Hi, this is my first Lora and its first version, any feedback and comments are well received.
Trained on NoobAI-XL V-Pred 1.0
Version 1.2 should be an upgrade. "mikasa ackerman" tag makes it look more alike the character. I will try to make a definitive version if I can expand the dataset more.
Both previous versions have the same dataset but different parameters, the 1.1 version isn't necessarily better than 1.0.
I used the following settings:
Sampler: Euler
Strength: 0.7 - 0.9
Guidance/CFG: 4