InCase style (LoRA) for Anything v3 compatible models and different versions. After several failed attempts, this is my first try on training a LoRA model. I have used the same images set than for my hypernetwork model : https://civarchive.com/models/5124/. I suggest to use LoRA weight between 0.6 and 0.7 to obtain a good rendering.
If you like this model, please add a small review or a sample of your work to share with the others all the possibility of this model.
Description
Trigger : <lora:ics_a3_lora:0.7>
FAQ
Comments (6)
Again, what are the differences betwene your models? I notice one is of your current uploads are still the original that I downloaded a while back but what abotu the rest? Different sizes so they must be different somehow.
hello, is there any additional settings to set up ? like clip skipping for example ?
Hi!. First off, I love your style loras, embeddings!
I would like to know if you trained with nudity or without and if you did, actual porn art or just topless for example?
This is amazing but is there any chance you can run a few other incase characters through it? it seems to have issues / wants to argue ALOT about making pointed ears for example.
What checkpoint should I use for these??? Thank you!
@GtsuyaStudio would you be willing to upload your dataset to Gdrive or something for this Lora (with or without captions, doesn't matter to me.)
I want to try experimenting with training an Incase based model but finding images has been a pain.
Thanks for reading / considering.
Love your stuff.
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.