I have been playing around with AI art generators. I think I have a good handle on creating decent characters so if anyone wants I can try and make custom character art for you for your game. If interested just rMail me the details. The more descriptive the better but also understand that AI art generation is very hit or miss. I often have to generate hundreds of images to get a couple of keepers.

BUUUUUT i can have it running while I am asleep or at work so its not a huge effort and it helps me practice tweaking settings.

Anyway what I need in the mailer is:
1) Text description. The more descriptive the better. Also consider the background, the art style, the lighting etc. Feel free to  offer thoughts on poses or body types or hair styles or whatever really. Just understand that some or none of it might directly get into the final image. Poses especially can be difficult, hands are notoriously problematic (the AI doesn't really know how many fingers go on a hand and often kinda wings it) and, as expected, nothing NSFW or naughty or lewd.

2) Let me know if I can post it here in general or not. i know everyone would love to see what comes out but i get some of these can be personal. I would love to be able to showcase it and illustrate how I can go from the text prompts into the final product but it's not a requirement. By default I'll post the details here but if you want it kept private just make sure to let me know.

3) Understand that sometimes it doesn't have the training material to do a specific thing. I have found that certain poses or combinations just don't work out. It is trained on hundreds of millions of art but it is lacking in many ways. I tried to do this for a Star Trek game for example and it just doesn't really get star fleet uniforms and I spent all day trying to do a proper klingon forehead but it just won't stop turning that forehead ridge into a headband or tiara.

4) Also understand high quality art is a tremendous effort. The stuff you might see online for AI art is often a combination of multiple generation processes and clean up stages. It is still a lot faster than human artwork but the highest quality stuff can easily take several hours on top of extensive training on how to write prompts and set settings. I am not an artist, I'm just...decent at getting the AI to kinda do what I want if I let it chug overnight.

Don't expect high resolution, don't expect perfection, don't expect it to be exactly what you see in your minds eye. That being said it can be really cool and for the purposes of a tiny portrait for the game or even just something put into your bio, it should be fine. The images are going to be small, probably 256x256 and that will help hide things like weird hands or bizarre tiny details.


5) I have no idea how popular this will be so I'll try to process these as they come in but if I wake up tomorrow with 1,000s of these  there might be a lengthy delay before you get your hands on it. If I can nail the prompt I can probably turn something around in a couple hours but if it is exotic or very specific it could take me days of playing with settings and stuff to figure it out. Today I tried to make a Klingon warrior and a "N'avi" style alien and after about 6 hours of work I have...acceptable images I can use:

Male Klingon Warrior (probably the best attempt at getting the forehead right...)
https://i.imgur.com/HDUnE6v.png

Female Klingon Warrior:
https://i.imgur.com/E1BB8En.png

N'avi or Kaminoan style alien in a toga:
https://i.imgur.com/2Fn06pV.png

This one came out a little to Navi for my tastes but still looks good:
https://i.imgur.com/Y9I9dpn.png

This is a failed klingon where the forehead became a headband that became the chestplate for another klingon. It still looks like a really cool composition though:
https://i.imgur.com/SDcSjUV.png

---

tl;dr: via rmail I need

1) Description of what your looking for. Character looks, clothing, art styles, background, poses & action. The more details the better but also know that many of those details are going to get warped or fall off.

2) By default I'm going to post the images and prompts (no names though) but if you really want it to be private let me know.

3) No adult stuff. PG only. Also try to shy away from being toooo deep into an IP. The AI can do stuff that looks "like" an IP (see the klingons above) but it is almost impossible to get it to replicate perfectly a Star Trek uniform or a Forgotten Realms religious symbol etc.

First attempt!

Prompt: male fantasy rogue, middle-aged, red hair, greying hair, thin beard, perfect smile, detailed face, dark pupils, detailed pupils, fantasy art, character concept, trending on artstation, black cloak, colorful clothing, scarves, gold earring, jewelry, braclets

Result: https://i.imgur.com/zisMAsA.png

praguepride that is AMAZING! I'm so envious!

All I do is create PC characters using Heroforge.

While my work will pale in comparison to Pragueprides, in the spirit of community networking, if anyone wants me to create a portrait image for your PC using Heroforge (which is far more limited in variation than praguepride's AI stuff) I'd be happy to do it.  The more detail you give me the more I can work with.

