Viridian's AI Experiments

Artificial Intelligence is here! Really! Anything created with AI assistance, including stories, should be posted here.
User avatar
sixgunzloaded
Posts: 918
Joined: Tue May 05, 2015 8:16 pm

Re: Viridian's AI Experiments

Postby sixgunzloaded » Sat Mar 25, 2023 9:51 pm

MadMax359 wrote:these are working so well i am getting picky about which ones to download! :twisted:


There's an easy solution to that. :D lol
How long did Tarzan watch before deciding to save Jill..?

Viridian
Posts: 1587
Joined: Wed Apr 15, 2009 10:03 am

Re: Viridian's AI Experiments

Postby Viridian » Sun Mar 26, 2023 2:10 am

sixgunzloaded wrote::shock: :shock: :shock:
The detail - especially when zoomed out - look almost photographic. Most of them certainly don't look like anime or manga anymore.

I've only spent a day playing with the Stable Diffusion models on Mage.Space. I've been trying to tweak my formula for SD. If it's a photorealistic look we want, especially with better quicksand/mud textures, let's take it to the next level.

There's a specific charm to digital illustration styles, but flip the switch to go all-in on photographic detail and we have a different game.
You do not have the required permissions to view the files attached to this post.
Viridian @ deviantART: http://viridianqs.deviantart.com/

Viridian
Posts: 1587
Joined: Wed Apr 15, 2009 10:03 am

Re: Viridian's AI Experiments

Postby Viridian » Sun Mar 26, 2023 4:25 am

sixgunzloaded wrote:Dude, if you could find a way to make the program give you the same character and bog/background every time, you could make one hell of a good looking storyline.

While this isn't feasible yet, given that the AI doesn't memorise what is the "same" but recreates patterns that it recognises, the inpainting tools present some more customisation. Earlier you mentioned that in my NovelAI renders, details will change between each version, and generating the right image is a bit of a juggling act. Mage.Space has a more flexible function where you can identify which parts to keep and which parts to re-render. Using the Susan Desert image, let's say I want to keep the character but swap out the background. I highlight the area I want to work with (or use the auto-background function).
Screenshot 2023-03-26 at 13-13-51 Mage Free Fast Unfiltered Stable Diffusion. Generate Anything You Can Imagine!.png

Then I put in my new prompt and it will only apply it to that section. In this case, I swap out the dry sand with something more gloopy.
Screenshot 2023-03-26 at 13-18-27 Mage Free Fast Unfiltered Stable Diffusion. Generate Anything You Can Imagine!.png

And let's re-do the face.
Screenshot 2023-03-26 at 13-20-49 Mage Free Fast Unfiltered Stable Diffusion. Generate Anything You Can Imagine!.png

And here's the end product:
a3ff096680f24d1b800c9a4cc886766c.png
You do not have the required permissions to view the files attached to this post.
Viridian @ deviantART: http://viridianqs.deviantart.com/

Viridian
Posts: 1587
Joined: Wed Apr 15, 2009 10:03 am

Re: Viridian's AI Experiments

Postby Viridian » Mon Mar 27, 2023 3:14 pm

Some comparisons between my original NovelAI renders and running it through Stable Diffusion. Honestly, while I thought I was onto something with NovelAI, I think I may have created some of the best digital renders of QS images. This is the sort of thing we dreamed of.
You do not have the required permissions to view the files attached to this post.
Viridian @ deviantART: http://viridianqs.deviantart.com/

Viridian
Posts: 1587
Joined: Wed Apr 15, 2009 10:03 am

Re: Viridian's AI Experiments

Postby Viridian » Tue Mar 28, 2023 12:14 pm

I tried to generate raw images in SD using my prompts without an image reference. The results, while less flexible and precise, are top quality.
You do not have the required permissions to view the files attached to this post.
Viridian @ deviantART: http://viridianqs.deviantart.com/

