great, Thank you. I did stuff around looking into some of the models but to be honest, christmas was approaching fast and the comics formed part of bigger present and the pressure was on. Job got done, but it's sparked my interest and so I have already got the next project in my head. Certainly take your comments into consideration. You know I might be an old fart, but this AI stuff is truly mind blowing. Thank you for your replies and content, it's great and usable. Cheers
Can I get the gen ID of a previous image and not just recent image??
Yes
@@metricsmule But how? I was unable to
@metricsmule This Great Video as usual do you have a prompt for comic strip-type Images I am interested in creating Financial Literacy Posts for Children. If you have them for sale please direct me where I can purchase them.
I have a few but I haven't tried them in awhile. Are you looking for an actual story-line comic strip or prompts to create animated-type comic characters, or both?
@@metricsmule both
Is Gen I'd same as seed
No, it is not. The Gen_ID is associated to the theme /style of the image.
The Seed is used to control the "randomness" (how different the new image will look). If generate an image of a red rose and get the seed number, you sometimes can use the same prompt (change red to blue) using the seed.
i.e. "an image of a blue rose" seed: 123456
That tells DALLE to use the same shape (limit the randomness) of the rose in the original image. Note: it doesn't work all the time.
dalle refuses to use gen_id ! what's the fix
?
Ensure that you're asking for the gen_ids not gen_id ;)
Great point.
watch from 1:55