Also, I have a 3d printer, so if you want me to PRINT a 3d resin print of your PC, I'm happy to do that, but I just don't want to (a) paint them; and (b) pay the $7.99 Heroforge charges to send me the STL and (c) pay for the mailing cost.

But the portraits are free!  :)

OK, I am curious, so here are two characters to try:

Character is a Fencer from roughly the three Musketeers era- but no Surtout.
20 year old Male with long auburn/brown hair tied back
Green eyes, tanned skin although natural pink/white elsewhere
Amidextrous.
Wears a rapier (with cutting edge) and a Main Gauche (Large Knife with large protective grip)
Leather helmet, light scale on Torso, bronze Armbands just above wrists, Blousy sleeves. woolen breeches and knee high soft boots.
Wide leather belt with small pouch, scabbards for rapier and main gauche.

Second Character is a Female Space ship Captain-
No uniform
35 year old tall, moderately full figured. [5’ 8” and 157 lbs]
Shoulder length black curly hair, held back with tie or band
Green eyes, alabaster white skin.
Grey ballistic suit (form fitting) with attachment ring for space helmet and pouch for First Aid Kit on left thigh
Assault boots
No gloves (on belt) and Survival watch on left wrist.
Belt with a gas mask, powered  holster for 10mm gauss pistol and magazine pouch for 10 spare clips or 10 shots and medium communicator.

Thanks

So this  is a little premature because it is only about 1/2 through the image rendering but here is a sample of what it can do:

Based on this description:



This is my favorite so far: https://i.imgur.com/7kq9wPS.png

The hawk on the shoulder is a big trouble. If I do this again I'll probably just have it flying overhead because the  AI likes to merge the "staff" and the "hawk" so here is a humorous reject showing what I'm dealing with XD

https://i.imgur.com/fEw1ogq.png

That being said, I really like the look of that, hawk-staff and all :D

Praguepride, how long does it take you to make these images?
What is the site you create these on?  I have SOOO gotta try!

I'm using a local instance of stable diffusion (automatic1111 version). It's free to use but it took me several days to get up and running because you will likely need to be somewhat proficient in Git and python to get it up and running. You also need a decently powerful computer, at least 4GB of GPU to do the processing.

After that it is spending a week or two getting used to the controls and settings so that it makes the things you want to do, doing research on things like GPT and CLIP so you can figure out what the  computer "sees" when you type in a prompt.

THEN after all that (lol) it takes about ~1 minute an image to render so I do batches of about 50 and then just pick the best ones. Other people instead do a lot of refining via other AI engines or photoshop. They will kitbash  the stuff they want and continuously re-render and cherry pick the best ones in a  feedback mechanism (take the best of the batch, feed that in as the starting point of the next round etc.)

BUUUT I just like to run ~50-100 images and take the best couple and call it a day :D

The mods tend to not like extensive linking to other websites here so google "git automatic1111 stable diffusion" if you want to go down this rabbit hole.

OTHERWISE if you look for web portals there are people hosting SD servers but those aren't as  easy to just "fire and forget"

edit: One more "staff as bird" but in a really  cool way: https://i.imgur.com/TeVe7rt.png
This message was last edited by the user at 21:17, Mon 17 Oct 2022.

Decided to do a 3D render for this one:



The AI decided it really liked green neon lights. You can kind of tell that based on my prompt it took "green eyes" and kind of...missed the eyes and added green everywhere. I can re-render in different styles if 3D renders aren't your cup o' tea and fix that but after some post-processing through GPFGAN to fix their eyes here are some good results, I think:

Black armor: https://i.imgur.com/vSwCh7T.png

Black & Green armor: https://i.imgur.com/2A7SD3r.png

Heavier armor and some curls are loose!: https://i.imgur.com/azTDQ0c.png

Space Iron Man Armor: https://i.imgur.com/XiplF9m.png

And here the AI decided that "grey armor" means white: https://i.imgur.com/35ffSpV.png

Final one: https://i.imgur.com/tWZ32Fj.png

I generated 50 images. It took ~ an hour and here are the ones that looked pretty decent.

I don't know when I'll get to the next one of yours. I did request an rMail so its easier to keep track of.

---

Also going forward if there is a style you WANT or DON'T WANT let me know. Typical broad style are 3D renders, digital art, anime, and oil painting. The AI is specifically trained on art so if there are specific artists or mediums that you want ("Make my character out of knit yarn!") then let me know.

OR converesly if you DON'T want a style let me know. I'm going to trend towards high quality, realistic 3D renders for sci-fi stuff and fantasy painting style for fantasy stuff by default.

