CivArchive
    Yatsu Murasaki v1 - V1
    NSFW
    Preview 1
    Preview 2
    Preview 3
    Preview 4
    Preview 5
    Preview 6
    Preview 7

    Yatsu Murasaki (Taimanin) - LoRA for SD 1.5

    This is a LoRA for Yatsu Murasaki from the Taimanin series. It has been trained to capture her distinct features, including her signature purple hair, blunt bangs, and red eyes, while maintaining high flexibility for different outfits and poses.

    Key Features:

    • High Fidelity: Accurately reproduces her facial features and hair style.

    • Outfit Versatility: While it recognizes her classic Taimanin look, it is flexible enough to handle casual clothes, office suits, or fantasy dresses.

    • Dataset Variety: Trained on a curated dataset including various angles, expressions (from smiles to shouting), and lighting conditions.

    Recommendations for best results:

    • Trigger Words: yatsu murasaki, taimanin (series)

    • LoRA Weight: 0.6 - 0.85 (Start at 0.7 for the best balance between character likeness and prompt adherence).

    • Sampling Method: DPM++ 2M Karras or Euler a.

    • CFG Scale: 7 - 9

    • Negative Prompt: Use standard negatives like (worst quality, low quality:1.4), blurry, distorted, extra limbs, bad anatomy, monochrome.

    Character Tags:

    To get the most accurate look, use: yatsu murasaki, purple hair, long hair, blunt bangs, red eyes, large breasts


    Tips for Civitai Creators:

    • Showcase Images: Upload a mix of her "classic" look and some creative prompts (like the Office Lady or Cyberpunk styles) to show the LoRA's flexibility.

    • Base Model: Mention which base model you used for training (e.g., AnyLoRA, Pony, or standard SD 1.5) so users know where it performs best.

    Description

    LORA
    SD 1.5

    Details

    Downloads
    2
    Platform
    SeaArt
    Platform Status
    Available
    Created
    4/5/2026
    Updated
    4/5/2026
    Deleted
    -

    Files

    Available On (1 platform)

    Same model published on other platforms. May have additional downloads or version variants.