AI Art: It is criminal to not post your prompt

The Something Awful Forums > Main > General Bullshit > AI Art: It is criminal to not post your prompt

Aertuun: Dec 18, 2012

ArtistCeleste posted:

Thanks for trying. They do look a bit, imperial? I think. I might be better off with photoshop.

I'll try Stable Diffusion. I'm messing with Canva too, but it's a pretty simple AI. I don't think you can feed it images.

Use Stable Diffusion with the Controlnet extension. With the canny or hue models it'll do exactly what you're after (keep the anvil profile / style while altering the overall style of everything else)

# ? Mar 20, 2023 08:34

Adbot: ADBOT LOVES YOU

# ? May 31, 2024 13:43

Ruffian Price: Sep 17, 2016

Tom Clancy is Dead posted:

Has anyone gotten Latent Couple to work for them in Automatic1111? I have done the whole mask and multiprompt thing but doesn't seem to be doing anything.

with these instructions

quote:

Use this in the Extensions Folder Command Window:
git clone -b feature/mask_selection https://github.com/ashen-sensored/stable-diffusion-webui-two-shot.git

Use this in the A1111 Root Folder Command Window:
git apply --ignore-whitespace extensions/stable-diffusion-webui-two-shot/0001-Adding-after_ui_callback-for-scripts.patch

I've got to a stage where the masks are applied, but it's more like inpainting part of a full picture. So if I put a red mask at the bottom of the image and assign a full body shot to it, the generation will only have legs :v:

I'm sure that's not how it's supposed to work but I didn't mess with it further, stacking controlnet is more than enough for now

# ? Mar 20, 2023 09:18

mobby_6kl: Aug 9, 2009; by Fluffdaddy

unzin posted:

Here's the algorithm you're probably talking about https://www.deryaakkaynak.com/sea-thru

This git has the actual report and a python script that implements it. https://github.com/hainh/sea-thru

Maybe someone will implement it in stable diffusion.

Awesome, thanks!

Seems pretty dead and apparently it's not a neural-net model so I doubt it'd get included in SD, but I'll try it out later.

# ? Mar 20, 2023 10:17

busalover: Sep 12, 2020

A redditor figured out that you can use ChatGPT as a tool for interactive fiction, pretty cool.
https://teddit.net/r/Futurology/comments/11w1ykw/the_matrix_is_here_using_chatgpt_4_to_play_as_an/
They're using Star Trek, but technically it could be used for any subject matter it's knowledgeable about. I'll try the Bible, could be a cool educational tool. If it works as promised.

# ? Mar 20, 2023 11:21

Roman: Aug 8, 2002

Roman posted:

Then tried "christian bale as the world's most dangerous fast food employee." Looks nothing like him but I got the A24 version of "Waiting"

I spend too much time on this poo poo.
Netflix presents You Never Really Wanted Fries With That.
A riveting modern fable of a fast food restaurant staffed entirely by retired hitmen haunted by their pasts.

Only registered members can see post attachments!

Roman fucked around with this message at 12:16 on Mar 20, 2023

# ? Mar 20, 2023 12:06

Ruffian Price: Sep 17, 2016

busalover posted:

A redditor figured out that you can use ChatGPT as a tool for interactive fiction, pretty cool.
https://teddit.net/r/Futurology/comments/11w1ykw/the_matrix_is_here_using_chatgpt_4_to_play_as_an/
They're using Star Trek, but technically it could be used for any subject matter it's knowledgeable about. I'll try the Bible, could be a cool educational tool. If it works as promised.

This worked pretty well in the non-chat older model, too. First time I was impressed with the long-format memory (especially compared to early AI Dungeon) was going through a CYOA book template, reaching a dead end and realizing the model accurately keeps track of which room description was on which page and puts the right labels on choices

# ? Mar 20, 2023 12:51

pixaal: Jan 8, 2004; All ice cream is now for all beings, no matter how many legs.

