CivArchive
    BBC Reverse Squatting (front view) [Real] | Wan 2.2 I2V - v1.0 [high]
    NSFW

    Take it easy 😱

    Okay, I'll go harder 🀀


    Wanna support me?

    Don't send Buzz, send me USDc or USDT through Polygon Network:

    0x3f1e1323fa90b4c9f5a386bfa5b1e72f5a6c25e9

    $0.10 or $10, any amount is welcome and helps me make more LoRAs for the community


    Workflow

    I don't use ComfyUI, so I don't have a workflow to provide πŸ€·πŸ»β€β™€οΈ

    The training is done on Wan 2.2 A14B Diffusers (BF16)


    Training

    • 400 steps on high noise

    • 1 edited video (80 frames)

    • HLR + ZCD (These acronyms were created solely to confuse you)

    • Lots of hope

    • $7 (😭😭😭😭😭😭)


    Usage (Low noise)

    This LoRA was trained only on High Noise, which means you'll have to use some Low Noise that knows what a penis is (I used the Low Noise from DR34ML4Y v2).


    Got an idea for LoRA? Then let's do it! (or at least try 🀠)

    In my repository, I created a model called "LoRA Requests", so if you want to request a LoRA, send videos and photos to help build the dataset for a public LoRA that will be posted here, that's the place to go.

    Open LoRA Requests on HuggingFace πŸ€—


    I2V:

    Prompt (Ongoing):

    The man moves his hips rapidly, making his penis go up and down repeatedly, going in and out of her vagina.

    Prompt (Cut to):

    Maybe not, bro, at least not here πŸ™‚πŸ€šπŸ»

    Visit this LoRA on HuggingFace for uncensored examples, I know you want them 😏


    High Noise LoRA Scale: 1.0

    Low Noise LoRA Scale: 1.0

    Shift: 4


    About HLR + ZCD

    This is a fast-learning technique, which makes LoRA less flexible. This can drastically reduce creativity, but it yields stable results while using few resources. Negative effects:

    1. It will probably only work at the same camera angle

    2. High chance of not responding to different prompts

    3. High chance of forcing the original characters of the training video Training only the High Noise reduces this chance, but it can force things that are VERY similar, like the same hair length, same color, etc.


    Download

    Download High Noise LoRA (On HuggingFace)

    Download Low Noise LoRA (DR34ML4Y_I2V_14B_LOW_V2)


    Help me creating more

    If you want to help me continue making LoRAs, or if you want me to make a LoRA for you, buy 5000 PlayCoins at Aisha-AI and transfer them to my account (account number 2).

    This helps Aisha-AI to stay alive and produce new LoRAs for you all πŸ’œ

    Description

    First version

    FAQ

    Comments (6)

    iluvlamiaMar 4, 2026
    CivitAI

    ltx 2 version?

    AishaAI
    Author
    Mar 4, 2026

    I don't have access to that model, so no, not for now

    jevencheungMar 6, 2026
    CivitAI

    Bro, I've tried many more examples, and the facial consistency of this Lora model, at least in my opinion, isn't as good as the other Lora models you uploaded before. Regarding your suggestion to add character features to the cue words, I've also tried that. Although the clothing, hairstyle, and features are preserved, the facial appearance doesn't look as similar to the original. Anyway, with the other Lora models you uploaded before, I didn't need to do this when using them in the same way, and the facial appearance still remained highly consistent with the original, didn't it?

    jevencheungMar 6, 2026

    Another point is that the "cut to" of this Lora model sometimes doesn't cut smoothly, but rather changes within the same shot through character movement transitions. Of course, the problem I mentioned above doesn't occur every time and can be resolved by drawing cards multiple times, but this situation is noticeable compared to other Lora models you uploaded before.

    AishaAI
    Author
    Mar 6, 2026Β· 4 reactions

    @jevencheungΒ To maintain consistency in appearance, you need to mention the person's face. This is a technique I recently learned and haven't yet applied to the prompts.

    I'll give an example:

    "The scene starts with a woman posing, she has a pretty face. Then the scene cuts to [...]"

    Amidst the details of her after the cut, you mention the woman and her pretty face. This will cause Wan to memorize her face and reuse it after the cut.

    When you say she has a pretty face, Wan understands: "Hmm, so that's a pretty face, ok" and when you mention it again later, it will use the reference it received. With this, I managed to maintain consistency in several LoRAs, not just mine, and it worked every time. If, by chance, it still doesn't obey, if you are using ComfyUI, you can force the weight of these tokens:

    "The scene starts with a woman posing, she has a (pretty face). Then the scene cut to [...] The same woman is naked, her (pretty face:1.2) has expression of pleasure [...]"

    jevencheungMar 6, 2026

    @AishaAIΒ Oh, I see! I've learned something new too! Thank you!