Happy #Super Sunday!
I post a Super (Hero/Villain) every Sunday
Update v2 - New version is trained on Animefull-final-pruned-fp16.
Training with Anime means its more stable and flexible but loses a bit of the original art style.
For example, in v1 I used a weight of 0.5 to get flexibility but in v2 the lora is flexible at weight 0.8. So details can be maintained while still being flexible.
This is an extreme example of the difference between AnyLora and Anime training because the dataset had a very unique art style.
Please checkout all my Super Hero/Villain Loras: https://civitai.com/user/CitronLegacy/models?tag=super
v1 Details
Weight: 0.5
To get the original art style use Weight: 0.7-1.0
Trigger: fire_lit_DC, green glow
Example prompt: fire_lit_DC, 1girl, green glow, upper body, long hair, smug, holding green fireball, magic