Cum on Face concept. Tried to make the cum drip slowly. Doesn't include penises.
Will improve cum results on face if you use this LoRA and and combine it with other fellatio or facial LoRAs.
Trained on 136 images.
Use this
A woman with a thick white semen on her face dripping down slowly. She is in a state of pleasure.Then customize to your liking.
Workflow is included in the samples.
If you enjoy these models and find them useful, please leave a thumbs up on both. It really motivates me to keep creating more! Feel free to also share some of the work you generated with them in this page; I’d love to see it.
Enjoy!
Description
FAQ
Comments (27)
Thank you in advance.
Looks amazing, thank you
I know it says t2v, but is it a solid ONLY t2v or can I work with this in I2V as well?
Didn't give it a shot to be honest. But technically it should work just like most of the T2V models, just upload a picture with a face loaded with cum. Technically it should work.
@f1985 It's okay to not know, Just wanted to ask and see. I'll do like you said and just try it. Worse happens, just doesnt work lmao
Testing your examples, they all seem to use the fp8 scaled model. When I test it with full fp16, the results tend to look much faker. Have you done any testing with fp16?
No.
This is art, my friend.
the trigger is a quite a long and specific sentence. Does it only work when writen exactly like that, or can words like "woman" or "pleasure" be changed?
give it a shot, just stick to "dripping down slowly" and I think you should be fine.
@f1985 Thank you, that works. I have another question. When using the LORA, my outputs are in slow-motion (does not happen with other LORAs). Do you have any suggestions as to why that happens?
@civitai_dane mostly has to do with the fact that the training set was all images and not videos, so there was no motion in the dataset to begin with.
I checked your workflow, and your lora's are not the same as the downloaded ones, atleast by name. Yours says wan22\fb\nsfw...etc Tried two different workflows getting interesting results...but not what is depicted here...
That's ok, just change the lora and add it from the path you downloaded the LoRA to. These things happen, but don't break the workflow.
can you share the workflow?
AMAZING WORK!
What did these girls do to deserve such artillery I wonder.
hope you can update your previous loras to wan2.2 i2v
You don't update a LoRA. You train your dataset: again.
I need a workflow for this, examples seems to be empty
works well, thanks
Why some of you creators set up those long sentences as triggers? It just makes complicated to use into custom creations...
I like to describe the concept and make it ready for use, so I can use it by itself some times, something I've personally gained from. You can disregard it and use natural language to describe what you want to see.
@f1985 then the whole phrase is not mandatory? If so, it's OK.
there are no real trigger words or sentences for Wan 2.2. Just describe what you want to see and the Lora will do what it was trained for.
Hi where is the workflow? Can you share more info on settings? cfg, lora strenght, steps etc. Theres no workflow on the files. Thanks!
Details
Files
Cum on Face_high_noise.safetensors
Mirrors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors
Cum on Face_high_noise.safetensors