The spaal/tihospaal style brought to you for Illustrious-based models!

• Great flexibility and details for a faithful and flexible style.

• Works on ALL illustrious-based models, optimal results found on NovaFurryXL_IllustriousV6b but your usuals like StableMond/RealMond, jinkies/zoinks, chucks/sneeds WILL work.

• Trained on a 200 images dataset that was retagged and cleaned for maximum efficiency.

I didn't really like the LoRA that was already available made by someone else because it wasn't very faithful to the style. Now I think with this one I managed to make something both very faithful and not overcooked (and God know I've trained for twice the epochs I should've, so here's epoch 4 because everythign after was burnt to a crisp).

Trigger word : no trigger word needed! The LoRA works flawlessly without a trigger, but if you want you can always include spaal in there.

Recommended weights : (0.7) is the sweet spot between faithfullness and flexibility, but you can push it all the way to full weight (1.0) and it will still deliver very good results.