It is indeed frustrating, some leeway should be given for the beta testing nature of the product, everything at this level in AI is cutting edge and we get to play with products as they are developed. Now I understand this can be a pain to deal with, but personally I’d rather have it in beta than not at all.
3 Likes
Very good point Fox. I should, and therefore will, be more understanding.
3 Likes
These aren’t perfect, but the prompts should put you on the right track. You want to avoid brand names or recent artist names, etc…
Use this exact prompt: “Sammy is a playful bear with wide eyes and jet black curls. Buster, a distinct robot, accompanies him. This is a landscape-oriented single photo illustration in b&w for a kids chapter book interio . Sammy and Buster are in a swamp.” n=1 Do not alter the prompt.
Use this exact prompt: “Sammy is a playful bear with wide eyes and jet black curls. Buster, a distinct robot, accompanies him. This is a landscape-oriented single photo illustration in b&w for a kids chapter book interio . Sammy and Buster are in a city.” n=1 Do not alter the prompt.
The robots (and bears) are off here, but with experimenting, you should be able to get something similar. As I mentioned, you want to keep some part of the prompt exactly the same then change the last sentence or scene/setting/actions.
Hope this helps!
3 Likes
What I’ve found useful is to leverage Custom Instructions.
Crate a fictional artist’s name.
In the about you section, put the description of the artist’s style. In the hope to respond section put the instructions to generate all images using the artist’s style, then in the message describe what you want to generate and add “in the style of <artist’s name>.”
It’s gotten pretty consistent results for me.
4 Likes
While still the same product - just another brand - i used your exact prompt via Bing/Copilot with a free/personal MS account, in Edge on MacOS Monterey (just in case someone wants to serve another example result for comparison on another system, this might help seeing if the platform or other variables make a difference in reality).
Dall-E 3 seemingly has a real long-term memory (vs. “just” the internet as a public reference and a median short term memory as well - aka makes use of comparative caching? The only alternative explanation would be heuristics are applied and the descriptive prompt you used is triggering creating functions that follow a so recognizable pattern that it edges on being “reproducible” (which btw i wish could be enforced somehow - in my personal opinion the holy grail - abstracting in real time with a predictable outcome + decision-making-ability that meshes all layers of processing/prompt-result-refinement).
Enough bla-bla, here is the result:
For comparison the difference when using the whole prompt, not just the content in quotes:
The focus is on the characters not the overall scenario when using:
vs.
… and yes, it is reproducible, i tested - why it does make a difference although is beyond me - any insights?
1 Like
I would like to emphasize the reproducibility-aspect. Could some of the more experienced users share some “templates” for prompting (meaning: the structure that makes it most reproducible).
On the same topic (earlier was mentioned, not to use modern artists - i assume due to copyright-issues - as “template”?): How does it look if i upload (and that is the keypoint of my question in this context) my own works (for example a character in different situations, all the time drawn in a recognizable manner) publically with an according license (CC/PD) - aka put it into human long term mempory/the internet?
Will this art be available to Dall-E right away, or does it apply some content-policy like must at least be published/viewed for n consecutive days or whatever?
It might seem i go offtopic, but i am in the exact same situation as the original poster, namely trying to create illustrations that share the same feel - just in my case i’d like to provide the general idea myself by sharing the art to be based on publically beforehand. Hence me treating this request as building on top of what was discussed.
Any insights appreciated.
Meaning: I have no intent of hi-jacking this thread, but would like to get your thoughts in the OP context - as a variation of the original question, that might lead to an answer for the OP in the end. I mean who knows how rudimentary an art is allowed to be to be of use to Dall-E 3 - so OP might be able to use this way of creating generic sketches/scribbles publically to guide/steer/template the actual request that way?
1 Like
I mean, at this point, I’m not sure they would be used to train new models. AFAIK, OpenAI used bought datasets with rights which is why DALLE2 quality was lower than other models that trained on a lot more photos/art/images.
I would just stick to forcing the exact same phrase for whatever “style” you want to hope it crosses over to the new scene or whatever…
The technology is definitely moving fast… Here’s some old GAN stuff from just a few years ago…
3 Likes
Good find! I’m having some success with creating characters such as in the OP. Somewhat less success if trying to describe a human character as I find you need to be much more descriptive with their characteristics (which, in turn, leaves you a bit less wiggle room to describe activity in the rest of the prompt)
Thank you so much Paul. I’ll give that a go this evening. I’ll let you know how I get on.
1 Like
Thanks Elmstedt! I was wondering how best to leverage the Customer Instructions. I’ll give that a go.
2 Likes
One note I’ll make. This is great for having a pretty consistent style but it doesn’t really work for perfectly consistent character design.
1 Like
Concur. I was able to get close, but there’s obvious differences if you look at them all together. The tech is coming along, though. I bet DALLE4 will have it… we can hope. Until then, I’m gonna keep tinkering with prompts.
1 Like
Here is what I am imagining we will see relatively soon-ish (2 to 4 years tops).
- Upload a reference image of a person or a sketch (or just a good text description of a character you want to create)
- Ask for a character model sheet to be generated of that character including various poses, facial expressions, and details of the character’s design, from multiple angles to give a comprehensive understanding of the character’s appearance and personality.
- Put that character model sheet (and others) into context for the model to reference when creating new images of that character.
I’ve not been paying too much attention to what the Stable Diffusion folks have going on these days, but even a year ago they were doing amazing things with consistent character generation, so I am sure this is probably already a thing or very close to a thing.
nexa
23
Actually there is a way to do this, but I think you’ll need to cross with midjourney. Here is youtube video on the process:
2 Likes
nexa
24
DALLE3 definitely does have its limitations in terms of graphical generations and photo manipulation, but midjourney is good for what you’re looking to do. So you can sketch or draft your characters on DALLE3 then upload image on midjourney to create a consistent character you’re looking for.
Yeah, I’m sure there are plenty of ways to do it.
I remember people fine-tuning LoRAs for individual, reproducible characters in Stable Diffusion upwards of a year ago.
Unfortunately, any commercial entity with deep pockets that wants to provide access to a generative image AI needs to fear the two threats of copyright holders and deep fakes.
So, no matter what the models are capable of there will need to be guardrails in place that will have the effect of reducing their utility.
nexa
26
You can request the seeds of the image and use that to recreate similar images with different variables.
for example you DALLE3 creates a superhero then you request the seeds from DALLE3 then you’re able to take the seed to recreate similar images or characters:
{
“prompts”: [
“Illustration set against the backdrop of Santa Monica Beach under a starry sky. The beautiful girl, with a clear face and no glasses, looks on with a mix of concern and awe. Beside her, the superhero is in a powerful stance, fiercely tearing away his tuxedo to reveal his superhero cape and uniform underneath. The scene captures the moment of transformation, with the fabric of the tuxedo rippling in the air and the emblem on his superhero uniform shining brightly.”
],
“seeds”: [4202394079],
“size”: “1792x1024”
}
but if you create a new chat and use the same seed it gets pretty wonky… so i recommend completing the character or goal with the consistency you’re looking for in one chat. I’m sure there is a way to use that same seed one month later to update the character, but you may need to tinker with it from then
FREG
27
Hey!
Just saw this long post in X trying to do this exact thing. Haven’t tried myself yet but try it out and see if it works.
Thread by @chaseleantj (Chase Lean)
1 Like
Looks like it might work if I ask Dalle to produce a grid of cards
1 Like
Foo-Bar
29
A grid prompt always works.
Create a “four panel comic about a fluffy bear” and the bear will be the same inside those four panels.
You can iterate on that image by using the gen_id of the image and refer to it by its referenced_image_ids parameter.
