Stable Diffusion, NovelAI, Machine Learning Art - AI art generation discussion and image dump

I am having a lot of fun with this program. One thing I am struggling with though is adding something to a scene that I already like. For instance, on this one that I generated, I am not able to add a katana leaning up against the tree. I tried making one in paint and running it through img2img but it didnt work. Oh well, I know this isnt a tech support thread but so far a lot of the other posts have really helped
 

Attachments

  • Blossom.png
    Blossom.png
    510.7 KB · Views: 123
It really does seem that the AI is only really trained in drawing anime women with big balloon tits, and even then, unless the anime character is facing straight forward, it struggles with figures.
That's unless you give it a really good prompt, it's as only as good as the information you give it.
It seems good at landscape, but rather because there's a lot going on in those pictures and you need to sit there and gawk at it for a while to pick apart where it goes wrong, the fine details it struggles with, but overal it's not too bad, bit samey... If you refine the images a lot, it becomes visually quite nice to look at, though some times it doesn't even know what it's trying to portray, it still looks cool.

Give it a prompt that is just nonsensical, and it'll come out with some truly horrific shit, give it a prompt where it is asked to create something that isn't common anime women or human characters, like mech warriors or knights in armour, or maybe even animals among different animals, it becomes uncanny, like it gets confused.
 
Maybe I’m going crazy, but people saying this will be a giant era-changing development sort of said the same thing about 3D printers (particularly being able to 3D print guns). Now 3D printers are out of the spotlight and barely discussed in MSM despite the massive promises hey have/had.

How does this relate and differ to 3D printers and us entering the “machine age”? Will this really be big or will it fade into the bg like 3D printing?

Just mentioning this because I had a shower thought of combining both this AI art and 3D printing to create AI sculptures and/or physical items/devices only a machine can come up with…
I mean, you're comparing an incredibly basic-level new technology with a ton of potential that is also currently limited in utility by about a thousand different very important things (material, size, cost, time, etc) to recreate actual physical objects out of something completely different than the physical objects currently exist as.

To replacing pictures made by a guy usually using a computer to a computer making the same images with the same material....with a computer...

So, yea, I think a program anyone can download on their computer (for free) and type enough words to generate things that would take other "trained humans" hundreds of hours to create (with the same material, for free) will absolutely have a huge effect. No one is going to spend a thousand dollars and wait a month for a guy to draw some backgrounds when they can do it themselves in half an hour for free (and likely better). Literally the only roadblock currently is your GPU cost and your willingness to read a tutorial. Both things can be sidestepped using an internet subscription, and will be more and more irrelevant in the future. Vs. spending thousands of dollars, hundreds of hours, and learning a new program to print some physical trinkets. lol.


An actual comparison would be machine-line automation (a machine doing literally the same job a human was with machine speed and precision), which absolutely wiped out hundreds of thousands of jobs and the need for people having to readapt. Yea bruh, maybe you are crazy. Or just a lil slow.
 
Last edited:
  • Like
Reactions: NoReturn
It really does seem that the AI is only really trained in drawing anime women with big balloon tits, and even then, unless the anime character is facing straight forward, it struggles with figures.
That's unless you give it a really good prompt, it's as only as good as the information you give it.
It seems good at landscape, but rather because there's a lot going on in those pictures and you need to sit there and gawk at it for a while to pick apart where it goes wrong, the fine details it struggles with, but overal it's not too bad, bit samey... If you refine the images a lot, it becomes visually quite nice to look at, though some times it doesn't even know what it's trying to portray, it still looks cool.

Give it a prompt that is just nonsensical, and it'll come out with some truly horrific shit, give it a prompt where it is asked to create something that isn't common anime women or human characters, like mech warriors or knights in armour, or maybe even animals among different animals, it becomes uncanny, like it gets confused.
People are just putting too many prompts all at once instead of few more accurate prompts, refining the image then adding gradually in a more layered way. If an image is slightly off what you want use photoshop or other software to fix it.
 
I am looking to enchance game screenshots, make anime girls etc, which version ia best for that?

I tried Dall-E but the differences in space marines was too much.

