These Base Models were used for testing the Models:
Illustrious:
CAT - Citron Anime Treasure:
https://civarchive.com/models/131986/cat-citron-anime-treasure-illustrious-and-sdxl-and-sd15
Prefect illustrious XL:
https://civarchive.com/models/1224788/prefect-illustrious-xl?modelVersionId=1379960
Pony:
Sym-Pony World:
https://civarchive.com/models/565749/sym-pony-world
AniMics - Pony XL:
https://civarchive.com/models/631811/animics-pony-xl
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
Trigger:
LaNell
Appearance:
purple hair, side braid, blue green eyes
Original wardrobe:
dark grey and white maid uniform, beret, dark grey beret, white feather, pink hairbands, glasses, thin-framed glasses, white front panel, black tie, puffed short sleeves, black armbands, white detached sleeves, brown shoulder straps, belt, brown belt, square buckle, silver buckle, Apron, white apron, dress, dark grey dress, high side slits, dark grey thigh highs boots,
Settings:
LoRa Strength: 0.80 - 1.00
ClipSkip: 2
Steps: 20 - 30
CFG: ~7
Sampler: Euler, Euler_A or DPM++ 2M /w Scheduler Karras
for Full-Body-Images i recommend a FaceDetailer and/or an Upscaler to get better Faceresults
Description
Training:
some images from the first training were recreate with img2img and the results replaced the previous for more flexibilty in style
increased amount of repeats
changed Tag "blue green eyes" to "aqua eyes" (works sometimes better together with "green eyes, blue eyes")