LoRA of Fine Motion (umamusume)

NOT COMPATIBLE WITH AOM3A2, AOM3A3 model.

Recommend options

• LoRA weight 0.7~0.9

• Trigger words

  • base

    • fine motion \(umamusume\)

  • Racing competition clothes :

    • white gloves, white footwear, black skirt, white jacket, green shirt, long sleeves, white coat, open clothes, collared shirt, socks, high heel boots

  • Wedding costume :

    • official alternate costume, tiara, jewelry, wedding dress, necklace, veil, green flower, bracelet, white dress

  • Tracen school uniform :

    • tracen school uniform, purple shirt, white skirt, puffy short sleeves, sailor collar, bowtie, pleated skirt, sailor shirt, miniskirt, zettai ryouiki, white thighhighs

  • Casual clothes:

    • green skirt, belt, casual, white shirt

    • (optional) noodles, holding chopsticks, food, ramen, bowl, table

Use animefull based model

• Settings

  • DPM++ Series(SDE Karras, 2M Karras, etc.)

  • about 20 steps, CFG scale 3.5~6.5

  • kl-f8-anime2.vae

  • CLIP skip = 1

  • If using highres

    • Latent series

    • Denoising strength 0.50~0.65

LoRA training info

• trained on sd-scripts by kohya_ss. Thank you a lot!

• based on Animefull-pruned

• 112 images of fine motion

  • 4x low quality 35 images

  • 8x medium quality 37 images

  • 16x high quality 40 images

  • => 1 epoch = 1076 images

• replaced character feature tags with fine motion \(umamusume\)

  • horse girl, horse tail, brown hair, green eyes, etc.

• using aspect ratio bucketing

• resolution : 768x768

• rank=alpha=16

• using dadaptation

• batch size = 4, 6 epochs trained

• clip skip = 1

This LoRA may be compatible with animefull-based models (ex. AbyssOrangeMix Series...)

This LoRA can generate nsfw images of fine motion.

All uploaded images are generated by AOM3A1