Options
Go to last post Go to first unread
dm  
#1 Posted : Tuesday, December 27, 2022 4:29:54 PM(UTC)
dm

Rank: Member

Joined: 8/27/2018(UTC)
Posts: 19

Thanks: 69 times
Was thanked: 6 time(s) in 6 post(s)

I did not find any information about this subject in the community guidelines. I normally work in DAZ, but have recently been dabbling in Stable Diffusion AI rendering. Is it acceptable to post AI rendered images in Render Erotica, if appropriately label as such and adheres to community guidelines? My current AI work is more pinup style, but as Unstable Diffusion (hardcore adult trained) AI becomes more mature, I'm sure folks will find the renderings more entertaining.


thanx


 


davo  
#2 Posted : Tuesday, December 27, 2022 8:27:37 PM(UTC)
davo

Rank: Advanced Member

Joined: 9/15/2012(UTC)
Posts: 369

Was thanked: 281 time(s) in 165 post(s)
As long as no kiddo or beastiality you're good to go, you can post any art medium in the gallery. Also no copyright infringements for example star wars fan art, stuff like that.
D
thanks 2 users thanked davo for this useful post.
dm on 12/27/2022(UTC), Sutut on 12/30/2022(UTC)
CDDVD  
#3 Posted : Wednesday, December 28, 2022 7:51:48 AM(UTC)
CDDVD

Rank: Member

Joined: 8/21/2022(UTC)
Posts: 14

Thanks: 3 times
Was thanked: 5 time(s) in 4 post(s)
Would you mind giving us a brief overview of the program? Does it meet your expectations, etc?
I'm not interested in making images from scratch in it, but wonder if it would work to modify
existing renders.
Thanks!
RangerRabbit  
#4 Posted : Friday, December 30, 2022 1:54:50 PM(UTC)
RangerRabbit

Rank: Administration

Joined: 12/23/2019(UTC)
Posts: 204

Thanks: 54 times
Was thanked: 240 time(s) in 112 post(s)

I've created an entire AI gallery section to add AI images to. Please use these categories if you are creating using AI and do not cross over into the Rendered images categories. Thanks.  Have fun.


RangerRabbit
Renderotica Management
thanks 4 users thanked RangerRabbit for this useful post.
ContentModerator on 12/30/2022(UTC), Zaavaleta on 12/30/2022(UTC), dm on 12/30/2022(UTC), Thunder-3D on 12/30/2022(UTC)
Zaavaleta  
#5 Posted : Friday, December 30, 2022 3:09:51 PM(UTC)
Zaavaleta

Rank: Advanced Member

Joined: 4/21/2016(UTC)
Posts: 318

Thanks: 50 times
Was thanked: 88 time(s) in 63 post(s)

Originally Posted by: CDDVD Go to Quoted Post
Would you mind giving us a brief overview of the program? Does it meet your expectations, etc?
I'm not interested in making images from scratch in it, but wonder if it would work to modify
existing renders.
Thanks!


I'm no expert at it, but I'll try. Stable Diffusion is a program you can install on your computer that generates AI images from a text prompt you type in. There's Inpainting - selecting an area of an image and having AI add content inside it. There's Outpainting - expand the borders of an image and let AI create whatever it wants. SD Upscale - upscale an image and let AI generate whatever details it can't make out. Training - you can train SD with images of a person, object, place, or whatever and create images from it. There are a ton of other features I haven't even tried yet.


If you're working locally, you're limited by your card's VRAM. I have 12 Gb and it's good for 1280 x 720 renders, but above that I start running out of memory. There are online companies that offer SD instances running on Nvidia workstation cards(for a subscription), so they can offer 24, 48 Gb or more.


I'm not that good at inpainting, but here's my attempt.