Like if I wanted to add a KF logo to my avatar or make a really nice screenshot of a space marine into a realistic painting?

Something that keeps all the little armour details consistent.
 
Last edited:
  • Like
Reactions: NoReturn
I am having a lot of fun with this program. One thing I am struggling with though is adding something to a scene that I already like. For instance, on this one that I generated, I am not able to add a katana leaning up against the tree. I tried making one in paint and running it through img2img but it didnt work. Oh well, I know this isnt a tech support thread but so far a lot of the other posts have really helped
I've had the same issue with trying to get certain elements, like a fence along a wooded path. Right now it's probably the point where the human element comes into play when it comes to small details, overpaint it or photobash it in yourself.

edit: very quick, something like this? i am not very talented and did this with a mouse

kwkatana.png
 
Last edited:
I am looking to enchance game screenshots, make anime girls etc, which version ia best for that?

I tried Dall-E but the differences in space marines was too much.

Like if I wanted to add a KF logo to my avatar or make a really nice screenshot of a space marine into a realistic painting?
Hugging face model for normal things, NovelAI for anime. You can use hugging face for all things, but obviously NAI was specifically trained for anime girls (and boys) and excels at that.
 
Look I understand the concern but Null who said he had less then 3 days experience with this software took this original image and turned it into what is close to a Disney Princess in a short order of time. Imagine if this imperfect AI drawing was hit with some photoshop to fix some of the inconsistancies. This technology is pretty new to people and they are doing crazy things with it. At minimum this is a huge shortcut and time save.
View attachment 3749182View attachment 3749186

Like if you do not think this is an insanely useful tool for any fucking person you are in denial full stop. Instead of getting defensive about it I recommend you learn how to incorporate this into your skillset.
Turning any image you put into it into a generic weebshit sameface waifu is literally what this program is designed to do (other than spit out generic weeb porn) . It won't impact actual artists whose art is more intricate than "same face anime woman from different angle with different hair and bust size".
 
Turning any image you put into it into a generic weebshit sameface waifu is literally what this program is designed to do (other than spit out generic weeb porn) . It won't impact actual artists whose art is more intricate than "same face anime woman from different angle with different hair and bust size".
This technology is pretty new and people are just getting familiar with it. Long story short the art world is gonna get impacted. Nobody except shitty artists care. Its not a big deal bro just do not ever buy AI generated art, if a new game uses it do not buy it. Vote with your dollar, at the end of the day most people do not care except pearl clutchers afraid of change.
 
This technology is pretty new and people are just getting familiar with it. Long story short the art world is gonna get impacted. Nobody except shitty artists care. Its not a big deal bro just do not ever buy AI generated art, if a new game uses it do not buy it. Vote with your dollar, at the end of the day most people do not care except pearl clutchers afraid of change.
I don't care if the deviantart artists who draw sameface waifus get impacted by this at all, their art is shit anyways. I don't like it because it's used to make child pornography of real children, and apparently the database it was trained on contained real CSAM.
 
I don't care if the deviantart artists who draw sameface waifus get impacted by this at all, their art is shit anyways. I don't like it because it's used to make child pornography of real children, and apparently the database it was trained on contained real CSAM.
You know, I don't read threads that upset me, and then whine in those threads that I am upset. Instead, I click on a thread that's going to make me laugh. I suggest you do the same, so you can stop shitting up this thread.
 
AI arts are nice and all but they still sucked at drawing HANDS.
You know the funniest thing about that? Hands are a common thing that real artists struggle to draw; its often complained about in art circles, and even professionals sometimes draw wonky hands from time to time. I wonder why the AI is also struggling with hands specifically?
 
I don't care if the deviantart artists who draw sameface waifus get impacted by this at all, their art is shit anyways. I don't like it because it's used to make child pornography of real children, and apparently the database it was trained on contained real CSAM.
Lets be real, this means that people like Shadman cannot profit off of hand drawn CP, if you think there was not a market for computer or hand drawn CP before this AI program existed you are delusional. At least now nobody can profit off of drawing it anymore.
 
  • Optimistic
Reactions: a feel
Back