Giantess XL
This model is created as an easier-to-use version of Giantess-XL LoRA.
Trained on Animagine XL V3.
Please do not repost or distribute this model on other platforms without the author's permission!
Usage Guidelines
Please first follow the guidelines of Animagine XL V3.
Size Tags
building size, mega size, giga size, shrunk size, planet sizeLicense
Description
FAQ
Comments (6)
Excellent concept, I'm trying it out right now. Do you recommend using the Lora along with it or just stand-alone? It does throw a lot of runtime errors when trying to load it, however, and it seems very vram hungry, to the point where I can't run it unless I use --midvram and have to wait a lot per image. Do you think a more optimized version is possible? Thank you for your work.
Just stand-alone since that LoRA is quite difficult to use if you want to get a good result. You need to carefully adjust the prompts and block weights of the LoRA. And this model is aimed to make it easier to use.
The model is based on SDXL, so it consumes more memory than SD 1.5. And in SD 1.5, it's hard to generalize such concept.
! IMPORTANT !
https://github.com/hako-mikan/sd-webui-prevent-artifact
You should install and enable this to prevent artifacts (such as abstract, broken or Lovecraft-ish outcomes). Thanks.
Curious, the images seem to come out a lot more saturated than with animaginexl, is there a reason for this? Even when using just a basic non gts related prompt on both it comes out heavily saturated in comparison.
Can you do a regular Checkpoint model for this too, please?
Can you do one for like Real Life Giantesses as well? Also can you do one for Macro Furries both animated and real life?