Zaavaleta attached the following image(s):
1- trailer render.png
2 - sd output.png
thanks 1 user thanked Zaavaleta for this useful post.
CDDVD on 2/16/2023(UTC)
Zaavaleta  
#6 Posted : Friday, December 30, 2022 3:23:12 PM(UTC)
Zaavaleta

Rank: Advanced Member

Joined: 4/21/2016(UTC)
Posts: 318

Thanks: 50 times
Was thanked: 88 time(s) in 63 post(s)

Originally Posted by: RangerRabbit Go to Quoted Post


I've created an entire AI gallery section to add AI images to. Please use these categories if you are creating using AI and do not cross over into the Rendered images categories. Thanks.  Have fun.



Thanks for making that. Can you or the mod team clarify how much of an image needs to be AI to make it an AI image. For example, if I just reshape some trees or add mountains in the background, does that qualify? I mean technically you can use SD to create textures for 3d objects, so AI images could show up anywhere. Thanks.


Sutut  
#7 Posted : Friday, December 30, 2022 3:51:17 PM(UTC)
Sutut

Rank: Advanced Member

Joined: 9/15/2012(UTC)
Posts: 591

Thanks: 329 times
Was thanked: 100 time(s) in 88 post(s)

You could even draw stuff with a pencil and pen despite the site's 3D generated theme - of course keeping TOS as Davo reminded above!



Frankly 3DCG was the "Aiart" of the past, animators worrying it'd "Take their Jeerb!" and people having issues with "Is it art..." so we should be supportive.


 


I'll argue -and look at my gallery - that it's really neat to play with and we who use 3DCG can work with it inputting images, our old ones or new ones to make better results.  Furthermore the utilities can be downloaded, the AI trained - I'm looking into this option myself downloading datasets and playing with some AI - our piles of 3DCG images we made can be input to train our personal AIs.  The EULA allows any use of 2D images generated


 


Note I saw an issue with should this site have an AIart opt out?  Well I doubt Renderotica plans to or is capable of an AIart thing like*deleted*.  On the note those initial Stable Diffusion learning engines scrubbed the PUBLIC internet as any human illustrator does to look at tons of pictures to learn.  Somehow its 'stealing' when a machine does it even though a human does the same thing?  In any case Renderotica is behind a password wall and the AI was told to NOT bypass those for legal reasons so they likely did not scrub our work here.  Too bad, they'd get HANDS better perhaps...  One of my major issues is to get hands working -but painter + 3DCG hand + clone brush helps when the hand is too mangled to work or missing in a Midjourney painting....


Edited by moderator Saturday, December 31, 2022 11:49:45 AM(UTC)  | Reason: Not specified

dm  
#8 Posted : Friday, December 30, 2022 4:02:40 PM(UTC)
dm

Rank: Member

Joined: 8/27/2018(UTC)
Posts: 19

Thanks: 69 times
Was thanked: 6 time(s) in 6 post(s)
Thanks for creating a separate category for AI rendered images, which is why I sent a question form email requesting clarification many days before I posted the first image, but sadly did not get a response. I believe it only fair to have this catagory. I hope others explore the AI rendering possibilities, but for now it is very limited in scope. It would be great to share experiences and information with other AI rendering enthusiasts.
wurger  
#9 Posted : Friday, December 30, 2022 9:11:54 PM(UTC)
wurger

Rank: Advanced Member

Joined: 5/29/2015(UTC)
Posts: 202

Thanks: 26 times
Was thanked: 77 time(s) in 62 post(s)

Originally Posted by: Sutut Go to Quoted Post


You could even draw stuff with a pencil and pen despite the site's 3D generated theme - of course keeping TOS as Davo reminded above!



Frankly 3DCG was the "Aiart" of the past, animators worrying it'd "Take their Jeerb!" and people having issues with "Is it art..." so we should be supportive.


