Kotori Mizuki (Yu-Gi-Oh ZEXAL)

This Yu-Gi-Oh "heroine" (she hardly deserves this title due to only having dueled once against the protagonist after getting brainwashed by a villain as she's known to be a damsel of sorts supporting the protagonist barely doing anything to help him throughout the entire anime, it would've been cooler if she was either a proper duelist or being the first main female character in the history of YGH having Ena instead of Astral) has to be my personal favorite since I used to have a crush on her during my childhood when I saw its anime on TV, I used to take pics of her on TV in order to preseve her beauty and especially her costumes which I adored so much. I was kind of disappointed that she isn't as popular as Mazaki Anzu or Yuzu Hiiragi for example, being one of the more obscure and overlooked ones despite being a main one.

But thankfully now that I finally trained her via LoRA, hopefully many others of you can bring her to popularity and give her a second chance since I was looking forward the most to train her compared to the others I previously trained.

Trained as pivotal training with OneTrainer using its Additional Embeddings feature, where it trains each concepts as its own embedding as to not contaminate the text encoder, therefore each of her costumes are separate embeddings, so its required to also download those and place them all inside the Embeddings folder if you want to generate them (I reccomend creating a Costumes folder inside there and inside it create folders with names of characters I'll train via pivotal training in order to organize each costume embeddings properly and avoid confusion since I may name most of them with simple names if they don't have names of their own such as with gacha characters).

Costume list (their names are used for each embedding, so no need to use additional descriptive tags in prompt at all, although they can still be helpful if a costume doesn't generate properly):

1. heartland_school

2. heartland_swimsuit

3. heartland_gym_uniform (dataset needs improvement or considering to remove it due to having very few images)

4. cheerleader (dataset needs improvement)

5. dress (dataset needs improvement)

6. casual_1

7. casual_2

8. casual_3 (dataset needs improvement)

9. casual_4 (dataset needs improvement)

10. casual_5 (dataset needs improvement)

11. casual_6 (dataset needs improvement)