Stable Diffusion

[Sandman]

Just in case somebody have missed it: Stable Diffusion is another one of those "describe something and the computer makes the art for you" solutions. (DALL-E is another famous one.) The thing with Stable Diffusion is that it's Open Source, and it seems to be several levels better than the competitors.

There are so many webpages about this that I don't really know what to link to, but here's the webpage itself, which is a bit dry:

Stable Diffusion Public Release — Stability AI

We are delighted to announce the public release of Stable Diffusion and the launch of DreamStudio Lite.

stability.ai

Here's an article about it, together with an amazing example:

Stable Diffusion is a really big deal

If you haven’t been paying attention to what’s going on with Stable Diffusion, you really should be. Stable Diffusion is a new “text-to-image diffusion model” that was released to the …

simonwillison.net

And here's a picture from Wikipedia:

An image generated with Stable Diffusion from the prompt "Oversaturated burned light leak expired film photo of a woman's resting face submerged in a bubbly opaque liquid"


Note: I'm neither a user of this, nor do I a whole lot about it. I'm just amazed by it.

 

[Magma]

Wow... is that real (not the girl in photo) - but - is it possible that ???

 

[Sandman]

Yep, it seems so. Reminds me of this quote from Arthur C. Clarke:

“Any sufficiently advanced technology is indistinguishable from magic.”

 

[Magma]

until today thought that a "computer" with a help of software can easily create those graphics:

but this is really "magic" ... not trolling us ?

 

[Hamied Abou Hulaikah]

Amazing!
What is the conventional name of this (think-to-image)?
for example chat-bot response is built on AI conversation.

 

[Hamied Abou Hulaikah]

Ok, I figure it:
Model name: stable diffusion
Model url: https://huggingface.co/CompVis/stable-diffusion-v1-4
Type: text-to-image

trying to imagine ...

 

[kimstudio]

Nice! Are all these pixels generated by AI? or it just grabs some photos from all available photos on internet and combine part of them together?

 

[Hamied Abou Hulaikah]

Nice! Are all these pixels generated by AI? or it just grabs some photos from all available photos on internet and combine part of them together?

It is a hybrid of both.

 

[Sandman]

This isn't created with Stable Diffusion, but an alternative software. Still interesting.

An AI-Generated Artwork Won First Place at a State Fair Fine Arts Competition, and Artists Are Pissed

Jason Allen's AI-generated work "Théâtre D'opéra Spatial" took first place in the digital category at the Colorado State Fair.

www.vice.com

 

[rabbitBUSH]

ran the demo : an elephant will roam through water wallowing in steam
not quite as creative as one might have hoped but . . .

 

[Sandman]

not quite as creative as one might have hoped

Try finishing the prompt with something like "in the style of Picasso"

 

[rabbitBUSH]

"in the style of Picasso"

how about stormy daniels rather?

 

[Sandman]

Another interesting article for those interested in running Stable Diffusion on a more basic machine. Also contains some nice examples of prompts and resulting images.

WARNING: Some of the images depict tits, which might not be ok for everybody. (But the page also contain blueprints for a time machine, so there's that.)

Running Stable Diffusion on your GPU with less than 10GB of VRAM on Windows

Constant Meiring - Technology Product Builder

constant.meiring.nz

 

[rabbitBUSH]

WARNING: Some of the images depict tits

No that's alright as long as they fit in a C'-cup

 

[Sandman]

I'm no expert, but it looks like they would do that.