Hey praguepride! Don't know if you remember me or not, but I'd love to see the AI's attempt at Renee from one of our past games together!

Young Albino woman.
Bright neon-green eyes.
Messy shoulder length hair.
Extremely thin, willowy build.
Long neck.
High cheekbones.
Thin button nose.
A third eye in the center of her forehead.
Long white gloves.
White patterned sweater.
Shin-length white skirt.
Simple leather apron.

Would love to see her in both an anime and pixar style :D
This message was last edited by the user at 23:02, Mon 17 Oct 2022.

I forgot to share the prompts I'm using as well. A little insight into how to translate a description into something the computer can understand and then see how that effects the outputs.

Also feedback is welcome! I'm not actually making any of this so if you think it sucks just tell me! The feedback will help me write better prompts for next time.

Anway this is what I got:


Which I translated into this prompt:

Positive: professional painting of a (((beautiful female))) ((warrior princess)) holding a saber, ((pretty face)), (detailed pupils), (golden eyes),  ((full body)), (dark complexion), long straight black hair, leather armor,  (trending on artstation),  epic artwork, high resolution, (realistic skin texture), masterpiece

Nevative: (((Poorly drawn face))), (((cross-eyed))), (((blurry))), (long neck), (((deformed))), ((close up)), bad anatomy, bad proportions, ((malformed face)), (malformed body), ((fused fingers)), extra faces, extra fingers, extra arms, extra legs, amateur art, missing arms, missing legs, missing head, amateur art

The parenthesis helps its weight and the order also gives it priority. The way it works is it comes up with a random noise and then slowly pushes and pulls it pixel by pixel towards what it thinks you're asking for.

Anyway enough chatting, here are the results:

Not the way to hold a sword: https://i.imgur.com/MHz2dyS.png

She is holding an arm that is holding the sword: https://i.imgur.com/y08LlNG.png

And this is why I don't ask the AI to try holding things: https://i.imgur.com/KrYV9AQ.png

So close! so very close: https://i.imgur.com/LZ8flV9.png

Anyway I'm going to re-bake this one. See if I can't tweak some stuff and get something that is completely awesome and not just ~75% of the way there...

There we go: https://i.imgur.com/bzKkWMI.png

I said I wasn't going to do any post-processing but I did a little because it was a difficult pose and it was sooo close. So what I did was photoshop a better hand & sword over the flawed ones and then had it re-render (via img2img) the scene.

Then in the 30 images I re-rendered there was one that was almost perfect (had that headband jewel thing hanging out in space on the left) so I could just wipe that away because I had 27 other skies to pick from and here we go.

Pretty proud of this one. Hope it floats your boat!

Absolutely amazing!  And yes it floats my boat big-time!  Thanks so much Praguepride!

Praguepride, that is astonishing work!

On this journey we're going to talk about CLIP. CLIP is trained on a dataset of hundreds of millions of images so that if you type in a phrase like "warrior queen" or "albino woman" it will find images that look like that idea even if the images aren't specifically tagged that way. This lets you search for things that aren't even in the set of images.

This is part of the process of getting the AI to generate art so for this request:


I went to a clip search and  looked at what it would present with "albino" vs "albino woman" and so I ended up with albino woman as a keyword and bumping up green eyes to offset the dark or red eyes that are typically associated with albino pulls.

(full body) portrait of a young (albino woman), (((third eye open))), finely detailed features, professional art, ((thin)), (cute face), white shoulder length hair, (((bright green eyes))), wearing white gloves, long white skirt, leather apron, trending on pixiv fanbox, painted by makoto shinkai takashi takeuchi studio ghibli, akihiko yoshida

(((Poorly drawn face))), (((cross-eyed))), (((blurry))), (long neck), (((deformed))), bad anatomy, bad proportions, ((malformed face)), (malformed body), ((fused fingers)), extra faces, extra fingers, extra arms, extra legs, amateur art, missing arms, missing legs, missing head

But really the hardest part was the 3rd eye on the forehead. That is not a good feature to pull from because there just aren't enough images or specific terms to describe it and sadly the dataset images are "polluted" by generic spiritual symbols that water down and obscure any actual artwork that has a third eye on the forehead. No question about it, trying to put an eye where an eye shouldn't go is really really really hard. But I had a crazy idea that just might work...

https://i.imgur.com/JmmOwuL.png

Photoshop! Grab an eye from another render, do a little tweaking and you have something that's...good enough. I generated well over a hundred images for this one and never got even close to getting an eye to open up in the forehead. It kinda knows what I'm saying but just has zero reference to go on so it just skips over it. Anyway, that closes out this one.

