CivArchive
    Footjob_Position - v2.0
    NSFW
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined

    This is a textual inversion embedding for footjob position image generations,and you can use kinds of models for generation.the example image is not very well,but you can do better than me! :)

    AMD is not easy,but I have tried my best.

    the file should named footjob_position.pt

    AnyHentai or AbyssOrangeMix2 - Hardcore is recommended.

    VAE you can use kl-f8-anime2,or others.

    https://civarchive.com/models/5706/anyhentai

    use (Footjob_Position) to activate,need bracket!

    There is still many problem,like extra pennis,mosaic processing,wired position...

    here is the new footjob model with lora:https://civarchive.com/models/27060/footjobpose

    Description

    Improved the accuracy of footjob posture.

    Added more postures of footjob.

    The position of male and the body looks normal.

    Reduced the probability of insert(still exist)

    Standing footjob is exist,but the trigger rate is very low,you can add standing,girl standing,stand position to trigger it,I really hope you can try it out.

    Double girl footjob is exist too,it's still awful.

    AbyssOrangeMix,AnyHntai and Grapefruit (hentai model) highly recommended!

    If you like my work,you can support me on patreon,this is an appreciation and encouragement to me,I will spend more free time to make more models.

    FAQ

    Comments (3)

    VKTralalaMar 6, 2023
    CivitAI

    How did you trained this? I'm trying to train different poses with textual inversion but I cant get results, only can train characters or styles. Do you use captions for the training images?

    eroAlucard
    Author
    Mar 7, 2023

    Actually I haven't,just use lots of pictures to train it,but I want to give up TI,and use dreambooth lora,https://civitai.com/models/11842/paizuri-different-angles-and-types,you can see this lora model,it's the best way to train

    VKTralalaMar 7, 2023

    @eroAlucard I know about Lora, unfortunately I have a Mac, and currently is not possible train Lora en Mac. Anyway I will try with no captions, thanks