Example images are made with a variety of checkpoints and prompts. Nothing unusual.

• No trigger word

• Model only (no text/CLiP)

• Adjust strength to your own taste and what's in your chain

• Works better with photorealistic checkpoints

Recommend checkpoints

Juggernaut | Analog Madness | Photon | EpicRealism | CineDiffusion

Note

Did lots of experiments after v11 and found some settings that are much more reliable. It is more accurate and has better generalization. I will most likely create an updated version in the near future after doing further refinement trainings on the current weights and some merges.