Jason's Mixed Media XL brings the look of my mixed media pieces (collage, layered paint and found textures) into SDXL.
Trained on 45 photos of my own original artwork, captioned with my own titles for each piece. No AI images in the dataset.
How to use: put jmixed_style first in your prompt and start at 0.7. It carries over to new subjects; try it on a lighthouse, a cat or a ship.
Training: Juggernaut XL v9 base, kohya sd-scripts, LoRA dim 32 / alpha 16, UNet only, 768 px buckets, AdamW8bit, cosine, LR 1e-4, 6 epochs trained, epoch 4 released (452 steps).
Showcase: Juggernaut XL v9, Euler a, Karras, 24 steps, CFG 4.5, LoRA 0.7 with trigger, hires 1.5x.
The art is mine; the pictures made with this LoRA are AI. https://www.deviantart.com/signalofsolarum · https://signalofsolarum.substack.com
Description
First release. Trained on 45 photos of my own mixed media pieces, captioned with my own titles. UNet-only, epoch 4. Use the trigger, start at 0.7.