Sign In

Narita Taishin (umamusume)

Download

1 variant available

fp16 SafeTensor

Half precision, best balance • 18.12 MB

Verified:

Type

LoRA

Stats

3,693

Reviews

Published

Feb 16, 2023

Base Model

SD 1.5

Hash

AutoV2
7CB610B1B7

Trigger Words

narita taishin \(umamusume\)
default creator card background decoration
Donator Badge
mht's Avatar

mht

LoRA of Narita Taishin (umamusume)

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

Recommend options

  • LoRA weight 0.7~1.0

  • Trigger words

    • base

      • narita taishin \(umamusume\)

      • (optional)closed mouth, serious

        • These tags make her kawaii.

    • Racing competition clothes :

      • torn jeans, (asymmetrical clothes:1.1), belt, fur-trimmed jacket, jacket around waist, clothes around waist, midriff, long sleeves, tied shirt, open clothes, yellow shirt, mismatched footwear, asymmetrical footwear

        • asymmetrical tags might be need higher weight. (It is challenging for AI)

    • Casual clothes :

      • black headwear, white shirt, red jacket, long sleeves, open clothes, torn pants, denim, blue pants

    • Steampunk costume :

      • white shirt, long sleeves, center frills, frilled shirt, brown belt, corset, monocle, semi-rimless eyewear, fingerless gloves, official alternate costume, steampunk

  • 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 (bicubic antialiased)

      • Denoising strength 0.55~0.7

LoRA training info

  • based on Animefull-pruned

  • 108 images of narita taishin

    • 8x low quality 32 images

    • 16x medium quality 32 images

    • 32x high quality 44 images

    • => 1 epoch = 2176 images

  • replaced character feature tags with narita taishin \(umamusume\)

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

  • using aspect ratio bucketing

  • resolution : 768x768

  • rank=alpha=16

  • using dadaptation

  • batch size = 4, 9 epochs trained

  • clip skip = 1

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

All uploaded images are output of AbyssOrangeMix2-nsfw, kl-f8-anime2.vae