At one point you could tell chatGPT to be an actor that never breaks character and have it play as whatever and you could bounce stuff off it pretty well. It was great for exploring a setting I had in my head that's now a bit more fleshed out. Not sure I'll ever actually do anything with it but it was fun to see it come to life just a bit.

# ? Mar 20, 2023 13:18

IShallRiseAgain: Sep 12, 2008; Well ain't that precious?

pixaal posted:

At one point you could tell chatGPT to be an actor that never breaks character and have it play as whatever and you could bounce stuff off it pretty well. It was great for exploring a setting I had in my head that's now a bit more fleshed out. Not sure I'll ever actually do anything with it but it was fun to see it come to life just a bit.

chatGPT can actually do that again. Now that they have improved alignment, it doesn't spew the OpenAI policy every 5 seconds.

# ? Mar 20, 2023 13:49

lunar detritus: May 6, 2009

ArtistCeleste posted:

I'm really failing at this AI generation thing. I'm trying to get it to help me with my logo. I'm trying to take my current logo and replace it with an anvil that is more like the one I use every day.
This is my logo.
https://drive.google.com/file/d/1na_fhy4fg6gZaqrKBmxlWmu6BBjPbrgO/view?usp=sharing

This is my anvil.
https://www.instagram.com/p/B_xdxpXDmxM

It's would probably be called a double horn German pattern anvil. AI evidently thinks that's code for Nazi-like images with birds and various symbols.
Sorry I'm such a newb. Can someone help me?

I doubt this'll help you specifically but it's a fun demonstration of controlnet

# ? Mar 20, 2023 15:36

LASER BEAM DREAM: Nov 3, 2005; Oh, what? So now I suppose you're just going to sit there and pout?

Not sure where else to post this, but I feel like people here may understand. I spent the entire weekend troubleshooting llama in text-generation-webui with the creator of reddit.com/r/locallama and oobagooba. After 4 failed installs and a system format, we finally figured out that you need a specific dev version of cudatoolkit in order to build the 4bit support module in WSL Ubuntu along with a bunch of other little tricks.

After spending the weekend in Ubuntu shell I went to fire up my stable diffusion instance and typed this into powershell

code:

cd /c:\s<tab>

My brain is broken

# ? Mar 20, 2023 16:57

pixaal: Jan 8, 2004; All ice cream is now for all beings, no matter how many legs.

windows will actually accept linux commands you can ls in powershell to piss off anyone watching you work. did you just use ls in windows instead of dir?! it's quicker!

# ? Mar 20, 2023 17:07

LASER BEAM DREAM: Nov 3, 2005; Oh, what? So now I suppose you're just going to sit there and pout?

I actually already stumbled on that! Windows doesn't care for the forward slash to talk about root, though. Makes sense because it doesn't exist!

The path to systems drives when in Ubuntu WSL is

code:

(base) steph@DESKTOP:/$ cd /mnt/c/users/steph
(base) steph@DESKTOP:/mnt/c/users/steph$

# ? Mar 20, 2023 17:19

mightygerm: Jun 29, 2002

Ruffian Price posted:

This worked pretty well in the non-chat older model, too. First time I was impressed with the long-format memory (especially compared to early AI Dungeon) was going through a CYOA book template, reaching a dead end and realizing the model accurately keeps track of which room description was on which page and puts the right labels on choices

I�ve been meaning to test out loom to see how deep it can go with CYOA type stuff:
https://generative.ink/posts/loom-interface-to-the-multiverse/

# ? Mar 20, 2023 18:24

RPATDO_LAMD: Mar 22, 2013; 🐘🪠🍆

busalover posted:

A redditor figured out that you can use ChatGPT as a tool for interactive fiction, pretty cool.
https://teddit.net/r/Futurology/comments/11w1ykw/the_matrix_is_here_using_chatgpt_4_to_play_as_an/
They're using Star Trek, but technically it could be used for any subject matter it's knowledgeable about. I'll try the Bible, could be a cool educational tool. If it works as promised.

this is one of the oldest popular usages for this technology
AI Dungeon launched in May 2019, almost 4 years ago

# ? Mar 20, 2023 20:55

Comfy Fleece Sweater: Apr 2, 2013; You see, but you do not observe.

https://www.federalregister.gov/doc...al-intelligence

II. The Human Authorship Requirement

In the Office's view, it is well-established that copyright can protect only material that is the product of human creativity. Most fundamentally, the term �author,� which is used in both the Constitution and the Copyright Act, excludes non-humans. The Office's registration policies and regulations reflect statutory and judicial guidance on this issue.