 


I'll argue -and look at my gallery - that it's really neat to play with and we who use 3DCG can work with it inputting images, our old ones or new ones to make better results.  Furthermore the utilities can be downloaded, the AI trained - I'm looking into this option myself downloading datasets and playing with some AI - our piles of 3DCG images we made can be input to train our personal AIs.  The EULA allows any use of 2D images generated


 


Note I saw an issue with should this site have an AIart opt out?  Well I doubt Renderotica plans to or is capable of an AIart thing like *deleted*.  On the note those initial Stable Diffusion learning engines scrubbed the PUBLIC internet as any human illustrator does to look at tons of pictures to learn.  Somehow its 'stealing' when a machine does it even though a human does the same thing?  In any case Renderotica is behind a password wall and the AI was told to NOT bypass those for legal reasons so they likely did not scrub our work here.  Too bad, they'd get HANDS better perhaps...  One of my major issues is to get hands working -but painter + 3DCG hand + clone brush helps when the hand is too mangled to work or missing in a Midjourney painting....




There always seems to be some excuse from the pro-AI crowd to justify their fun. AI "free speech" arguments fail the instant you add "in the style of <insert artist name here>" in your text string.

Edited by moderator Saturday, December 31, 2022 11:50:28 AM(UTC)  | Reason: Not specified

Sutut  
#10 Posted : Saturday, December 31, 2022 11:35:16 AM(UTC)
Sutut

Rank: Advanced Member

Joined: 9/15/2012(UTC)
Posts: 591

Thanks: 329 times
Was thanked: 100 time(s) in 88 post(s)

Originally Posted by: wurger Go to Quoted Post


 


There always seems to be some excuse from the pro-AI crowd to justify their fun. AI "free speech" arguments fail the instant you add "in the style of <insert artist name here>" in your text string.


 



That's a lie being told by "Artists", many who can barely draw a stick figure, to repeat enough to try to make it the truth.


If the AIart programs did that they'd need to have the NSA's data center and computers for the task.  And they'd be blatantly cloning and get caught overnight.  Note real HUMAN artists steal all the time - and get caught very quickly since we have good pattern recognition.  Nick Simmons is a good famous example.  For AI it only clones that close when forced to, often by the "Artists" complaining about it who are REALLY mad that it spat their own stuff when they fed it in to 'prove' stealing by machine out but the machine spat it out better...


It's fascinating and terrifying but legit the machines were able to learn how to do art.  Yes, off existing art - so WHAT - humans do the exact same thing;  Create a database of stuff in their heads, expression from Van Gogh or Monet, realistic painting from everywhere, a shoulder from a photo for shading...  Then they combine the millions of actions to make a new painting, doing their best to not "Nick Simmons" make so close its almost tracing which can be done by a human by accident.  Picasso blatantly copied other artists, he'd visit them for a party, eat all the cheese, sleep with their wives or other lovers, see a  lonely painting with some paint and sketches  on canvas out of the corner of his eye -then 1 day later he'd have done BETTER than they could in their lifetime what they were trying to paint.  That's what we have in computers.  It's called "Machine Learning" as a better term than "Artificial Intelligence" but you can look it up, it is legit.


Note also a human's "Style" is NOT protected by Copyright. Roger Dean v James Cameron for instance.  That's good or Disney would own it all instantly.  Mark Bode would be ruined or forced Mike Diana style NOT to draw anything by Disney because the "Vulgar Underground" he and his late Father made interfered with the "Wizards" franchise they own now.  Legally you can imitate another's style or pay another to do so - there's neat documentaries on YouTube:  China Van Gogh - and look near the end of the long ones, they imitate anybody and save 1:1 existing work tracing or copyright characters its LEGAL proved again and again in court.  Any battles over AI will just likely weaken this, hurting artists - or if they win somehow (style is copyright protected?) then DISNEY will own it all instantly and make it harder.


I suggest any "Artist" in it as a 'career choice' get . out . NOW.  Those who do art (or other creative stuff) for its own sake just embrace it.  Thank God that it is software you can access easily for free or a small fee - and can download and run on your own PC or rented cheaply virtual machine space.



Trying to be nice, but ready to DRINK their TEARS and I'll elaborate if you wish.


Hellboy  
#11 Posted : Saturday, December 31, 2022 12:04:03 PM(UTC)
Hellboy

Rank: Advanced Member

Joined: 9/15/2012(UTC)
Posts: 2,158

Thanks: 55 times
Was thanked: 109 time(s) in 77 post(s)

Originally Posted by: wurger Go to Quoted Post


AI "free speech" arguments fail the instant you add "in the style of <insert artist name here>" in your text string.



