Simply makes the "Cleft of Venus" visible through articles of clothing. AKA Camel Toe or Crotch Cleavage. Works equally well from both a frontal and rear view. Plays well when used with another Lora. Two of the showcase clips were made in combination with another lora. Feel free to offer constructive feedback and any clips using the lora are appreciated.
Description
FAQ
Comments (7)
anyone can help with this lora?
what workflow do you guys use?
I'm happy to help. All of the videos in the showcase were made with this workflow: https://civitai.com/models/1092466/hunyuan-2step-t2v-and-upscale Which diffusion model are you using? Can you send me your workflow?
@voyeur19 thanks so much! i will give a try!! i have a 4060 ti 16 gb, do you think i will be able??
@pretsbr It should work on that graphics card. I haven't tried it with ComfyUI, I use Framepack-Eichi, which does a much better job of handling VRAM, plus you can create up to two minute clips. Works great.
I am extremely happy with this LORA, although I would highly recommend NOT using a strength of 1.0 - even 0.8 tends to produce nudity. I am having great success with this LORA at 0.6
I'm glad you are happy with it. Interesting. I haven't noticed that issue with the LoRA. I went back and checked my dataset. 12% did have bare breasts.
It may be happening due to stacking LORAs. None in and of themselves are nude LORAs, but I am sure they were probably trained with some nude images. Might just be a weird result that happens when stacking LORAs, but as I said, it works great if I lower the strength. Thanks again!