Dunno if this was discussed because I'm like 180 pages behind, but this seems interesting, and still evolving.

Basically, no copyright protection on anything you make with AI. Soo... if you're using it for something that might need copyright, beware (videogames, comics, etc come to mind).

They cite an example of a comic book/graphic novel made with AI, but only the writing is protected because the images were made with Midjourney...

Another section implies MAYBE you can have some copyright protection

It begins by asking �whether the `work' is basically one of human authorship, with the computer [or other device] merely being an assisting instrument, or whether the traditional elements of authorship in the work (literary, artistic, or musical expression or elements of selection, arrangement, etc.) were actually conceived and executed not by man but by a machine.� [23]

In the case of works containing AI-generated material, the Office will consider whether the AI contributions are the result of �mechanical reproduction� or instead of an author's �own original mental conception, to which [the author] gave visible form.� [24]

The answer will depend on the circumstances, particularly how the AI tool operates and how it was used to create the final work.[25]

This is necessarily a case-by-case inquiry.

# ? Mar 20, 2023 21:23

KillHour: Oct 28, 2007

Comfy Fleece Sweater posted:

Basically, no copyright protection on anything you make with AI. Soo... if you're using it for something that might need copyright, beware (videogames, comics, etc come to mind).

My reading on this is that the specific thing you make is under copyright, but the raw assets that were generated by AI are not. Which is incredibly dumb, but probably has no real bearing because except in really trivial situations where the assets are used unmodified, nobody will be able to get the original assets to "steal" them, and you're under no obligation to provide them either. It's the same as if you used some public domain art in your project.

# ? Mar 20, 2023 21:33

Cousin Todd: Jul 3, 2007; Grimey Drawer

Christ not again...

# ? Mar 20, 2023 21:41

Cabbages and VHS: Aug 25, 2004; Listen, I've been around a bit, you know, and I thought I'd seen some creepy things go on in the movie business, but I really have to say this is the most disgusting thing that's ever happened to me.

The government, or any government, does not have the resources, will or interest to investigate any of this on a "case by case basis". We'll probably get a bunch of dumb precedents that are terrible for actual artists, middling to bad for the average person, and highly favorable for Disney.

And, as always, from left to right:
Cabbages with paint pens ; cabbages using this as img2img input and trying to apply "my only colors but more photorealistic" ; cabbages using this as img2img input and then applying progressively more and more Giger prompting ; cabbages simply using SD with text input to make a cartoon picture of a penguin with a bunch of specific characteristics

To me, the layperson who pushed pigments/buttons that resulted in these images, the idea that I'm more or less responsible for the creation of any one of them is pigheaded hogwash; these things existed because I wanted them to, they were "done" when I decided they were, and they're all some amalgamation of things some part of me wanted to express based on what I know about the world. Stuff that comes out of SD is constrained by the productive limits of the model(s) it uses and the ability of an operator to understand and work with them; stuff that comes out of paint pens is constrained by the palette available and ability of an operator to understand and work with them.

I'm not, at all, saying that drawing poo poo by hand is the same as pooping out batches of SD, but these activities have become so interconnected for me that I am rarely doing one and not doing both at this point. I guess I can just be an interested bystander in the escalating legal battles, at least until my own job is slurped by GPT.