People have always tried to reference, inspire from or just copy other artists' styles. A style can't be copyrighted.


UserPostedImage
DAZ STUDIO USER SINCE 2006
thanks 1 user thanked Hellboy for this useful post.
Sutut on 12/31/2022(UTC)
RangerRabbit  
#12 Posted : Saturday, December 31, 2022 12:08:40 PM(UTC)
RangerRabbit

Rank: Administration

Joined: 12/23/2019(UTC)
Posts: 204

Thanks: 54 times
Was thanked: 240 time(s) in 112 post(s)

Originally Posted by: Zaavaleta Go to Quoted Post


Originally Posted by: RangerRabbit Go to Quoted Post


I've created an entire AI gallery section to add AI images to. Please use these categories if you are creating using AI and do not cross over into the Rendered images categories. Thanks.  Have fun.



Thanks for making that. Can you or the mod team clarify how much of an image needs to be AI to make it an AI image. For example, if I just reshape some trees or add mountains in the background, does that qualify? I mean technically you can use SD to create textures for 3d objects, so AI images could show up anywhere. Thanks.



I think you need to use your best judgement here. If the main focus of the image is AI generated, then use the AI categories, if you are using AI to generate textures or shading or background, then I think it would be fine in the Rendered User categories. Ask yourself the question, are you using AI like a post production editing tool? Or are you using AI to do the bulk of the creating? 


RangerRabbit
Renderotica Management
thanks 1 user thanked RangerRabbit for this useful post.
Zaavaleta on 1/1/2023(UTC)
davo  
#13 Posted : Sunday, January 1, 2023 2:01:28 PM(UTC)
davo

Rank: Advanced Member

Joined: 9/15/2012(UTC)
Posts: 369

Was thanked: 281 time(s) in 165 post(s)
Art is art, I'd hate to think that people actually filter out artwork based on whether it was hand drawn, sketched, rendered or AI? Do we really need to know? How the artist composites and uses the tools available is the the part that makes the artist the artist. When people inquire it sparks conversation, which I'm sure the artist loves because feedback is always appreciated. We had an interesting discussion a while back about photo-real versus graphic art styles of art and it was eye opening. Just and observation.
Bumper2  
#14 Posted : Tuesday, February 14, 2023 5:38:58 AM(UTC)
Bumper2

Rank: Advanced Member

Joined: 11/27/2016(UTC)
Posts: 325

Thanks: 1250 times
Was thanked: 183 time(s) in 139 post(s)

I've been thinking about this question for several weeks now and have two minds about it. If you'll allow me, here are my thoughts.


I think the crux of the 'AI' vs 'Anti-AI' debate rests on what people not only qualify as art, but in what is felt to be 'too close' to an original - and also, what qualifies as an artist. AIs are taught by feeding them a ton of images, then you put in several keywords and let them go nuts. In a way, it's akin to a photo filter - let the AI decide what to choose and display. It becomes even more of an immediate relief than browsing for something on the web. Busty girl, naked, desert, green hair, spread eagle, large labia, overhead. You put it in and the results form exactly (or relatively close) to what you want. The database scans millions of files, picks the best choices, and makes them work together. You get the picture you want, and it's usually pretty good.


The 'Anti-AI' crowd is angry because it's both combined previous works of actual people who've taken hours to make various images, and years to hone their craft, and distilled it into a couple minutes. That is valid. I've personally spent several full minute-by minute years of my life trying to make the art I wanted to see and be seen. I've learned lighting, composition, phrasing, movement. Does all my effort count for nothing simple because you can type in a few words and get a picture? I think many artists also rile at the fact that they've spent their own money on various products, whether it be paints and supplies or assets here on 'rotica, and AI artists can get a free pass to all of that. We spend a section of our lives in both training and materials. That's time we'll never get back. It IS a bitter pill to swallow that with a few button presses others can get the same results, and post the amalgamation of it as 'their art'.


However...


