Going to give you a little hint my friend. When you ZIP up the file, the HASH Code that CivitAI records for the model is of the zip file. not the model file. So even if people create 100000 images using a model from the zip file, none of them are associated with this assets.

It is because the HASH Code of the Model that A1111 or even ComfyUI pick up is of the model file. So that hash is saved with the image. When that image gets uploaded to CivitAI it reads the Hashcode, and can't find it on the site. All because you ZIP up the models instead of uploading them as safetensors.

suggestions: hair pulling, curl with fingers, comb, shaving... and other actions :)

i fucking love pubic hair

By the way guys, for Wan2.2 this is the best pussy lora so far.

Severely degrade image quality

Just to help some new folks:

Light4-step high - 0.9 -> LoRA+CLIP Loader (LoRA 1.5+CLIP 0.5)
Light4-step low - 0.7 -> LoRA+CLIP Loader (LoRA 0.5+CLIP 1.5)

Connect your +&- Prompt's CLIP to the Input of your High LoRA+CLIP. Then connect its output to the Low LoRA+CLIP.

If you want faster movements per second, increase your Steps to 1.5x your FPS.

For testing: 320x320 @ 12fps, 20 Steps 0-10 / 10-20, length 33. (2 secs)

That should be your basic Wan2.2 'test bed' for nearly anything.

i was not looking for "some hair", i am trying to recreate the forest moon of endor. we are a far away from pony LMAO

Did you cut out the female faces in your dataset? Seems like it's affecting my female faces quite a lot turning it on... Regardless, this is the only pubic hair WAN LORA that exists (wtf), so I'm grateful for it still!

Anybody can tell me how we use this lora ? I put a woman without hairy sex and start frame and after 5 secondes, no hairy sex !!