Cabbages and VHS fucked around with this message at 21:58 on Mar 20, 2023

# ? Mar 20, 2023 21:56

Moongrave: Jun 19, 2004; Finally Living Rent Free

Comfy Fleece Sweater posted:

Basically, no copyright protection on anything you make with AI.

https://www.reddit.com/r/StableDiffusion/comments/11tls9x/no_matter_how_much_you_capitalize_your_work_will/

# ? Mar 20, 2023 22:26

IShallRiseAgain: Sep 12, 2008; Well ain't that precious?

Cabbages and Kings posted:

The government, or any government, does not have the resources, will or interest to investigate any of this on a "case by case basis". We'll probably get a bunch of dumb precedents that are terrible for actual artists, middling to bad for the average person, and highly favorable for Disney.

And, as always, from left to right:
Cabbages with paint pens ; cabbages using this as img2img input and trying to apply "my only colors but more photorealistic" ; cabbages using this as img2img input and then applying progressively more and more Giger prompting ; cabbages simply using SD with text input to make a cartoon picture of a penguin with a bunch of specific characteristics

To me, the layperson who pushed pigments/buttons that resulted in these images, the idea that I'm more or less responsible for the creation of any one of them is pigheaded hogwash; these things existed because I wanted them to, they were "done" when I decided they were, and they're all some amalgamation of things some part of me wanted to express based on what I know about the world. Stuff that comes out of SD is constrained by the productive limits of the model(s) it uses and the ability of an operator to understand and work with them; stuff that comes out of paint pens is constrained by the palette available and ability of an operator to understand and work with them.

I'm not, at all, saying that drawing poo poo by hand is the same as pooping out batches of SD, but these activities have become so interconnected for me that I am rarely doing one and not doing both at this point. I guess I can just be an interested bystander in the escalating legal battles, at least until my own job is slurped by GPT.

Its really only going to stop the really lazy people that try to sell automatically generated content with no human input. Its pretty easy to bypass. Especially if you consider the fact that its really easy to muddy the waters and make it unclear what is AI generated and what isn't.

# ? Mar 20, 2023 22:32

RPATDO_LAMD: Mar 22, 2013; 🐘🪠🍆

you can sell uncopyrighted stuff just fine
you just can't sue people for pirating it

not like most small independent artists are doing anything about piracy anyways

# ? Mar 20, 2023 22:33

deep dish peat moss: Jul 27, 2006

Problem: AI cannot generate pictures of planets.

Epiphany: Computers are dumb, cold, logical machines. AI does not know what a planet is, but it does know what a map is, and it knows what a sphere is.

Galaxy-Brain Solution: [description of map] wrapped tightly around a sphere floating in outerspace

How did it take over a year for this to dawn on me :doh:

# ? Mar 20, 2023 23:35

Ruffian Price: Sep 17, 2016

RPATDO_LAMD posted:

AI Dungeon launched in May 2019, almost 4 years ago

This was July 2020:

following, as you might imagine, the same discussion that gets rehashed every minor update. Even the "moon moon, but it's multiple paragraphs" explosion on the forums happened already in February 2019 with the public Talk to Transformer interface for GPT-2 (124M parameters :allears:

how quaint)

# ? Mar 20, 2023 23:59

busalover: Sep 12, 2020

Another multi-modal video generator

https://research.runwayml.com/gen2

Input:
A low angle shot of a man walking down a street, illuminated by the neon signs of the bars around him.

Output:
https://i.imgur.com/NHyC2cb.mp4

# ? Mar 20, 2023 23:59

cinnamon rollout: Jun 12, 2001; The early bird gets the worm

deep dish peat moss posted:

Problem: AI cannot generate pictures of planets.

Epiphany: Computers are dumb, cold, logical machines. AI does not know what a planet is, but it does know what a map is, and it knows what a sphere is.

Galaxy-Brain Solution: [description of map] wrapped tightly around a sphere floating in outerspace

How did it take over a year for this to dawn on me

These are really great

# ? Mar 21, 2023 00:20

BrainDance: May 8, 2007; Disco all night long!