When I saw 'The Black Lace Cafe' originally, within an hour of you posting it up, I immediately liked it and added it to my favorites. It's quite simply stunning. AI art can grab things from everywhere to create an exceptional and polished image.Many of your pinups, such as 'Tatoo Punk' are gorgeous. It IS art, though not what many deem it to be. There's a video from MovieBob here that does an amazing job of breaking down the feelings of both sides, albeit from a video game perspective. As the Roger Ebert quote he uses goes: "Anything can be Art. Even a can of Campbell's soup. What I should have said is that games (AI drawings in this case) cannot be high art, as I understand it." Any new medium takes time to be accepted and right now AI art is in it's baby steps. The legal issues of AI art are incredibly complicated. LegalEagle does an excellent job of breaking them down here, as well as what's to come in the future. 


For me, I generally have no problem with it. My basic thought is that as long as there's no copywrite infringement and no use of my intellectual property (i.e. Nadine, Trixie, Tusk, Alyssa, et all) I'm ok with it. Chrissy is a great 'girl next door', but that trope has been used for a VERY long time. What differentiates her is the personality I've included and her exact look. Someone remaking her in their own image would fall afoul of me. However, someone typing in a prompt of 'girl next door, blonde, young, lesbian, bondage' wouldn't, even if it was similar to my style and look, simply because it's not the same character.


This in no way will sway the opinions of many, and is probably too much for a fun site like 'rotica, but hopefully it will give a few varied opinions that helps people see the depth of the issue.


p.s. Note to the mods, the links are youtube videos. I intended them as clarification of my thoughts only, and as a research source.


CDDVD  
#15 Posted : Thursday, February 16, 2023 5:53:55 AM(UTC)
CDDVD

Rank: Member

Joined: 8/21/2022(UTC)
Posts: 14

Thanks: 3 times
Was thanked: 5 time(s) in 4 post(s)
Thanks Zaavaleta,
I've got the thing running on my dedicated graphics computer now. My intention is to use it for some of the 'grunt work' presently done in Photoshop. Cutting out hair is a pain despite PS's advanced tools. I'm also training it on my massive, hard drive choking collection of lingerie, especially the vintage stuff. I'd like to see how it handles dressing nudes in lingerie. I played a bit with fixing ripped and torn scans from vintage mail order catalogues and it worked perfectly.
Thanks for the introduction!
Zaavaleta  
#16 Posted : Saturday, February 18, 2023 8:28:31 AM(UTC)
Zaavaleta

Rank: Advanced Member

Joined: 4/21/2016(UTC)
Posts: 318

Thanks: 50 times
Was thanked: 88 time(s) in 63 post(s)

@CDDVD, hey, thanks for the post. It inspired me to try something I really hadn't done before. Lately I've been learning how to train with textual inversion. I finally got a few decent images of *hyperlink deleted* and *hyperlink deleted*


So far I haven't gotten Dreambooth working, but maybe soon. I can't seem to convince textual inversion that girls can have penises.


Hopefully this doesn't violate the site's terms of service, training on actual people. I figure "They're porn stars, they do porn. This increases their brand." If it was Mr. Rogers and Betty White, then maybe I could see a problem. Maybe.


 

Edited by moderator Saturday, February 18, 2023 12:05:21 PM(UTC)  | Reason: Community Policy Violation - Links and/or Site References

Zaavaleta attached the following image(s):
06982-1804941110-photo of (daisytaylor15_0.9) with short hair wearing starfleet uniform, realistic, solo.png
06497-2764735214-a photo of female starfleet officer in uniform, realistic, (daisytaylor_0.8).png
03942-612186388-photo of young voluptuous woman wearing starfleet uniform, (dillionharper_2.1_0.8), sci-fi, realistic.png
06772-1656541671-photo of young woman (naked_1.3) in restaurant, (dillionharper_2.1_0.8), sci-fi, realistic.png
Users browsing this topic
Guest
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.

Powered by YAF.NET | YAF.NET © 2003-2024, Yet Another Forum.NET
This page was generated in 1.001 seconds.

Notification

Icon
Error