Kingpin - Life of Crime Style Lora [ILXL/Pony]

Update 13.04.2025 - Epoch 5 for ILXL 1.1 has been released, the purpose of this version is to be as highly compatible with many keywords, scenes and character related loras as possible. It is the current main version of the model.

I highly recommend Ikastrous 8.1 checkpoint, it does wonders with this model.

• Recommended sampler for this is Euler A

• Lora weighting parameters are between 0.7 to 1.3 strength, 0.7 gives really amazing results from my testing at 7/8 CFG Scales, going higher in terms of lora weighting beyond 1.0 will make it look like in-game kingpin screenshots, which was my main focus, but regardless the images with such settings are available in the gallery so feel free to have a look around and try it out yourself!

• I appreciate those who have given it a try and I'd would love to see your images that make use of this model!

________________________________________________________________________________________

UPDATE 29.03.2025 - PonyXL Version has been released for those who use Pony based models, like PonyDiffusion, AutismMix and so on.

• Use it between 1.1 - 1.3 lora weight strength / 5.5-7 CFG scale, Euler A at 31 steps

• It may have issues... I generated the images with both Forge and Fooocus to see how both handle it, images are available in the gallery and I think this is as best as it'll get on Pony. I will provide more images later.

________________________________________________________________________________________

UPDATE: 23-03-2025 - ILLUSTROUS XL 1.1 VERSION RELEASE!

• It's better than 1.0, but I am letting 1.0 stay, while 1.1 release is the main version now

• 1.1 is rebuilt from scratch with a new and carefully picked data that should do wonders, it no longer has issues with burnt text in a left upper corner and you can now use it with 1.0 - 1.2 lora strength! Give it a try!

for ILXL 1.1 - I recommend using Euler A sampler with 21/31 steps, Lora weight: 0.8 to 1.2

I also heavily recommend Ikastrious in use with 1.1 version of the lora

https://civitai.com/models/874216?modelVersionId=1229496

"Yo man, got a buzz? Get back to me if you get a Buzz, huh?"

OLD DESCRIPTION (IMAGES SHOWN HERE ARE OUTDATED, SEE 1.1 FOR UPDATED ONES):

Hello, this is a Lora model based on the looks of a criminally underrated videogame that old-school players ||such as myself|| have played day and night. It should also satisfy the needs for those who want to feel the touch of Quake 2 era engine visuals and overall feel. Or if you want to, go mess around with it!
________________________________________________________________________________
Activation Prompt: ``kingpin``
Recommended weight strength: 1.4, i strongly suggest 1.9/2.0 with this one, please see the image examples i posted on this model page!

Basic recommended prompt: kingpin, realistic
________________________________________________________________________________

Without loras: 1.1-1.4 strength

With loras: 1.4-1.9 strength

Going higher will make it very close to the in-game look (usually starts at around 1.9 and 2.1 weighting scale) while lower will make it look like it's from an art deco styled comic book.

Both look cool either way!
________________________________________________________________________________

Recommended Settings:

Sampling Method:
Euler A (41 steps), I found it to be the best sampler for this Lora.
DPM++ 2M Karras (45 steps),
DPM++ 3M SDE (63 Steps)

________________________________________________________________________________
HiRes Fix (I always use it after txt2image generation in most of my images):

Denoise values ranging from between 0.48 up to 0.51
Upscaler: 4x-UltraSharp

HiRes Steps: 45 should be fine, I use 63.
________________________________________________________________________________
CFG Scale: 4 to 6
_______________________________________________________________________________

I recommend this checkpoint for this model.
Hassaku XL (Illustrious) - v1.3 - Style A by Ikena
https://civitai.com/models/140272
________________________________________________________________________________

Notes from the author:

• Due to a slight carelessness, there will be a very tiny text portion in a left upper corner if you're going to increase the intensity of the Lora. I will fix it when I update the Lora but i'll need to find time to do it but for now, it is great for what it is!

• it Might produce females with 1boy/male focus tags, use 1boy, muscular in the prompting and for negatives put 1girl, female focus, hopefully this should do the trick

• Kingpin: Reloaded was not used in this, thankfully.

Thank you for checking out this model page. Have fun generating and have a great day!

I'll appreciate any image that uses this model, I want to see all kinds of crazy madness that you will come up with!