Fan Art Kingdom of Deception and Noxian Nights [fan art, ai models, waiting for updates, future plans, healthcare]

[KGARrokh]

Some fan art from this week) I just monitoring new styles and concepts, then testing them with KOD LoRA.
All art in this post generated with CivitAI generatior (only txt2img, no additional processing).

Spoiler: Realistic (PonyRealism checkpoint)

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

Spoiler: Boris Vallejo style (love this fantasy style)

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

Spoiler: Ai Love Higokko style

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

Spoiler: Some fun art

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

Spoiler: Dark Ambience style

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

Spoiler: Sabia+Neve

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

Spoiler: Other pin-ups

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

Spoiler: Other

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

 

[KGARrokh]

All art in this post was generated with ForgeUI.

Spoiler: pack 0 (ControlNet+Depth)

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

Spoiler: pack 1 (ControlNet+Depth) more failure, then success)

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

Spoiler: pack 2 (ControlNet+Depth)

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

Spoiler: pack 3 (ControlNet+Depth)

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

Spoiler: Other (i spent a lot of time trying to fix hands in last one image, but failed)

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

 

[Enigmanic]

Well it is a minotaur a fantasy race that have absurdly large dicks,that it is challanging to fit is the whole fun XD

That's a common trope but it's not written in stone. Either way, they can have large thick members without being stupidly huge.

 

[Dripstein]

Made another mini-story with Ylva this time around

Spoiler: Ylva's Interrogation Tactics

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

 

[..a..]

i would post on civitai but civit blocking image. dont know why
maybe if people share image prompt + used models so easier for community effort to creative similar image and maybe make game mod. I try to use seeds and not random because some seeds better than other seeds. easier to create similar image when you get good seed.

...

thank you for training model, can get good pictures with it.

Spoiler: Neve bonding with katherine horse

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

 

[KGARrokh]

i would post on civitai but civit blocking image. dont know why
maybe if people share image prompt + used models so easier for community effort to creative similar image and maybe make game mod. I try to use seeds and not random because some seeds better than other seeds. easier to create similar image when you get good seed.



thank you for training model, can get good pictures with it.

Spoiler: Neve bonding with katherine horse

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

If you generate via A1111, ForgeUI or CivitAI, then all necessary data (resources, settings, prompts) is included in your image as additional info (for jpg and png formats).

Spoiler: Example

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

 

[..a..]

If you generate via A1111, ForgeUI or CivitAI, then all necessary data (resources, settings, prompts) is included in your image as additional info (for jpg and png formats).

Oh ok thank you! only use civitai text to image. didnt know

 

[KGARrokh]

CivitAI now can fix faces with 3 options (face fix, high res fix, 2 previous combined), but in my few tests i always get different, from original, images.

Spoiler: Original, face fix (denoise 0.15):

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

Spoiler: Another example: original, variant, face fix (denoise 0.05):

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

 

[KGARrokh]

Also i tested a lot of styles (250+), here is source with only KOD LoRA:


And some results, that i like:

Spoiler: Part 1:

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

4, 7, 33, 47, 57 - is my favorite)

 

[KGARrokh]

Spoiler: Part 2:

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

69, 75, 76 - favorite here)

Spoiler: Links to styles in order of appearence:

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

Some styles are mixed (few different LoRA), mostly last images, but i don't this "others" LoRA's significantly changes results.

 

[Magnelo]

Spoiler: Part 2:

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

69, 75, 76 - favorite here)

Spoiler: Links to styles in order of appearence:

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

Some styles are mixed (few different LoRA), mostly last images, but i don't this "others" LoRA's significantly changes results.

While testing new styles, do you use a model + trained by you KoD lora combined? If so is it always duo or are you trying more methods? Assuming you're still on civit not ForgeUI

 

[KGARrokh]

While testing new styles, do you use a model + trained by you KoD lora combined? If so is it always duo or are you trying more methods? Assuming you're still on civit not ForgeUI

The idea was to check how my LoRA work with other styles, so it's duo for most images, and five for some others images.
I use CivitAI because i don't want to download all this styles, and don't want to use other methods besides simple txt2img (only prompting, no fixes or other postprocessing), most of the time.

 

[Magnelo]

The idea was to check how my LoRA work with other styles, so it's duo for most images, and four for some others images.
I use CivitAI because i don't want to download all this styles, and don't want to use other methods besides simple txt2img (only prompting, no fixes or other postprocessing), most of the time.

Also, is Civit lagging for you so much after the last update with face fixes?

 

[KGARrokh]

Also, is Civit lagging for you so much after the last update with face fixes?

CivitAi lagged before and lagging after update) Almost everyday something does't work right, this is annoying. Most incompetent web-developers)

 

[..a..]

CivitAI now can fix faces with 3 options (face fix, high res fix, 2 previous combined), but in my few tests i always get different, from original, images.

Spoiler: Original, face fix (denoise 0.15):

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

Spoiler: Another example: original, variant, face fix (denoise 0.05):

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

I get very same face denoise pictures. nothing change only face.

Spoiler: Sabia caring for katherine horses

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

denoise between 0.4-max
this image denoise 0.4
You can remove all original promps, only leave what you want expression and other things you want change in face. only leave original prompt for color eyes, color hair.
probably best result is get good image, run image upscale, then face fix for different expressions. sometime facefix give good right eye, sometimes only good left eye, is easy to combine in photo edit program.

 

[KGARrokh]

More styles.


Prompts for this styles is more diverse, this was my old experiments.

Spoiler: Styles:

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

Spoiler: Links:

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

 

[KGARrokh]

Memes and some fun concepts.

Where is the update Sabia?)

Spoiler: Fun:

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

 

[KGARrokh]

Pin-Ups, clothing and posing concepts.

Spoiler: Pin-ups:

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

 

[KGARrokh]

And finally lewd concepts (mostly light, close to pin-ups; actually maybe some of them are pin-ups).

Spoiler: Lewds:

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

 

[KGARrokh]

Forgot about this "style") Great colors, volume, water and not good eyes, but it mostly PDXL6 then LoRA's achievment).

Spoiler: Cthulhu:

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