Trained with 242 images on the AAM AnyloraAnimeMix checkpoint. The samples were generated with AAM and Yuzu.
Recommended settings:
Sampler: DPM++ 2M Karras / Euler a - 35 steps
CFG: 7
Resolution: 512*512, 768*768, 512*768, 768*512
Lora weight: 0.65-0.8
HighRes: 4x_NMKD-Siax_200k, upscale by 2x, 25 steps, 0.35 denoising
Clip skip: 2
Basic prompt: noela, 1girl, white hair, long hair, blue eyes, wolf ear, wolf tail, hair flower, puffy sleeves, bloomers, navel
Costumes:
Ballroom dress: noela, pink ballroomdress, 1girl, blush, smile, open mouth, ahoge, hair bun, white hair, wolf ear, tail, choker, wolf girl
Swimsuit: noela, 1girl, white hair, twintails, long hair, blue eyes, wolf ears, wolf tail, wolf girl, pink frilled swimsuit, one-piece swimsuit, hair flower
Gym uniform: noela, gym uniform, 1girl, white hair, ponytail, long hair, wolf ears, tail, blue eyes, white t-shirt, blue shorts
Notes:
- I used the fastnegativeV2 embedding for the samples