CivArchive
    Skinny preggo - v1.0
    NSFW
    Preview 22030616
    Preview 22030620
    Preview 22030613
    Preview 22030617
    Preview 22030614
    Preview 22030626
    Preview 22030624
    Preview 22030625
    Preview 22030630
    Preview 22030629
    Preview 22030627
    Preview 22030628
    Preview 22030631
    Preview 22030632
    Preview 22030633
    Preview 22030635
    Preview 22030634
    Preview 22030648
    Preview 22031072
    Preview 22031066

    This embedding creates images of pregnant bellies of thin and young girls.

    Description

    FAQ

    Comments (7)

    yessy_boemJul 30, 2024· 1 reaction
    CivitAI

    safetensor please? : )

    siurspaces750544
    Author
    Jul 31, 2024· 2 reactions

    Done :-)

    yessy_boemAug 2, 2024· 1 reaction

    @siurspaces750544 Thaaaanks!

    Bit_ShifterJul 30, 2024· 2 reactions
    CivitAI

    I appreciate the effort, but this is easily prompted on most photoreal models by default. "slim woman, skinny, petite, pregnant" in the prompt and "(fat, obese:1.4)" in the negative.

    Prompting, instead of using embeddings, can potentially decrease token use. Doing so increases prompt adherence, realism, and generation accuracy on most photoreal models.

    However, something I do like about this embedding is that navels look significantly more realistic. That's rare and certainly makes it worth using!

    One other suggestion, this needs a .safetensors conversion. Pickletensor (.pt) is an insecure file type that can allow for embedding malicious code.

    siurspaces750544
    Author
    Jul 30, 2024· 1 reaction

    Thanks for the detailed review!

    It was because of the navels that I started this business.

    Please tell me how to convert .pt in .safensors?

    Bit_ShifterJul 31, 2024· 1 reaction

    The only script I've found for converting is on HuggingFace's GitHub page...

    https://github.com/huggingface/safetensors/blob/main/bindings/python/convert.py

    One other way is to create embeddings directly in the .safetensors format using OneTrainer. I've found it easier to use than Kohya or the trainer in A1111 and generally produces better results once you get the hang of it. It has far more customization and tools than either while keeping a very clean, organized UI. It's all I use now for training... https://github.com/Nerogar/OneTrainer

    siurspaces750544
    Author
    Jul 31, 2024· 1 reaction

    @Bit_Shifter thanks, bro :-)

    TextualInversion
    SD 1.5

    Details

    Downloads
    111
    Platform
    CivitAI
    Platform Status
    Available
    Created
    7/30/2024
    Updated
    9/4/2026
    Deleted
    -
    Trigger Words:
    Skinny_preggo

    Files

    Skinny_preggo.pt

    Mirrors

    HuggingFace (1 mirrors)
    CivitAI (1 mirrors)