Sign In

Maruzensky (umamusume)

675
5.0k
16
Verified:
SafeTensor
Type
LoRA
Stats
5,006
Reviews
Uploaded
Feb 22, 2023
Base Model
SD 1.5
Training
Steps: 1,146
Epochs: 6
Trigger Words
maruzensky \(umamusume\)
Training Images
Download
Hash
AutoV2
4A3E92D990
0
0
0
0

LoRA of Maruzensky (umamusume). Put LoRA file in your stable-diffusion-webui/models/lora folder and write LoRA notation in prompt to applying LoRA. <lora:your_loha_name:weight>

All of my works is free. If you'd like to support me, feel free to buy me a coffee. 👍

Recommend options

  • LoRA weight 0.55~0.75

  • Trigger words

    • base

      • maruzensky \(umamusume\)

    • Racing competition clothes :

      • red thighhighs, red jacket, pleated skirt, red skirt, long sleeves, thigh strap, choker, red shirt, open jacket, white bow, miniskirt, open clothes, high heel boots

    • Casual clothes :

      • sleeveless dress, white dress, straw hat, floral print, flower

    • Summer costume:

      • swimsuit, official alternate costume, heart-shaped eyewear, eyewear on head, sunglasses, black bikini, cleavage, navel, jewelry, flower, frilled bikini, necklace, thigh strap, ponytail, bracelet, collarbone, sandals

Use animefull based model

  • Settings

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

    • about 20 steps, CFG scale 3.5~6.0

    • kl-f8-anime2.vae

    • CLIP skip = 1

    • If using highres

      • Latent series

      • Denoising strength 0.50~0.65

LoRA training info

  • based on Animefull-pruned

  • 80 images of maruzensky

    • 4x low quality 29 images

    • 8x medium quality 21 images

    • 16x high quality 30 images

    • => 1 epoch = 764 images

  • replaced character feature tags with maruzensky \(umamusume\)

    • horse girl, horse tail, brown hair, aqua 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...)

All uploaded images are generated by AOM3A1