This Lora can significantly improve the stability of generating sexy movements videos for characters.
It's very easy to use and doesn't require any specific trigger words. Just write down the action you want。
Please refer to the review below for a comparison video between using and not using Lora
for example:
stroking the neck with both handsgrabbing the breastsstroking the buttocksboth hands to touching from head to breasttouching the breasts from hips up and grabbingbreasts bouncingThis Lora is trained on i2v_14B_480P and can run on i2v, but has not been tested on t2v.
Have fun~~!!
Description
FAQ
Comments (15)
only complaint is that it makes everything slow motion
Indeed, this is due to the WAN itself. You can try to increase the total number of frames generated, such as 81 frames, 129 frames, and speed up from 16 frames to 24 or 30 frames in the later stage.
try in prompt 15fps. Fast motion.
And Neg. prompt Slowmotion
Well, slow motion is more likely to be stable, so that might be the reason why it works. Maybe lowering the LoRA weight could also help?
I can't load this lora
What's a good weight to have for anime/cartoons?
0.8~1
The LoRA does not show up in the used resources list anymore when posting a WAN video, I believe due to CivitAI changing their WAN filter categories a few days ago.
I think so, so I modified the classification of this Lora yesterday
What's a good strength range for i2v?
0.8-1
Can we please get at least the prompts used in your sample videos?
I added prompt words on the example image
请问这个lora是挂载到高噪声还是低噪声模型的后面?
这是wan2.1的lora,在wan2.2上效果不明显
Details
Files
P003-Sexy-Dance-i2v-v10-000010_converted.safetensors
Mirrors
P003-Sexy-Dance-i2v-v10-000010_converted.safetensors
P003-Sexy-Dance-i2v-v10-000010_converted.safetensors
145_P003-Sexy-Dance.safetensors
P003-Sexy-Dance-i2v-v10-000010_converted.safetensors
149_P003-Sexy-Dance.safetensors
P003-SD-i2v-v10-000010_converted.safetensors
P003-Sexy-Dance-i2v-v10-000010_converted.safetensors