Fan Art Kingdom of Deception and Noxian Nights [fan art, ai models]

[goatwhisperer3]

Look mom, no hands!

I got it running. Thanks for the help guys.

Will probably download a shitton of that lora stuff and then proceed to experiment.

Spoiler: Follow up questions

You don't have permission to view the spoiler content. Log in or register now.

 

[KGARrokh]

Look mom, no hands!

I got it running. Thanks for the help guys.

Will probably download a shitton of that lora stuff and then proceed to experiment.

View attachment 4759568

Spoiler: Follow up questions

You don't have permission to view the spoiler content. Log in or register now.

1) Not necesserily. After i found interesting image, i just turn up upscaler or ADetailer, or both in txt2img, then just regenerate image once more (i don't like switching tabs).

Upscaler allows you to enchance image, but not always work.

Spoiler: Example before upscale:

You don't have permission to view the spoiler content. Log in or register now.

Spoiler: Example after upscale:

You don't have permission to view the spoiler content. Log in or register now.

Spoiler: Settings that i prefer:

You don't have permission to view the spoiler content. Log in or register now.

I use img2img tab when i have reference image (anime screenshot or other ai image).
Can't remember when i last used Inpaint tab)


2) Camera.
This is how AI differs from good old 3D) You don't have direct camera controls, only indirect by tags and hope that ai "hears" you.

Spoiler: Useful tags that may help:

You don't have permission to view the spoiler content. Log in or register now.

You can combine this tags or place them in negative prompt.

3) Illustrious and Pony both are based on SDXL1.0. Illustrious checkpoint should work with your UI, but if not i recommend reForgeUI. Later versions of Forge are not very stable, at least in my experience.

 

[Mamaragan]

Look mom, no hands!

I got it running. Thanks for the help guys.

Will probably download a shitton of that lora stuff and then proceed to experiment.

View attachment 4759568

Spoiler: Follow up questions

You don't have permission to view the spoiler content. Log in or register now.

Yes, img2img, is a good start to modify a generated picture. But the inpaint tab, just next to it, give you more control to what you want to edit.

For perspective, use danbooru tag like facing viewer, cowboy shot, from below, from above, from side, wide shot, distant, and so on.
Here's a list of tags to use

You must be registered to see the links

You can also download this extension, which autocomplete tags

You must be registered to see the links

You should also be familiar with “weight”. Weight reinforce a prompt and make it stronger. For example, if you want to make a perspective stronger, or anything really, select a prompt, and then ctrl + arrow up or down to decide if you want to enhance or weaken a tag. Never go beyond a certain threshold, or it's going to fuck with the render. Just experiment with increment.

Pony and Illustrious are both based on SDXL, so you can use an illustrious model like WAI or HassakuXL with forge no problem.

And just in case you didn't know. Use specific resolution from which all those SDXL based model are trained.

 

[Mamaragan]

 

[KGARrokh]

Upscaler can't fix creature eyes, so "he" added details for boobs) And another experiments with KOD LoRA.

 

[Mamaragan]

 

[big_shl0m0]

View attachment 4768225 View attachment 4768227 View attachment 4768228 View attachment 4768231 View attachment 4768230 View attachment 4768258

really loving her bitchy resting face in the last two images

 

[KGARrokh]

My attemps to use Sora from OpenAI:

Firts image looks like fail. Second is much better, environment looks really natural. Sword is weird, but this my fault, i typed greatsword instead of just sword. You can Sora by yourself at sora.com (3 images per day for free users).

It's really interesting how much VRAM needed to run this generative AI, to better realize gap beetwen Sora and Flux (or similar models).

 

[KGARrokh]

Sora supports other languages too! Most amazing that Sora can generate more then one character at image, 7-8 characters is possible. Architecture is also a lot better, then with SD AIs. Plus supports video generation, which is aslo great.

