CivArchive
    Preview 89579440
    Preview 89579529
    Preview 89580097
    Preview 89580130
    Preview 89580180
    Preview 89580260
    Preview 89584074
    Preview 89584122
    Preview 89584182
    Preview 89584477
    Preview 89585460
    Preview 89585501
    Preview 89585574
    Preview 89594360
    Preview 89594434
    Preview 89594503

    Work in progress. This is initial drop of 'Vivi Luvva'. This will be another series of loras similar to 'That Bitch!' and 'Ima Luvva'. Currently at the text embed stage. Other loras will be added as they are trained/created. This is mainly for character preview. I hope you like her... she has, long dark brown hair, pale fair porcelain skin, large brown eyes, long eyelashes, slight rouge, slim hips, long legs, perky medium-small breasts, 23 years old, "vivi luvva" from korea.

    I'll upload loras for sdxl, flux, hunyuan video and wan2.1 video. Right now just creating a training dataset.

    Notes:

    sdxl-base version strengths 0.8-0.9, any lower loses asian eyes.

    wan2.2 lora tested on wan2.2_t2v_*_noise_14B_fp8_scaled models w/ Wan2.2-Lightning_T2V-A14B-4steps-lora_*_fp16 loras and wan2.2-t2v-rapid-aio model. lora strength set to 1.0.

    hunyuan lora tested with hunyuan t2v and framepack-lora ready.

    Description

    initial lora trained against sdxl-base-1.0. 53 images 1024x1024. 90 epochs 5300 steps. strength 0.7 - 0.9. less than 0.7 model loses asian eyes.

    LORA
    SDXL 1.0

    Details

    Downloads
    147
    Platform
    CivitAI
    Platform Status
    Available
    Created
    7/21/2025
    Updated
    4/19/2026
    Deleted
    -
    Trigger Words:
    v1v1l2vv4

    Files

    vivi-luvva-training-dataset-v1.0.zip

    Mirrors

    v1v1l2vv4-sdxl-base-v1.0.safetensors

    Mirrors