A LoRA made with DALL-E scaly furries.
Make your scalies more scaly.
V2 is out, better dataset, compatible with many scales colors.
Strengh : 0.7-1.0
V1 Know issue : The scales color can be random if no color is specified.
Description
FAQ
Comments (8)
dino plz now
While I really like the added scale detail this model provides, it has the strong tendency to add multicolored scales - even more so in the refinement stage - which I'm not able to counter in any way. =/
I know what the problem is. Decrease the strengh of the LoRA to get rid of the multicolored scales.
I recommend 0.7 for strength.
@Anzatiridonia Unfortunately, as far as I can recall, it was happening at lower weights than that even. ~shrug~ I'll give it another shot or two. Maybe try remove it when I do my img2img hi-res/inpaint thing.
@anomedra Have you tried using a specific scale color?
@Anzatiridonia I always specify scale color in positives. Even tried various negative prompts for other coloring. As said, I'll give it another shot to see if there's a way around it.
@anomedra maybe a V2 is needed without the rainbow scales (with larger dataset)
@Anzatiridonia v2 seems to have done the trick. Nice! Thank you.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.