The interesting thing is that ultimately, you had to do pretty much what I did to generate the original portrait image I used for the game.

http://www.carakav.com/Renee1.jpg

The process was to basically do a bunch of google searches until I found a base image that was 'close' to what I wanted (but not perfect: in this example, the hair is too long, and the original woman was not albino), then I essentially went into Photoshop and had to clone-stamp the eye and make hue/saturation adjustments to get the skin tones right.

Thanks so much for your work on this!

Yeah. The process of how the AI works is fascinating to me and one of the reasons I did this was because I did want to see the limits of the AI.

It can do wild and amazing things but there are definitely ideas (like an eye in the middle of the forehead) where the AI goes "lol no".

I suspect part of it is the medium. I might take another shot at it but trying different art styles. Part of the problem could be that anime is too "clean" so it is hard to deviate from the very fixed template of face structure whereas with "fantasy art" or "digital art" or "3D render" you have more room for variation.

With some very minor photoshopping it was able to do this: https://i.imgur.com/eElb2y4.png

I'm letting another batch back to see if it comes up with anything better but that is pretty damn good given how much it struggled with the 3rd eye.

And for those curious, this is the original "pre-fix" but keep in mind it's kinda nightmarish...kinda.

Check at your own risk: https://i.imgur.com/RCxU4eF.png

More horrors that will haunt your dreams:

I have all eyes and I must scream: https://i.imgur.com/90BJAff.png

A russian nesting doll of horror: https://i.imgur.com/RQovySF.png

I SEE YOU!!!! https://i.imgur.com/72j4OdH.png

Which AI generator are you using? It looks like DreamStudio... I personally use Midjourney because I simply like the way it works better.

In reply to praguepride (msg # 17):

Oh dear XD

To be fair though, the portrait generator seems to have done a good job capturing who Renee is on the 'inside' :D

IE: A strange and creepy weirdo.

Another request for ya:



Positive Prompt: (full body) portrait of a (barbarian woman), finaly detailed peatures, epic fantasy art, (muscular), wielding a giant sword, (trending on artstation), masterpiece, Xena Warrior Princess, at night, full moon, stars in the sky

Negative Prompt: (((deformed hands))), (((fused fingers))), (((Poorly drawn face))), (((cross-eyed))), (((blurry))), (long neck), (((deformed))), bad anatomy, bad proportions, ((malformed face)), (malformed body), extra faces, extra fingers, extra arms, extra legs, amateur art, missing arms, missing legs, missing head


So with this one, once again "holding swords" was a problem and in the final one I had to do some minor clean up to get the sword blade attached back on the rest of the sword. Someone asked me about "failures" so here are a couple amusing ones:

Don't step on it! https://i.imgur.com/PmwRHSu.png

Blades! Blades everywhere! https://i.imgur.com/zCbJZ7Q.png

Who's hand is that? https://i.imgur.com/I6gZETZ.png

I used that last one, dropped a big sword on top of it to cover up the phantom hand and cleaned up that weird blade thing on her right leg and voila!

Final Copy: https://i.imgur.com/ySGTMw4.jpg

I'm running a local copy of Stable Diffusion so I get to tweak the settings. I really like Midjourney's output. My guess is they stripped out a lot of low quality art and images but it's expensive and I like having it run locally in the background. I can have it generate hundreds of pictures overnight and just sift through them in the morning for the keepers.

Got another one hot on the press:



I turned that into this prompt:

Positive: ((full body)) portrait of a male fantasy warrior,  brown_hair, brown_cloak, metal_armor, (wielding a sword),  (André the Giant WWE), (Big Show WWE), (orc), (trending on artstation), high resolution, masterpiece

Negative: (((deformed hands))), (((fused fingers))), (((cross-eyed))), (((blurry))), (long neck), (((deformed))), bad anatomy, bad proportions, (malformed body), extra faces, extra fingers, extra arms, extra legs, amateur art, missing arms, missing legs, missing head


Final result after once again some minor photoshop to "fix" the sword/hand issue: https://i.imgur.com/UpqUHin.jpg


edit: For those paying attention you'll note I removed all the "ugly" and "deformed face" negatives to push him to a more brutal and orcish look as requested.

