40 training dataset images were generated using the following base model, sampler settings, and LoRA combination

Base Model: LuminarQMix vPred (V‑Prediction enabled)

→ https://civarchive.com/models/1125276/luminarqmix-vpred-noobaixl-illustrious-xl-merge-model?modelVersionId=1474942

Sampler: Euler

Scheduler: Simple

LoRAs & Weights:

- Style : Konya Karasue – weight 0.4

https://civarchive.com/models/1015964?modelVersionId=1969394

- Style: Blue Archive Flat Color – weight 0.53

https://civarchive.com/models/642656?modelVersionId=955906

- Control LoRA Collection - NAI vpred fix – weight 1.0

https://civarchive.com/models/99619/control-lora-collection?modelVersionId=1965505

- [SDXL] Artist Style: Mika Pikazo – weight 0.3

https://civarchive.com/models/275291?modelVersionId=1573142

- 薄塗り / USNR STYLE V1.0 – weight 0.2

https://civarchive.com/models/176554/usnr-style

- Color Temperature Slider – weight -0.3

https://civarchive.com/models/1093089?modelVersionId=1227713