CivArchive
    Corneo's Runny Makeup/Mascara Embedding - 1.0.0
    NSFW
    Preview 40011
    Preview 40018
    Preview 40017
    Preview 40016
    Preview 40015
    Preview 40014
    Preview 40013
    Preview 40012

    This is a textual inversion embedding to add makeup/mascara running from the eyes, as often seen in fellatio/irrumatio images, but can be used elsewhere too. Use corneo_runny_makeup to activate it.

    It can come out pretty heavy by default (i.e. a LOT of black mascara running down the face quite thick) so you may want to tone it down by de-emphasising the tag such as (corneo_runny_makeup:0.9). Also as with some of my other embeddings, the eyes themselves sometimes need some work in image gens that use it, so add some other tags to prompt how you want the eyes to look, and if that fails, try inpainting.

    Example images produced with OrangeCocoa available here https://civarchive.com/models/4355/orangecocoa-5050-mix

    Check out my other civitai models for more characters and sex positions! https://civarchive.com/user/Master_Corneo

    If you enjoy this embed, I genuinely love seeing your creations with it - you can attach them to a review if you're not aware (but not comments sadly)!

    Description

    FAQ

    Comments (6)

    LadyNuggetsJan 26, 2023
    CivitAI

    hi! i´m kinda new, where i save the file? i´m using A1111 Webui

    Master_Corneo
    Author
    Jan 26, 2023· 1 reaction

    In the embeddings folder, then use the filename (without extension) as part of your prompt

    LadyNuggetsJan 27, 2023

    @Master_Corneo Thank you! You are awesome! ♥ Where can I learn how to do this? I'm fascinated! Do you have any kind of guide from where you have learned? Thank you very much anyway !

    Master_Corneo
    Author
    Jan 27, 2023· 1 reaction
    HaiFeb 7, 2023
    CivitAI

    Can you explain a bit about how you train your embeddings? I've been trying to train an embedding on the grapefruit model, but my image outputs quickly deteriorates to noisy garbage. I have 19 images, and my learning rate is 0.005:1000, 0.0005. Also tried more aggressive learning rates, but the loss seems to get to 0.2 immediately, so I dialed it back.

    I've flipped the images, and generated danbooru txt files for all of them, and gone through each one, trimming out nonsense tags.

    Can you leave a few tips on what might be wrong? Need more files? I've seen people train amazing things on just 5 files, so I don't get it.😅

    pizazaMar 9, 2023· 5 reactions
    CivitAI

    can I ask for a safetensor version please?