---

Fun fails follow:

I SUMMON...MYSELF! https://i.imgur.com/VD5r2un.png

Boomerang sword! But does it always come back? https://i.imgur.com/QXgw8Pm.png

Sometimes the AI gets just...so close to perfection. If only that sword was actually in his hand it would be an amazing composition: https://i.imgur.com/YMHco3F.png
This message was last edited by the user at 18:58, Tue 18 Oct 2022.

Alright another one:



So I discovered when doing this one that if you put any color in (like silver hair) no matter what the AI is now thinking "hmmm...silver" and will put it everywhere creating things like this:

You Had ONE Job, AI! Silver Hair: https://i.imgur.com/WkLa7ng.png

But if I put colors next to EVERYTHING (silver hair, black cloak, brown leather) then it helps "re-correct" the AI

So anyway here's the prompt I picked from that description:

Positive: ((full body)) portrait of a male fantasy warrior, (André the Giant WWE), (Big Show WWE), brown_hair, brown_cloak, metal_armor, (wielding a sword), orc, epic artwork, high resolution, masterpiece, (trending on artstation)


Negative: ((deformed hands))), (((fused fingers))), (((cross-eyed))), (((blurry))), (long neck), (((deformed))), bad anatomy, bad proportions, (malformed body), extra faces, extra fingers, extra arms, extra legs, amateur art, missing arms, missing legs, missing head

And here are some cool images that popped out the other end:

Close up: https://i.imgur.com/9p9nH7D.png

Cloaked: https://i.imgur.com/Lz1XwhV.png

Good Job, AI: https://i.imgur.com/obO4esr.png

These are absolutely fantastic - even the fails are entertaining fails. Well done!

...this is where I start hoping for enough free time in the next few days to get in on the rMail train for shiny AI artwork. Keep up the good work!

Ditto.  Following along eagerly.

In reply to praguepride (msg # 8):

I like them all, but the final one is the best in my opinion, the green glowing armor is a distractor though.

Could I pester you to try this one too please?

Good Job, AI: https://i.imgur.com/obO4esr.png  is a good start

Character is a Fencer from roughly the three Musketeers era- but no Surtout.
20 year old Male with long auburn/brown hair tied back
Green eyes, tanned skin although natural pink/white elsewhere.
Scar over right eye & cheek [Saber cut]
Amidextrous.
Wears a rapier (with cutting edge) and a Main Gauche (Large Knife with large protective grip)
Leather helmet, light scale on Torso, bronze Armbands just above wrists, Blousy sleeves. woolen breeches and knee high soft boots.
Wide leather belt with small pouch, scabbards for rapier and main gauche.

Perhaps more detail than you need, as a waist or mid-chest up rendering would be all I want.
This message was last edited by the user at 19:31, Wed 19 Oct 2022.

So the last couple of  prompts I tried  worked out horribly. That's what I get for experimenting. I do have this one though from that previous batch that might be a fit for ya:

https://i.imgur.com/bYnRu9y.jpg

So just as I'm playing with this more it is important to realize that the AI struggles with accessories. Things like swords or armbands or belts are tricky because while there are plenty of artwork where it is positioned correctly, that isn't always tagged as such (if there is a rogue in epic battle against a red dragon's fire breath you probably aren't going to tag every bit of clothing he's wearing). Instead a lot of that imagery comes from "fashion shots" pulled from ebay where it is people selling or showcasing items without anyone wearing them. When you search through the reference document a result of "sword" returns tons of pictures of just swords lying on tables which is why the AI has such a hard time putting  them where they need to go. It knows what a sword is but it just struggles to figure out "oh...they're meant to be in a hand because 90% of the sword images I have are just them lying around..."



(Full body:1.5) drawing of a male warrior, (handsome face:1.3), (detailed pupils:1.5), (brown hair), hair tied back, (green eyes:1.3), brown leather armor,  (trending on artstation:1.3), masterpiece, high resolution, (realistic skin texture), (early modern period), rapier, (scale armor), cloak, (sword belt)

(deformed hands:2), (Poorly drawn face:2), (cross-eyed:1.5), (blurry:2), (ugly:1.2)), (human centipede:1.2), (Long neck:1.2), (deformations:1.2), (close up:1.5), (fused fingers:1.5), (headless:1.5), missing hands, missing arms, missing legs, missing face