Don't know how to describe her folded ponytail to Sora)

Looks decent, but failed to cast shadow from dragon on village.

 

[Jinsoyun]

"Can we have Sabia from KoD?"
"We have Sabia from fucking Dark Souls at home".

Jokes aside, the last couple of pages were especially good.

 

[KGARrokh]

"Can we have Sabia from KoD?"
"We have Sabia from fucking Dark Souls at home".

Jokes aside, the last couple of pages were especially good.

All I can offer now is a fantasy illustration with Sabia. For now I'm testing the models I downloaded on similar (fantasy) prompts.

I'll generate something sexier with Sabia later.

 

[Mamaragan]

 

[zorin0815]

All I can offer now is a fantasy illustration with Sabia. For now I'm testing the models I downloaded on similar (fantasy) prompts.
I'll generate something sexier with Sabia later.

That reminds me of the Witchblade Comics! (

You must be registered to see the links

)

 

[miltonlucifer]

guys. guys.
I just stumbled upon this thread, by checking the progress on KoD on f95 and patreon. and I'm sorry if i'm off topic, but can I just say, you are all amazing. watching this whole thread page by page has been fanfuckingtastic and super hot.

Bravo. I feel like you could all create a new spinoff of the game with 10times more content than the original if you all worked together. so much hotness around here. Sabia and the blonde lass ( forgot her name ) are my favourite spicy game characters by far ( along with Saria from saria reclaimed, and Karryn from Karryn's prison. I seem to have a thing for purpled hair-moody arrogant sluts. Weird. New fetish unlocked I guess ) . We don't have enough mature female game characters that have normal boobs /athletic body proportions. they always end up with unreasonable breast size , a mom belly or stupidly thick thighs and ass.

anyhow, this has been a joy to look at every single pic on every single post of this whole thread.
cheers to you all.

 

[Enigmanic]

...I feel like you could all create a new spinoff of the game with 10times more content than the original if you all worked together..

I can write a new story in my sleep. Just need a dedicated coder and a couple artists.

 

[goatwhisperer3]

weekend so i played around again.

Spoiler: small rant

You don't have permission to view the spoiler content. Log in or register now.

 

[Mamaragan]

Can you post all the prompt you used?
Do you use any upscaling technique?

 

[goatwhisperer3]

unless ForgeUI has some sort of "open last project" or anything like that I doubt I still have the original prompts anymore

 

[Mamaragan]

unless ForgeUI has some sort of "open last project" or anything like that I doubt I still have the original prompts anymore

The prompt is embedded within the picture metadata. Generally, you put a generated picture in the png tab, and it gives you all information you need. As long as they weren't overwritten, that is.
You can click on that blue arrow to load the last parameters, prompt included.

 

[KGARrokh]

View attachment 4782008

weekend so i played around again.

Spoiler: small rant

You don't have permission to view the spoiler content. Log in or register now.

You need to realize that prompt is not strict instruction to generative AI. Every tag in positive/negative prompt increases/decreases probability to get desired result, if model know what tag means. There is a lot of tags that does't produce any response signal form model (example something weird like this "5K22v" or some rare words).
High strength values can cause many problems like hair color change, extra hand, overfitting noise and so on.
If AI doesn't know something, then you can't produce it (example Sora doesn't know how Sabia looks, so it produces image with wront hairstyle). Some tags may invoke more concepts than you want (triggers is a good example, they used to invoke all relative to LoRA concepts with differents strength).

If you want some rare view angle (concept), then you probably need to find/create some reference, then use it with img2img and/or controlnet. Also you may try to create specific LoRA to this view angle (concept), this should increase probability to get proper result.

"... and so on until me and mr stable diffusion meet on common ground but i feel robbed"
I have tried to create some scenes 200+ times and nothing) Now i try 10 times in average (with good prompts more, with bad less), if i don't get ddesired result, then i switched to another scene.

Current generation of AI's is another kind of lootboxes with some parameters to control)