CivArchive
    shantaeAI (Style embedding) - 1.0
    NSFW
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined

    This is a style embedding inspired by the artwork of the Shantae games. It's nothing major, but if you like the sometimes very amusing / funny style: this is for you!

    Also check out my other embeddings here.

    Trained on waifu-diffusion-v1-3-float32.ckpt, but works best when used in connection with DreamShaper 3.3 or Anything 3.0.

    I used Anything-V3.0 VAE or vae-ft-mse-840000-ema-pruned. Both work fine.

    The keyword is the filename (without the .pt) you put into your embeddings folder (usually shantae_styleAI).

    Description

    Initial release.

    Dataset images: 13
    Model used: waifu-diffusion-v1-3-float32.ckpt
    Number of vectors: 4
    Batch size: 13
    Gradient steps: 1
    Learning rate: 0.005:800, 0.002:1300, 0.0005
    Steps: 1100
    Latent sampling method: once

    FAQ

    TextualInversion
    SD 1.5

    Details

    Downloads
    1,347
    Platform
    CivitAI
    Platform Status
    Available
    Created
    1/29/2023
    Updated
    4/24/2026
    Deleted
    -
    Trigger Words:
    shantae_styleAI