Recommendations:
LoRA Weight: 0.5-0.7
LoCon Weight: 0.6-0.85
CLIP SKIP 0-1
DivineEleganceMix | model
Hi-Res Fix
Additional train info:
Trained on: AbyssOrangeMix3.
Description
A new training approach.
Better outfit drawling.
Better face drawling.
Smaller size.
Now LoRA!
FAQ
Comments (10)
get donked knucklehead
your mother
"Cry some more!" [Mimicking a baby crying] "Weeeeeeeeh!"
that moment AGAIN when you try to get this lora running correctly, you are mad like crazy,, scream like crazy and visit this page again to tell your negative thoughts to the creator of this lora.... and you are totally speechless and relaxed all of a sudden, because you read
L Y C O R I S
and you dumb idiot put it in the lora folder AGAIN....
Lycoris now works in Lora folder too, just update WebUI 1111, or use InvokeAI
i will never understand why creators think that only lowering the weight of a model will give you the perfect result. NOT lowering any weight will get you the best result, BUT you need the correct prompts. and those prompts are "official style" and "style parody". these give you exactly the images on how a model is trained on....
A suitable model for style also affects, I think this is obvious to anyone. And also trigger words, high res fix and etc.
@GameSpy i never used hires-fix as i've seen what it does with eyes....
im hella new to this but im gonna start using those for sure, i normally just mimic enhancement words like i see in others promts i.e "high res, HDR, 8k, 4k, realistic, masterpiece..." if there are more you know of please let me know!
@blueballed05163 just visit my gallery and you will see results that won't let you close your mouth
this is legit very pretty
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.