ghost
Sign In

Schmanzy SDXL - Generate Art For Colouring

411
3.2k
5.9k
88
Verified:
SafeTensor
Type
LoRA
Stats
3,157
5,900
Reviews
Published
Aug 3, 2023
Base Model
SDXL 1.0
Training
Steps: 2,600
Epochs: 1
Trigger Words
orn8
orn8 design
orn8 lineart design
orn8 design of detailed
schmanzy
Hash
AutoV2
6925C027B9
default creator card background decoration
Rasali's Avatar
Rasali

Hello everyone! This model is for making ornate and intricate line art that you could colour. The dataset is inspired by adult colouring books that have a tendency to be detailed in a way to make you want to colour in all the littles spaces. The dataset is made up of images generated with SD1.5.

Triggering the style: orn8, orn8 design, orn8 lineart design, schmanzy

If you are having trouble getting the LoRA to generate black and white lineart, I recommend using "photo", or "photograph" in the negative prompt. Sometimes adding "blurry" to the negative prompt can help make the lines crisper.

I've only made images with this LoRA in ComfyUI using SDXL 1.0. You might get some really cool images in the newer checkpoints from the community. Overall the LoRA seems pretty flexible and can make interesting images with simple and complex prompts.