NOTE: V2 is currently processing and looks like it will fix the problems with medium-wide shots with the current previews I have gotten at the half way point. It should also be better at different resolutions and canvas shapes/poses.
This Lora helps create consistent images of Prof. Augustine Sycamore. While it is trained for realisitc Illustrious checkpoints it can do anime styles with lower weights or good prompts.
Recommended Strength: (0.5 -1) 0.7 is the most consistent across checkpoints.
Pros: hair texture, face consistency, realism
Cons: Long Shots, faces at long range
The sample images all used the same prompt, generation settings, and strength. This prompt was not optimized for any checkpoint and is pretty general. Tweaking prompt will yield better results.
Only an eye Adetailer was used on these images.
This is my first Lora and any tips and tricks are appreciated. I hope you enjoy it!