V4
Trained on NewDream-illustrious (personally, it handles loras better compared to other models). Different dataset, but style-wise, the same, or at least what I intended for it to be.
Trained with colored data.
Trigger words:
gr4ph1t3v4, sketch
gr4ph1t3v4, sketch, monochrome, black and white
V3
Trained on the regular illustrious base model that is the civitai default.
Dataset is not the same, but it is similar.
The good range is: 0.5 to 0.7, with and without hires fix
Trigger words: gr4ph1t3v3, sketch
V2
Trained on NTR MIX Xiii
Same-ish dataset as v1, but, the lower the strength goes for this, the more colored it can be, which I think is neat.
The good range is: 0.7 to 1.0
Trigger words: gr4ph1t3, sketch
V1
Just a simple sketch lora for Pony. I wanted a sketch style for pony, but the one that is built into it, it not sufficent for me.
I don't know if it works with nsfw.
Civitai's auto tagger really wanted 'traditional media', so you can probably slap it on and get some interesting results with it too - it's not trained on it, as it caused my previous sketch lora to fail.
USE STRENGTH OF 0.5 to 0.7!!
Description
FAQ
Comments (5)
You should not constrain strength range for your models, just add recommended range in description and preset default strength in model settings. I want to try strength = 1 or higher, but can't because of this feature.
Model is great)
Pretty sure the version description says "usage tips". Strength constraints aren't possible, unless it's the way it's trained, which lies with civitai.
@Fatbuns
You literally constrained strength by yourself in range [0.5; 0.7], because you set "Min Strength" = 0.5 and "Max Strength" = 0.7 in "Edit Version" page) Because of this CivitAI generator truncates all strengths values that are outside range [0.5; 0.7].
@search_bot Ah, didn't know that. I fixed that.
They should fix that, as it's misleading.
Really adore the detail and stability that this has at the recommended weights. Draws gorgeous armor and eyelashes.