LASER BEAM DREAM posted:

Not sure where else to post this, but I feel like people here may understand. I spent the entire weekend troubleshooting llama in text-generation-webui with the creator of reddit.com/r/locallama and oobagooba. After 4 failed installs and a system format, we finally figured out that you need a specific dev version of cudatoolkit in order to build the 4bit support module in WSL Ubuntu along with a bunch of other little tricks.

After spending the weekend in Ubuntu shell I went to fire up my stable diffusion instance and typed this into powershell
code:
cd /c:\s<tab>
My brain is broken

Which dev version of CUDA? Do you have more details like what goes wrong if you use the wrong version?

I'll probably be writing more guides and updating my guides to include stuff about 4bit models (finetuning them is in the early stages though) and I always try to include any of the differences between baremetal Linux and WSL, so, after my 15 hour building deepspeed marathon, thanks for the heads up on this lol.

# ? Mar 21, 2023 00:34

Tagichatn: Jun 7, 2009

Looks like being good at prompting can net you $300k a year! https://jobs.lever.co/Anthropic/e3cde481-d446-460f-b576-93cab67bd1ed

Time to bust out my best prompts for that interview, ((((big breasts:1.8))))), goblin
Negative: hands, feet

# ? Mar 21, 2023 00:45

HOMOEROTIC JESUS: Apr 19, 2018; Verily I say unto thee, That this night, before the cock crow, thou shalt deny me thrice.

Tagichatn posted:

Looks like being good at prompting can net you $300k a year! https://jobs.lever.co/Anthropic/e3cde481-d446-460f-b576-93cab67bd1ed

Time to bust out my best prompts for that interview, ((((big breasts:1.8))))), goblin
Negative: hands, feet

how dare you put feet in the negative prompt

# ? Mar 21, 2023 01:17

KakerMix: Apr 8, 2004; 8.2 M.P.G.

There is an entire industry being born right before our eyes in both interacting with and researching these AI systems that are bursting onto the scene.
I think it's cool as hell, change is good. :colbert:

# ? Mar 21, 2023 01:22

LASER BEAM DREAM: Nov 3, 2005; Oh, what? So now I suppose you're just going to sit there and pout?

BrainDance posted:

Which dev version of CUDA? Do you have more details like what goes wrong if you use the wrong version?

I'll probably be writing more guides and updating my guides to include stuff about 4bit models (finetuning them is in the early stages though) and I always try to include any of the differences between baremetal Linux and WSL, so, after my 15 hour building deepspeed marathon, thanks for the heads up on this lol.

This reddit post is a fantastic resource, and details almost everything you need for a fresh install. The first entry of the troubleshooting section has the cuda toolkit info. I tried to post it and hit the Cloudflare code blocker.

If you get any errors, search for them in the body of the reddit thread. I've had very little luck getting support outside of reddit or the official GitHub issues section.

The only thing not documented I ran into is that I must fix part of the CUDA driver after every restart of the system or WSL as documented here. I created a shell script to automate it.

code:

cd /usr/lib/wsl/lib
sudo rm libcuda.so libcuda.so.1
sudo ln -s libcuda.so.1.1 libcuda.so.1
sudo ln -s libcuda.so.1 libcuda.so
sudo ldconfig

Side note: Wow, I've never been blocked from posting code before! When you hit the error it reads like you've been IP banned from SA.

LASER BEAM DREAM fucked around with this message at 01:34 on Mar 21, 2023

# ? Mar 21, 2023 01:29

deep dish peat moss: Jul 27, 2006

KakerMix posted:

There is an entire industry being born right before our eyes in both interacting with and researching these AI systems that are bursting onto the scene.
I think it's cool as hell, change is good.

It's really cool and neat just trying to understand how AI interprets things and how to solve the problem(s) of "ungenerateable concepts", like in my last post with planets. I've been puzzling on how to generate pictures of planets for literally over a year and I've consistently run into problems: just by its very nature the dataset is full of photos of planet earth and World Maps of planet earth that dominate the association to words like "planet" or "world" or "globe", making those words useless when trying to generate those things when the subject is anything but Earth. Similarly, the word "moon" or associated words like "lunar" always try to turn the prompt into a landscape scene with a moon in the sky, because there's a monumental amount of photos of the moon that dominate any association to the word "moon" - generating a fictional sci-fi moon was impossible.

