Tuned to work whether or not feet or a penis are in the initial I2V frame input.
I recommend starting the weight of the high model at 1.0 (low should always be 1.0), and tuning down to 0.75 or 0.5 if the motion is short and stilted. High noise weight with I2V frames that already have feet and penis in them often work better at 0.75 or lower.
Description
FAQ
Comments (18)
same aise main hi penis insert to vagina waala bana do please zoom out wale m
Drugs are bad.
can you make armpits suck
This LoRa dataset lacks training data on the vu lva. Therefore, to expose the vu lva, you must use other LoRa datasets. Currently, WAN DR34ML4Y - All-In-One doesn't work. The low-noise version of WAN POV Miss ionary Insertion works, but it's not perfect. It's recommended to have a first frame image that already shows the vulva.
The action of the Lora works fantastically however I'm running into issues where the penis and feet are rendered in like a haze or a bit artifacty/blurry not at all like the preview shows with it being very crisp and in focus, can anyone offer any advice on how to clean it up, its mainly when the feet and penis enter the shot
You are using both the high and low noise models? That problem sounds like the low noise model wasn't included.
@thegipper yeah using both, does it matter in which order? used both at strength 1.0
@HomeBoxOffice I'm not sure what you mean by order? The Low Noise lora needs to be loaded into the low noise model and the high noise lora needs to be loaded into the high noise model, are you only using one base model?
Finally, someone has made a LORA for long-toed footjobs. You have my gratitude! I can't wait to get started with this.
It doesn't bring up the feet even if the picture have feet, it is just doing handjobs
her the same! doesnt work!
Same here, it does the motion without the penis. Anything to know to force the penis in?
same. Anyone here able to fix it?
Can you share a workflow?
The AI displeases you. You must praise the AI first, then it will gen the precious footjob.
same
FYI, this LoRA effectively helps the model understand what a p*nis is, even without a trigger word.
Details
Files
wan_2.2_i2v_footjob_high_v1.0.safetensors
Mirrors
wan_2.2_i2v_footjob_high_v1.0.safetensors
wan_2.2_i2v_footjob_high_v1.0.safetensors
wan_2.2_i2v_footjob_high_v1.0.safetensors
wan_2.2_i2v_footjob_high_v1.0.safetensors
wan_2.2_i2v_footjob_high_v1.0.safetensors
wan_2.2_i2v_footjob_high_v1.0.safetensors
wan_2.2_i2v_footjob_high_v1.0.safetensors
wan_2.2_i2v_footjob_high_v1.0.safetensors
wan_2.2_i2v_footjob_high_v1.0.safetensors
Footjob_HIGH.safetensors
high.safetensors