CivArchive
    EyesPonyXL - v1.0
    NSFW
    Preview 41050480
    Preview 41050484
    Preview 41050481
    Preview 41050486
    Preview 41050487
    Preview 41050485
    Preview 41050483

    Worked on getting high detailed anatomy shots to build an overall body LoRA. These are the parts of the whole.

    Description

    "detailed eyes, even eyes, eyelashes, moist eyes, sharp eyelids, sharp eyes, sharp iris, sharp pupil" will get you results, use them all or only some. Go Crazy, it's AI!

    FAQ

    Comments (5)

    thatdidnttakelongNov 20, 2024· 2 reactions
    CivitAI

    Keep up the good work! Eyes are def an issue with many realistic/semirealistic generations. Excited to see your progress!

    kyvonpayden596Nov 20, 2024
    CivitAI

    I wouldnt touch eyes with lora in Pony or SDXL. They alter the character look. You can prompt high detailed eyes, and everything, or use hires.

    inpantspro
    Author
    Nov 20, 2024· 3 reactions

    If you change the eyes you just need to be aware of hair color preferences and other small details. I use a character LoRA for my personal image generation so having detail LoRA to help guide the results won't change their appearance much even at strength 1. This was also trained only on eyes and the space in which our eyes sit, not full faces, which helps counter some of those other unwanted changes.

    kyvonpayden596Nov 20, 2024· 4 reactions

    @inpantspro Ok, i tried it. I had bad memories about eyes and detail loras from SD1.5 era.

    This is fuckin good, not specially for eyes, but everything. Its a skin detailer, realism slider, body parts detailer together. Doesnt change face, only minimal. Hah good work

    inpantspro
    Author
    Nov 20, 2024· 4 reactions

    @kyvonpayden596 Appreciate that you checked it out, and that it helped you achieve better results on your generations. I've been trying endlessly to come at Pony from unexplored angles. I think we all need to remember that AI already has a good idea of what is in an image, so when it comes to training we need to tell it only what is of value in an image and not every other little detail. I know what an apple looks like, but if I need to know that it is red, I need to focus on the color not that it is an apple.

    LORA
    Pony

    Details

    Downloads
    4,780
    Platform
    CivitAI
    Platform Status
    Available
    Created
    11/20/2024
    Updated
    5/13/2026
    Deleted
    -

    Files

    EyesPonyXL-14.safetensors

    Mirrors

    CivitAI (1 mirrors)
    TensorFiles (1 mirrors)

    Available On (1 platform)

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