NOW ON FANVUE, CREATING FULL LENGTH ADULT CONTENT WITH SOME OF YOU'RE FAVORITES: https://www.fanvue.com/zikickix
UPDATE: V3 adds some requested fixes including furry and better anatomy on feet. It's not perfect, but it's a work in progress. Also added some more wide angle shots, images with intense blacks, and added more hand/anatomy photos to the dataset.
This uses my trained SinGlo lora/checkpoints from Illustrious mixed in with Anima Base!
Works very well with natural language.
YOU NEED:
qwen base encoder & qwen image vae
All gallery images should have metadata so you can drag and drop to see the exact things you need along with prompts and settings.
SETUP:
Steps: 30 - 50
CGF: 4 - 5
Sampler: Eular/Eular A (haven't tested others)
Description
Kept the original dataset and tags while adding more images that focused on anatomy, movement/action, and a few scenery images. Now trained on the Base model.
FAQ
Comments (4)
Two small tests with prompts I've already used, and I must say that—even though I already liked the previous version—I consider the style to be a very, very good addition to any setup :)
Thank you! I def wanna make this and all my other checkpoints the best it can be. If there's any prompts you come across that are failing or need work please let me know.
@Zikickix
I love testing checkpoints and models in their pure form without any LoRAs, and since I test a lot of them, I always take notes to remember which one to use at a certain scenario. I’m happy to share them with you, but please keep in mind that I'm just a beginner, so they might not be super deep. They’re really just my personal notes!
So after testing 200-220 gens...
Prompt following (9/10): It tends to follow longer prompts well, but occasionally adds random elements after a certain length (200+ tokens).
Tattoos (10/10): Excellent rendition of tattoos.
Lace stockings (9/10): Handles this small detail very well, whereas other models usually struggle with it.
Feet (5/10): Hit or miss, especially when combined with any form of legwear.
Anime characters (1/10): Poor performance, but to be fair, the model is not meant for this.
Furry (6/10): It seems to have a large fan base for this, though it tends to add hair randomly.
Styles (10/10): There is a distinct style baked into the model—you either love it or leave it. Personally, I love it.
Lighting (7/10): Great for darker setups with vibrant lighting, which I'm a fan of, though it occasionally lacks depth in the shadows.
@_Fischi_ This is perfect thank you. This will help this checkpoint and my others. I have zero feet in the dataset along side furry. I also know that camera angles need to be worked on. Thanks again.