---

I learned that I can just attach weights to things so I don't have to use tons of paranthesis and it allows me to more easily tweak the settings. On the upside I'm getting pretty consistent images but the downside is by so tightly controlling things I am getting very much what feels like the same images. I feel like I'm losing the creative spark within the AI. Anyhoo here is a bunch of solid renders. Sadly CLIP didn't have a lot of good results for things like "muskateers" or "early modern period" so it's kinda just more dudes in armor but they do look really good so see if any of these work for ya:

So close to a proper grip: https://i.imgur.com/adlHLNX.jpg

Amazing detail on the armor on this one: https://i.imgur.com/OVd8u6r.jpg

Just pretend he's holding a sword XD: https://i.imgur.com/BgLdnWX.jpg

Probably the best "grip" I've seen it naturally generate for things like this: https://i.imgur.com/ne7puKf.jpg

Now that is a "game face": https://i.imgur.com/d7CECuq.jpg

If Dr. Strange wore battle armor: https://i.imgur.com/kKgPeo3.jpg

So, what is the end-game here? Are you hoping to make money off the service? Or will you do it gratis on this site going forward, first come-first serve, as time allows? Or is there a third option I'm not seeing?

I sent in my description and loved the fast response/first mock up. But I wouldn't actually use it as a game avatar unless there was the chance for follow up and further edits.

To be clear, I'm not expecting that- wouldn't ask for it- and don't want to waste your time or seem ungrateful. Just curious.

I'm just doing this to practice making stuff in a semi-useful way. I would never look to charge for something like this. I have seen some artists looking at offering services to "clean up" your AI generated art but that's not me.

I'll do this as long as I'm interested and people are interested.  The end game is just it's fun to play around with and I just wanted to share it around.


And yeah with your request sadly without a strong "muskateer" look there isn't much I can do :D

Well thank You for your efforts, I can  touch up one of them enough to use  I am sure some of these good ones could be used on RPOL for characters. I assume they would be happy to add them too.

I hope this goes without saying but let me be clear I waive all copyright and ownership of any images I post here. Feel free to use & abuse them as you see fit.

Another request:



This was an interesting one. Had to tinker with it to find the right blend of "orc" and "baboon"

Are you horny?: https://i.imgur.com/GHdVotk.png

These ones look more like just primitive baboons: https://i.imgur.com/0RTvQw1.png

I think this is the best I can get for this prompt for  now: https://i.imgur.com/ZwcgeTv.png

https://twitter.com/Fahad4014/...DoFu5ZahEpQ66MIMPwYA

In case your local media didn't run this story, this is a magnified image of an ant's face. Thought this might be a fun thread to share it on.

ETA: If the link doesn't work or you'd prefer not to click on it, you can research the photographer instead- his name is Eugenijus Kavaliauskas.
This message was last edited by the user at 07:17, Sat 22 Oct 2022.

Praguepride, it has been rules that AI generated art can't be copyrighted.
Also, be aware many artists see this as a way to get around their intellectual property rights, and is being seen by many as a way to push artists out of the equation.

https://kotaku.com/ai-art-dall...copyright-1849388060

I highly doubt any of this information is going to stop anybody, but at least try to see where artists are coming from.

So the actual training model isn't restricted against copyright but the AI just naturally has problems exactly duplicating something that fits a rigid pattern. For example it can make star trek LIKE things but you can't really tell it to make a perfect star trek insignia unless you spend the time to train it on hundreds or thousands of images to cement that look.

Many public facing models like Crayon, DALLE, Midjourney put on filters that weed out copyright or other illegal/NSFW but it's not that the model itself 'can't' do it, but you put checks in place to kick out things. Violence is a big one because nobody wants the publicity of having their AI model publish violent material either.

As for the artists, they're kind of SOL. The copyright laws are still being developed but similar issues involving training AI datasets have gone heavily towards the researchers. While the artists didn't give permission to have their work incorporated, training an AI model isn't directly competing with the artists so it typically falls under "Fair Use".

In addition if I go to a gallery and see an artist's work and then go home and draw a picture in that same style I'm not breaking the law unless I am directly lifting stuff from their work.

AI image generation does not copy work, but is inspired by it. I do agree that there are going to be tons of lawsuits and case law that has to be published and current trends could break against AI researchers providing avenues for artists to get their works out of training sets but honestly, the cat's out of the bag. Even if they stop it tomorrow there are so many open source models floating around that are already trained on these same pieces of art that it's kind of too late.