Then the actual solution is the most obvious-once-you-see-it "just be literal" solution ever - describe it in purely geometric 3d modeling terms: a texture wrapped over a sphere. It's just a kind of word puzzle that's very addictive for me to solve: how to describe something in a truly unambiguous way.

Anyway, here's a topographical map of a fictional forest moon tightly wrapped over a sphere floating in outerspace:

And it's not even like this solution is unique to v5, it has a pretty high success rate in v4 too!

deep dish peat moss fucked around with this message at 02:08 on Mar 21, 2023

# ? Mar 21, 2023 02:03

Soulhunter: Dec 2, 2005

Women laughing at salads: A story in 12 grids.

Bonus Xenos laughing at salads:

Plus one hybrid:

# ? Mar 21, 2023 02:08

Confusedslight: Jan 9, 2020

Those are incredible. 10/10

# ? Mar 21, 2023 02:21

BrainDance: May 8, 2007; Disco all night long!

LASER BEAM DREAM posted:

Side note: Wow, I've never been blocked from posting code before! When you hit the error it reads like you've been IP banned from SA.

I had that when I tried to post my guide, too. It was wget and then any http after it for me. Removing the http part didn't hit the filter for some reason, which is weird because they still work without it so I'd imagine any hypothetical attack still would, too.

# ? Mar 21, 2023 05:24

KwegiboHB: Feb 2, 2004; nonconformist art brut
Negative prompt: amenable, compliant, docile, law-abiding, lawful, legal, legitimate, obedient, orderly, submissive, tractable
Steps: 32, Sampler: DPM++ 2M Karras, CFG scale: 11, Seed: 520244594, Size: 512x512, Model hash: 99fd5c4b6f, Model: seekArtMEGA_mega20

Comfy Fleece Sweater posted:

https://www.federalregister.gov/doc...al-intelligence
[i]
II. The Human Authorship Requirement

Thanks for posting this. I just read it however it's worth me re-reading it and thinking on. I'll have a better reply tomorrow.

Tagichatn posted:

Looks like being good at prompting can net you $300k a year!

I refuse to believe this is a thing until someone can actually handle the ultimate prompt
https://en.wikipedia.org/wiki/Archimedes%27s_cattle_problem

If thou art diligent and wise, O stranger, compute the number of cattle of the Sun, who once upon a time grazed on the fields of the Thrinacian isle of Sicily, divided into four herds of different colours, one milk white, another a glossy black, a third yellow and the last dappled. In each herd were bulls, mighty in number according to these proportions: Understand, stranger, that the white bulls were equal to a half and a third of the black together with the whole of the yellow, while the black were equal to the fourth part of the dappled and a fifth, together with, once more, the whole of the yellow. Observe further that the remaining bulls, the dappled, were equal to a sixth part of the white and a seventh, together with all of the yellow. These were the proportions of the cows: The white were precisely equal to the third part and a fourth of the whole herd of the black; while the black were equal to the fourth part once more of the dappled and with it a fifth part, when all, including the bulls, went to pasture together. Now the dappled in four parts were equal in number to a fifth part and a sixth of the yellow herd. Finally the yellow were in number equal to a sixth part and a seventh of the white herd. If thou canst accurately tell, O stranger, the number of cattle of the Sun, giving separately the number of well-fed bulls and again the number of females according to each colour, thou wouldst not be called unskilled or ignorant of numbers, but not yet shalt thou be numbered among the wise.
But come, understand also all these conditions regarding the cattle of the Sun. When the white bulls mingled their number with the black, they stood firm, equal in depth and breadth, and the plains of Thrinacia, stretching far in all ways, were filled with their multitude. Again, when the yellow and the dappled bulls were gathered into one herd they stood in such a manner that their number, beginning from one, grew slowly greater till it completed a triangular figure, there being no bulls of other colours in their midst nor none of them lacking. If thou art able, O stranger, to find out all these things and gather them together in your mind, giving all the relations, thou shalt depart crowned with glory and knowing that thou hast been adjudged perfect in this species of wisdom.
Steps: 64, Sampler: DPM++ 2M Karras, CFG scale: 11, Seed: 3034435310, Size: 768x768, Model hash: 99fd5c4b6f, Model: seekArtMEGA_mega20

