This LoRA was trained on artwork by the amazing artist thebigslick.

The LoRA averaged out the style a bit (so you get something between painterly and flat color), but the main thing is it's super detailed.

Also, the artist had a pretty clear focus on big dicks and veiny ones at that, hehe.

Dataset size: 167 images (tried to pick only high-quality and relatively recent ones).

The LoRA seems pretty stable so you can use it at 0.8–1 strength.