Artificial Intelligence Generated ART

neosapien said:

MC Escher's Relativity was perhaps the first art I saw and was like that's cool. I wonder who did that?

Funny story… my wife was in the hospital getting induced. The nurse hooked her up to the Pitocin drip and they kept coming in every so often to check. With nothing happening. So I'm bored, nervous, anxious looking at everything and was like shit, I ain't a medical worker but that's hooked up to nothing. So the nurse comes in again and I'm like "hey I'm not a nurse or nothing but it's looks like there's an MC Escher staircase going on with those tubes". And she was of course like "wut?" And I was like she ain't hooked up to the machine you stupid bitch!! How could you not get my obscure art reference!!??? OK maybe I didn't say the last couple parts but yeah. They never hooked her into the machine lol. And looking at all the tubes running to nothing reminded me of Relativity lol.

Click to expand...

Dr.Amber Trichome said:

wow, lucky for her you caught that the nurse fucked up. What an idiot! Thats an interesting comparison.
I love his work and have tried to copy by pencil some of his designs, but they are too difficult to reproduce by hand for me.
So... after digging into the program a bit more the other evening I somehow got to a page that had the Pro Night Cafe stars.
These artists are deep into it and have gotten so good at it they have become Pro and have links to sell prints and whatnot of their creations. There are some really heavy hitting superstars and I found one guy who did this very trippy ship floating in air and somehow i was able to hover over his text prompts that revealed his secret to how he sorta kinda got there. At least the text prompts i read were the last in his training of the image. Those words revealed quite a bit to me as to how one gets a really sweet image. What struck me as interesting was that he used 35 mm camera and many artists names and other people's names who i have no idea. What i should have done was screenshot that image and investigated deeper. Putting in text prompts to increase optimal pixel density and adjust resolution seems to benefit the image. Maybe using like 8K UHD (7680 x 4320) .

Click to expand...

Dr.Amber Trichome said:

wow, lucky for her you caught that the nurse fucked up. What an idiot! Thats an interesting comparison.
I love his work and have tried to copy by pencil some of his designs, but they are too difficult to reproduce by hand for me.
So... after digging into the program a bit more the other evening I somehow got to a page that had the Pro Night Cafe stars.
These artists are deep into it and have gotten so good at it they have become Pro and have links to sell prints and whatnot of their creations. There are some really heavy hitting superstars and I found one guy who did this very trippy ship floating in air and somehow i was able to hover over his text prompts that revealed his secret to how he sorta kinda got there. At least the text prompts i read were the last in his training of the image. Those words revealed quite a bit to me as to how one gets a really sweet image. What struck me as interesting was that he used 35 mm camera and many artists names and other people's names who i have no idea. What i should have done was screenshot that image and investigated deeper. Putting in text prompts to increase optimal pixel density and adjust resolution seems to benefit the image. Maybe using like 8K UHD (7680 x 4320) .

Click to expand...

neosapien said:

That's some amazing detective work. Your web-fu is impressive! That's super interesting about all the text prompts and what illicits a better creation and shit!

Click to expand...

Dr.Amber Trichome said:

Thanks, that’s how I roll dude.
View attachment 5196416

Click to expand...

neosapien said:

@Dr.Amber Trichome

A couple I've done as of late…

View attachment 5237996View attachment 5237997View attachment 5237998View attachment 5237999

Click to expand...

Dr.Amber Trichome said:

Those a crazy insanely amazing. You truly have a gift for doing these! I don’t have any desire to do Ai anymore. I want to get back into pencil drawings. taking this intro to Ai class this past semester and at first I was totally into it but as time went on lost all interest and found a renewed spirit and avenue in my long held career .Ai given me an new appreciation for the primitive human being. My Ai teacher failed my final without hesitation. I got a big zero for what I thought was a totally rad submission. It was a drawing of a neural network. She said it was very artistically creative but not what she was looking for. What a fucking bitch! It is a design for my invention of earbud that are designed to pick up brainwaves that trigger anxiety and tune into people’s intuition to help them make better decisions. View attachment 5238025

Click to expand...

neosapien said:

That's crazy awesome! Fuck that bitch. Just jaded and jealous. That's a crazy good drawing and a crazy good idea. Don't let it discourage you. You are brilliant. Telling someone what is or isn't good art is like telling someone what is or isn't beautiful. Who are you to tell me what my feelings are? You are really talented. I could never draw let alone think of something like that. I can barely make a stick figure lol.

I still think the AI stuff is cool. Partly cuz it's new I'm sure. But it really is the same argument as "hey look at this custom hand carved craftsman walking stick" versus "a machine in China pumped a million of these things out".

Keep doing you. Cuz I think you're great.

Click to expand...

xtsho said:

Cool stuff. I've been working with AI and Machine learning lately. Specifically the backend technology. I have an AI image generator running locally that functions the same as the online generators people are using. It would be a fairly simple task to deploy it into the cloud environment and incorporate it into a website.

I haven't used any of the online generators so I don't know how many parameters are available for users but I have a wealth of parameters I'm in control of and able to change that affect the results.

I took the OP's initial image, did an image to image generation while also an interrogation of the clip/image. The interrogation of the clip is just a text interpretation of the image. The image is processed and the application if trained can determine what the image is. If it's a car the text data that would be returned would be "car" and possibly the model and year of the car depending on how much data was used to train the AI engine. It's all about data.

I took the OP's image and did an image to image generation.

The text returned was very close to describing the image that was posted by the OP in the original post. This is the text that was generated:

"a penguin is standing in a room with neon lights and a man in a black suit is standing in the background"

There's a penguin and something in the background.

View attachment 5238262

Here's the photo after I did the image to image generation. Now it definitely is "a penguin is standing in a room with neon lights and a man in a black suit is standing in the background". Pretty cool.

View attachment 5238263



Here's some images I generated and the words I used.

Bougie Food


View attachment 5238265

Click to expand...

neosapien said:

That's cool. You're light years ahead of me with that stuff. As I've claimed since the beginning, I just type some text in a box and click some preset buttons. That doesn't make me an artist. Nor a computer guru. But you got some skills. I'm the very definition of an end user lol. The food picture is pretty damn realistic. Feel free to post some more stuff as time permits you. I'll certainly be interested in your work.

Click to expand...

High-Resolution Image Synthesis with Latent Diffusion Models - Computer Vision & Learning Group

By decomposing the image formation process into a sequential application of denoising autoencoders, diffusion models (DMs) achieve state-of-the-art synthesis results on image data and beyond. Stable diffusion.

ommer-lab.com

GitHub - AUTOMATIC1111/stable-diffusion-webui: Stable Diffusion web UI

Stable Diffusion web UI. Contribute to AUTOMATIC1111/stable-diffusion-webui development by creating an account on GitHub.

github.com

GitHub - Stability-AI/stablediffusion: High-Resolution Image Synthesis with Latent Diffusion Models

High-Resolution Image Synthesis with Latent Diffusion Models - Stability-AI/stablediffusion

github.com