Good evening nerds, I have come with a new LORA.
Has been tested for a resolution under 512, all preview images are as well
Excuse me stranger...got any spare change?
Ko-fi: https://ko-fi.com/keyennorman
Yo, it's time for some prerequisites.
1. Weights should be at 0.6-0.7 for the most part
2. Unsure how well it works in realistic models
3. A minimal dataset has been used for this of course not by choice. I enjoy the challenge that comes with doing so and will do more. But do not hesitate to give your boy feedback as it could greatly increase the quality of whatever I decide to put out.
4. Please take a look at ref images, I highly recommend using the tags: Adult/large body type if you do not wish for the elf to have small proportions akin to the Skylanders art style. The ai also has troubled consistently generating Stealth Elf's signature white eyes which is a bit of a let down or a plus for some people.
This has been tested with only one or two models, really hope my experience is similar when it comes to others. You may have to fiddle around for a sweet spot but this LORA is indeed capable of great outputs. Can sometimes be other lora compatible.
I love you, now for the rollerbrawl lora...
Description
FAQ
Comments (4)
this lora does not need any trigger... as you can see in my example images, you can load it as is and get the original "stealth elf".
but i have one question:
was this trained on game screenshots? as you can see in this horrible pixelated image, it mostly looks like a screenshot. that's what made curious.
Mostly from the “Skylanders Academy” Series that was on Netflix.
@RevolutionIsComing ah, okay.. but all in all a very gob job on this lora. it is perfectly trained.
@stapfschuh Awww appreciate it