Viridian
Posts: 1587
Joined: Wed Apr 15, 2009 10:03 am

Re: Viridian's AI Experiments

Postby Viridian » Tue Mar 28, 2023 12:30 pm

And now the turning point. No image reference. First-shot full-body QS generation from Stable Diffusion.
You do not have the required permissions to view the files attached to this post.
Viridian @ deviantART: http://viridianqs.deviantart.com/

User avatar
MadMax359
Posts: 4214
Joined: Sun Apr 19, 2009 5:07 am
Location: north carolina

Re: Viridian's AI Experiments

Postby MadMax359 » Tue Mar 28, 2023 12:59 pm

wow, things are progressing fast!
The strong do what they want, the weak do what they must

Viridian
Posts: 1587
Joined: Wed Apr 15, 2009 10:03 am

Re: Viridian's AI Experiments

Postby Viridian » Thu Mar 30, 2023 12:10 pm

Comparison of renders - NovelAI vs Stable Diffusion
You do not have the required permissions to view the files attached to this post.
Last edited by Viridian on Thu Mar 30, 2023 12:56 pm, edited 1 time in total.
Viridian @ deviantART: http://viridianqs.deviantart.com/

User avatar
MadMax359
Posts: 4214
Joined: Sun Apr 19, 2009 5:07 am
Location: north carolina

Re: Viridian's AI Experiments

Postby MadMax359 » Thu Mar 30, 2023 12:54 pm

so much better!!
The strong do what they want, the weak do what they must

Viridian
Posts: 1587
Joined: Wed Apr 15, 2009 10:03 am

Re: Viridian's AI Experiments

Postby Viridian » Fri Mar 31, 2023 9:01 am

Documenting my current process with NovelAI and Stable Diffusion.

I'm starting off with a very early version of my Quicky Sanders character many years ago, which I found again. This is very obviously an inexperienced and crude drawing, prior to my development in digital art.
chased_raw.jpg

Popping this straight into SD as an image reference is tough. The AI often doesn't recognise the features that make this a human in a mudpit, partly because of how badly is drawn, but also because of the style. I'm trying to create a photorealistic output, and a flat 2D drawing doesn't transition well. It might be easier to turn this into another, higher quality 2D art style, but I want the photorealistic effect. I need to turn this into a form that SD can work with. This is where my experience with NovelAI comes in.

NovelAI's model, though based on anime, can generate some decent photorealistic styles with the right prompts. The best part is that it allows very easy editing. You could otherwise make the edits yourself using a Photoshop-style tool (and if you're really good, you can skip this step entirely and just create a photomanipulation and run that through SD). The crucial part is that you want to get the mud texture looking enough like a mud texture. A few ripples with NovelAI's editing tool and I get something like this.
chased_novelai.png

This is workable, though I cannot emphasise enough how the mud texture will influence how SD will perceive the setting. Depending on the texture you want, you may want to edit the image to have the right appearance. The quality of the character and background isn't important, as SD will work with that.
chased_sd1.png

You can really see how easily Stable Diffusion works with the output from NovelAI. It is a closer approximation to what I want, thereby removing the randomness. SD has done a decent job with the mud texture, being close to the NovelAI output, though the background, trees and mud could do some more work. I use a mask to work with the mud and background and run that a few times.
chased_sd2.png

This is already looking miles better. I try to run it a few more times as a whole image to see if I can squeeze more photorealism. The returns diminish at this point, but I ended up with this.
chased_sd3.png

It's not the most photorealistic output I've attained. I actually had far more realistic ones, though it ended up turning the character too much into the Asian side and it didn't match the mud texture. Ultimately, considering I started from a basic digital drawing, the transformation is mind-boggling good - and if I wanted to, I could have made it better.
You do not have the required permissions to view the files attached to this post.
Viridian @ deviantART: http://viridianqs.deviantart.com/


Return to “AI Creations”

Who is online

Users browsing this forum: No registered users and 0 guests