It's equivalent to the RIAA or MPAA trying to crack down on the free share of music or films with the addition that the law is fuzzier because people are trading around not exact copies but music/films "kind of like" existing material.

It's interesting to watch and as I work with AI/ML professionally I have a keen interest to stay on top of the case law.

tl;dr: So far in the past law suits have not gone in favor of people suing AI researchers so I doubt the artist community will be able to stop it either. Progress marches on etc. etc.

I just wanted people to be aware of the issues if they are going to use it.

Totally fair. In respect to the artists I specifically do not put artists "names" into my prompts even if it is legal doesn't mean it is moral.

How convoluted can the description be? What can be done with something like:

Character is a tall, well-built, hybrid Vulcan Orc, trained to play a mandolin and sing to make a living. Training further to be an assassin, he incorporates a lightsaber into his mandolin which he uses to execute his targets. Dressing as a courtier, he frequents the courts of the nobles and plies his trade in the darkness; the only light upon his kills is the glowing blade springing from his mandolin.

You can put that into a prompt and it might work out well but the  AI doesn't have a strong association of how to put the "pieces" together. So if you put "trained to play the mandolin" it will likely just key onto mandolin so you don't really need the other parts. That being said the NLP (natural language processor) is very powerful so if you put "trained to  play the mandolin" it might do a picture of someone practicing the mandolin.

Just for fun I put that entire phrase into a quickie generation and this is what came out:

https://labs.openai.com/s/BGhGPIjJz03VUJSUJY4Ymm6D

So you get:

- Vulcan (pointy ears)

- Orc

- Playing  the mandolin

And  that's about it. Maybe the other stuff had an impact but at least  from our perspective most of the other stuff fell off.

https://labs.openai.com/s/TNxNwlQBjAlSyd1TVpQ6n9EF

That is another result and you can see it added "dressed as a courtier" and  "blade" even  though it skipped lightsaber.

This is why some engines use weighting so you can have more control over how you pull it in one direction or another.

It makes sense (to me, at least) that a lot of that information would be dropped by an AI...it's describing a behavior, not an appearance (everything after 'dressing as a courtier', pretty much, and even that is questionable).  If you were describing a specific scene, perhaps, but this particular AI is generating portrait work, not capturing the characters in media res, so the words preceding the description of what he does, where he does it, and how he does it would likely be interpreted by the AI (if it's got any sort of semantic filtering) as not associated with the actual appearance of the character and therefore extraneous information that can be ignored.

If it was rephrased, perhaps, it would be more easily incorporated, but I suspect that 'dressing like a courtier' is rather vague for an AI...a courtier from where?  And when?  Of what rank?  In whose court?  All of these could massively impact what 'dressing like a courtier' means, so without specifics to clarify, the AI doesn't really devote much 'effort' to realizing that part of the description.

In reply to praguepride (msg # 33):

Thanks! That last one actually looks pretty good.

I took a quick break from doing requests to take care of some personal projects of my own and figured I would share them. I have a sci-fi crew and I generated images for all of them.

Ex-military captain: https://i.imgur.com/3d6VWo4.jpg

Femme Fatale XO: https://i.imgur.com/HoMplee.jpg

Grizzled "I'm too old for this" engineer: https://i.imgur.com/MnIdYuo.png

Tough no-nonsense orbital drop trooper: https://i.imgur.com/luW7NzD.png

Spoiled rich young VIP: https://i.imgur.com/6pbWKVr.png

When you're back to taking requests, how about this one?  You might recognize her.  :-)

extremely attractive gnome female
 sorceress
 short black hair
 skin the color of coffee with cream
 green eyes
 red dress
 carrying a wolf puppy
 night sky background

Feel free to post it here, if you like.
This message was last edited by the user at 21:24, Tue 25 Oct 2022.

In reply to praguepride (msg # 43):

Those look really cool! Well done

So I'm still baking and will look at the results but from the very first batch came this which I thought looked awesome:

https://i.imgur.com/Lt39zqk.png

Don't know if it necessarily fits but I'm happy it got the wolf pup in there without being an absolute mangle.

Awesome!  Now you're going to have to let Nezaglomie keep the wolf pup to train and wear that cool-looking armor.

