I received a report that the LoRA I trained is compatible with the illustrious model. Use of the illustrious model is recommended.

Test Report:

Overall, the body proportions, hair shape, and facial contours are well trained. However, the further it deviates from the recommended settings, the more severe the eye distortion becomes. When the CFG value is set above 7, the position and ratio of the eyes and mouth tend to become misaligned.

For best results, we highly recommend using this in combination with the Kagami Hirotaka Style LoRA.

Recommended Settings:

Sampler: DPM++2M Karras

Steps: 30

Resolution: 768x1024

CFG: 5

LoRA weight: 1 or (0.6 and Kagami Hirotaka Style LoRA:0.4)

The sample images are generated with a LoRA weight of 1

Core Tags : Lo_uratsubaki, black hair, long hair, high ponytail, hip-length hair, long side locks, purple ribbon, hair ribbon, long dangling ribbon, green eyes, large breasts

School uniform: school uniform, white shirt, sleeveless vest, plaid skirt, red shirt ribbon, black pantyhose, loafers

Archery club: hakama, white top, black chest protector, tabi socks

Gym uniform: sports uniform, white shirt, short sleeves, v-neck, bloomers, knee socks, black socks, sneakers

Swimsuit: swimsuit, black swimsuit, green trim, two-tone, barefoot

Cheerleader: cheerleader uniform, long-sleeved shirt, cropped top, pink top, white shirt, pleated skirt, miniskirt, pink skirt, striped skirt, white socks, ankle socks, pink sneakers

Casual wear: hooded jacket, striped jacket, white and orange, high-top sneakers, white laces, miniskirt, blue skirt

Bunny girl: bunny suit, black bodysuit, bunny ears, bow tie, cuffs, fishnet stockings, high heels, sexy outfit

Training Base model : animeillustriousxl

Training resolution : 768x1024

Network Rank : 32

Network Alpha : 16

epoch 8

Dataset size : 37