Original link: https://sololo.xyz/model/614-scragglymessy-SoloLoRA

The results from the ZIB version were a bit unexpected. It actually reproduces the source better than the ZIT version, but I still prefer how the ZIT one looks overall. That’s probably due to my overly aggressive training settings on ZIB.

When this version is used on ZIT, the LoRA and base model kind of balance each other out. The result is different from both the ZIB output and the earlier ZIT version. It’s not bad though, so I decided to release it as well.

Many of the sample images were generated with ZIT.

You can start the prompt with “A sketch of”.

Strength for ZIB: 0.7–1.0

Strength for ZIT: 1.0–1.5

Trigger Words: ScragglyMessy