Alright this was a good batch, lots of good ones to pick from (including the above one)

Holding the white wolf: https://i.imgur.com/ldaB0R8.png

Did I say Green Eyes? I meant green HIGH-lights: https://i.imgur.com/xQkXCNQ.png

She's giving you the evil eye: https://i.imgur.com/rs96U0T.png

Who's a good boy! You're a good boy: https://i.imgur.com/DZG2Mst.png

Golden is the new green: https://i.imgur.com/cVrvr1i.png

It looks good but I just imagine her getting up an hour before everyone else to carefully do her eye makeup before the rest of the adventuring party wakes up: https://i.imgur.com/HcH1Qtr.png

Fun Fails:
Wolves are hard so I just gave your girl some  dog ears: https://i.imgur.com/YzsW8Mc.png

Wow!  Thanks!  I definitely like this one the best:  Golden is the new green: https://i.imgur.com/cVrvr1i.png

Most of the others don't look enough gnome-like.  She's supposed to be only 2 ft 2, so no long neck and even a wolf puppy should seem pretty big in comparison.

Because of how over-indexed WoW is, any time I mentioned "gnome" the thing got weird. If I'm up for it I might take another crack at it by re-jiggering some things around.

Another batch cooked up last night:



One of the things I've found useful is to find a famous celebrity that kinda meets the look you're going for and dump it at the very end with minimum weights to kind of "pull" the picture together. The AI can struggle if you put the celeb too high because most of the pictures it has for celebs are "red carpet" pics that can result in some bizarre (and hilarious) compositions but if you put it at the end so the AI thinks "like this guy...but subtle" it can be useful.

I tried it with a few different celebs:

Clive Owen: https://i.imgur.com/x2R7QWC.png

Jeremy Runner: https://i.imgur.com/gYq9TzN.png

The requestor suggested Daniel Craig and I think that was the right call:

Some final version:
https://i.imgur.com/GuxZvEp.png

https://i.imgur.com/rdWF8pS.png

https://i.imgur.com/cfOsTUv.png

https://i.imgur.com/bNGVX5F.png

https://i.imgur.com/XO3UffF.png

They all have varying degrees of resemblance but I think are distinct enough that it isn't just "Daniel Craig AS..." maybe more like his his brother or cousin....

Those turned out really good! (Yeah, it's my character...)

I may have to add one of these to the character description in the game...I'm still not completely sold on Daniel Craig as an archetype, but it's going to be hard to find someone else that beats some of those!

What is the name of the App that creates these? Which is the best.

There's a ton of information (maybe too much!) in these links, both on different apps and tutorials:

https://www.reddit.com/r/StableDiffusion/wiki/index/
https://www.reddit.com/r/Stabl...sion/wiki/tutorials/

Hoooooo boy. "Which is the best" is just asking for trouble.


Right now there are 3 Big Boys in the paid "AI Art As A Service" world: Dreambooth, Midjourney, and NovelAI.

NovelAI is known mainly for doing anime styles. Midjourney is primarily known for it's more evocative artistic flair although with the right prompts it can do various styles and photorealistic pictures. Dreambooth is kind of the "steady workhorse" where it doesn't necessarily specialize in anything but it does everything pretty well and also has a lot of features to help the user feed in an image (or images) so it can, say, put you into the picture.

Then if you are technically savvy and have a beefy enough computer you can run Stable Diffusion (the main model behind it all) on your local machine or use a Google Colab account to basically borrow free CPU cycles to do your own images but with the INFINITE SETTINGS you need to tweak to do it yourself (the paid services typically take a lot of the settings out of your hands so you just have to focus on the 'prompt').

There are others: DALLE-2 is pretty cool, Craiyon is a freebie one that is fun to play with and more and more are popping up every day but DB, MJ and NA are the three big boys for paid services.

As a warning paid AI doesn't look too expensive at  first ($15 gets you like 200 images through MJ for example) however it is very addictive and you can burn through a LOT of those credits as you're tweaking and refining prompts. Dreambooth gives you like 50 credits for free as a demo, MJ gives you 25 for free as a trial and I burned through both of those in a couple of hours.

Thank you for the excellent review and guidance!

 Just as a heads up for anyone with a significant visual issue, and this is just so far, but the big three previously mentioned don't work with a screen-reader.   This probably isn't a huge surprise, but in order for you to not beat your head against the wall, and instead let me do it for you, this just FYI.

For you fully sighted folks, disregard.

Me to.

So it's been awhile and I'd like to say that both my skills and the tools available have improved dramatically. Looking back at my original works here even just a few months ago the difference is night and day.

https://i.imgur.com/nwaaV8z.png

https://i.imgur.com/q0BIyqT.png

https://i.imgur.com/y5hk5yA.png