(I have a long way to go myself)

# ? Mar 21, 2023 05:55

Tunicate: May 15, 2012

KwegiboHB posted:

Thanks for posting this. I just read it however it's worth me re-reading it and thinking on. I'll have a better reply tomorrow.

I refuse to believe this is a thing until someone can actually handle the ultimate prompt
https://en.wikipedia.org/wiki/Archimedes%27s_cattle_problem

If thou art diligent and wise, O stranger, compute the number of cattle of the Sun, who once upon a time grazed on the fields of the Thrinacian isle of Sicily, divided into four herds of different colours, one milk white, another a glossy black, a third yellow and the last dappled. In each herd were bulls, mighty in number according to these proportions: Understand, stranger, that the white bulls were equal to a half and a third of the black together with the whole of the yellow, while the black were equal to the fourth part of the dappled and a fifth, together with, once more, the whole of the yellow. Observe further that the remaining bulls, the dappled, were equal to a sixth part of the white and a seventh, together with all of the yellow. These were the proportions of the cows: The white were precisely equal to the third part and a fourth of the whole herd of the black; while the black were equal to the fourth part once more of the dappled and with it a fifth part, when all, including the bulls, went to pasture together. Now the dappled in four parts were equal in number to a fifth part and a sixth of the yellow herd. Finally the yellow were in number equal to a sixth part and a seventh of the white herd. If thou canst accurately tell, O stranger, the number of cattle of the Sun, giving separately the number of well-fed bulls and again the number of females according to each colour, thou wouldst not be called unskilled or ignorant of numbers, but not yet shalt thou be numbered among the wise.
But come, understand also all these conditions regarding the cattle of the Sun. When the white bulls mingled their number with the black, they stood firm, equal in depth and breadth, and the plains of Thrinacia, stretching far in all ways, were filled with their multitude. Again, when the yellow and the dappled bulls were gathered into one herd they stood in such a manner that their number, beginning from one, grew slowly greater till it completed a triangular figure, there being no bulls of other colours in their midst nor none of them lacking. If thou art able, O stranger, to find out all these things and gather them together in your mind, giving all the relations, thou shalt depart crowned with glory and knowing that thou hast been adjudged perfect in this species of wisdom.
Steps: 64, Sampler: DPM++ 2M Karras, CFG scale: 11, Seed: 3034435310, Size: 768x768, Model hash: 99fd5c4b6f, Model: seekArtMEGA_mega20

(I have a long way to go myself)

I remember that one

https://www.youtube.com/watch?v=FavUpD_IjVY

# ? Mar 21, 2023 06:17

TIP: Mar 21, 2006; Your move, creep.

update on my video AI experiment

prompt: spiderman in new york city, illustrated by Marcos Martin, marvel comics, sharp, professional, award winning
input video:

higher resolution and now with no frame limit!

this is really making me want a new videocard with lots of ram, with this technique I'm stuck rendering each frame at 384x384 and I feel like I could get even more coherent results with higher resolutions

# ? Mar 21, 2023 08:09

Dick Trauma: Nov 30, 2007; God damn it, you've got to be kind.

Tagichatn posted:

Looks like being good at prompting can net you $300k a year! https://jobs.lever.co/Anthropic/e3cde481-d446-460f-b576-93cab67bd1ed

Time to bust out my best prompts for that interview, ((((big breasts:1.8))))), goblin
Negative: hands, feet

Wait, what does "goblin" do? Make... goblins?

# ? Mar 21, 2023 09:59