41–50 of 86 posts
Earlier quoted context omitted.
I think it's just an informal term for things that seem to require human-level AI.
Ah, don’t care for it in that case. Seems like it’s cashing in on the formality associated with algorithms research.
But there is a paper about it: https://www.aaai.org/Papers/Symposia/Spring/2007/SS-07-05/SS...
Is there a way to have current AI tools maintain consistency when generating multiple images of a specific creature or object? For example, if there are images of 'Dr. Venom' they need to look similar, or if there are images of the same space ship.
Yes, right now you have 3 options: - dreambooth, ~15-20 minutes finetuning but generally generates high quality and diverse outputs if trained properly, - textual inversion, you essentially find a new "word" in the embedding space that describes the object/person, this can generate good results, but generally less effective than dreambooth, - LORA finetuning[1], similar to dreambooth, but you're essentially finetunin…
...but, all of these can't maintain consistency.
All they can do is generate the same 'concept'. For example, 'pictures of batman' will always generate pictures that are recognizably batman.
However, good luck generating comic cells; there is nothing (that I'm aware of) that will let you generate consistency across images; every cell will have a subtly different batman, with a different background, different props, different lighting, etc.
The image-to-image (and depth-to-image) pipelines will let you generate structurally consistent outputs (eg. here is a bed, here is a building), but they will still be completely distinct in detail, and lack consistency.
This is why all animations using this tech have that 'hand drawn jitter' to them, because it's basically not possible (currently) to say: "an image of batman in a new pose, but that is like this previous frame".
So... to the OP's question:
Recognizable outputs? Yes sure, you've already been able to generate 'a picture of a dog'.
New outputs? Yeah! You can train it for something like 'a picture of 'Renata Glasc the Chem-Baroness' now.
Consistency across outputs? No, not really. Not at all.
Is there a way to have current AI tools maintain consistency when generating multiple images of a specific creature or object? For example, if there are images of 'Dr. Venom' they need to look similar, or if there are images of the same space ship.
You'd have to scale the resolution on these waayyy down to not see the usual janky, smudgy, sometimes nightmare-inducing details. I seriously have never understood why what gets published in these blog posts isn't just lower res especially since this is precisely about old video game graphics.
[flagged]
You'd have to scale the resolution on these waayyy down to not see the usual janky, smudgy, sometimes nightmare-inducing details. I seriously have never understood why what gets published in these blog posts isn't just lower res especially since this is precisely about old video game graphics.
I usually don't notice those. It's only when someone mentions it.
Is there a way to have current AI tools maintain consistency when generating multiple images of a specific creature or object? For example, if there are images of 'Dr. Venom' they need to look similar, or if there are images of the same space ship.
https://huggingface.co/docs/diffusers/training/text_inversio...
I was trying this recently with the Sierra Christmas Card from 1986![0] The images that I generated were[1], and I was trying to tweak the model parameters with different denoising and CFG scales. When you get the parameters just right you can preserve the composition of the input image very well while still adding a lot of detail. This isn't a completely automatic process though, with Stable Diffusion you have to pr…
[0] https://github.com/Stability-AI/stablediffusion#image-modifi...
Earlier quoted context omitted.
Yes, right now you have 3 options: - dreambooth, ~15-20 minutes finetuning but generally generates high quality and diverse outputs if trained properly, - textual inversion, you essentially find a new "word" in the embedding space that describes the object/person, this can generate good results, but generally less effective than dreambooth, - LORA finetuning[1], similar to dreambooth, but you're essentially finetunin…
> Is there a way to have current AI tools maintain consistency when generating multiple images of a specific creature or object? ...but, all of these can't maintain consistency. All they can do is generate the same 'concept'. For example, 'pictures of batman' will always generate pictures that are recognizably batman. However, good luck generating comic cells; there is nothing (that I'm aware of) that will let you ge…
As for consistency of character details, I think that will depend on how many images you use to train dreambooth etc. and how varied those images are.[1]
Earlier quoted context omitted.
Yes, right now you have 3 options: - dreambooth, ~15-20 minutes finetuning but generally generates high quality and diverse outputs if trained properly, - textual inversion, you essentially find a new "word" in the embedding space that describes the object/person, this can generate good results, but generally less effective than dreambooth, - LORA finetuning[1], similar to dreambooth, but you're essentially finetunin…
> Is there a way to have current AI tools maintain consistency when generating multiple images of a specific creature or object? ...but, all of these can't maintain consistency. All they can do is generate the same 'concept'. For example, 'pictures of batman' will always generate pictures that are recognizably batman. However, good luck generating comic cells; there is nothing (that I'm aware of) that will let you ge…
For the animation stuff where you need frame to frame consistency, the new diffusion based video models show that it's possible [1][2]. These are not open source yet as far I know, but it's highly likely that we'll get them within a few months.