Another user requested lora. This one produces nice puffy innie pussies for your enjoyment.
Trigger word is: "1nn13". It was also trained with phrases like:
a nude woman with a shaved innie pussyfocus is on her shaved puffy vaginaThis lora may need to be recooked as something just seems a little off with it. Too many loras stacked with it and you'll start to see weird motion glitches. If it gives you too much trouble, try lowering the weight. I use it at a default of 0.8.
Use with the Hunyuan Goonmaker workflow to achieve similar results: https://civarchive.com/models/1315959?modelVersionId=1485540
Description
V1 trained on 71 data set images
FAQ
Comments (24)
Beautiful =D
ah damn finally...
Been waiting for someone doing this lora for 2 months.
thank you
This is actually really good for getting "problem characters" to be nude generally.
I got those weird motion glitches when stacking it with just one lora.
Just a heads up I wanted to check your workflow settings but nothing happens when I drag the video to comfy even though the workflow is in the metadata.
Thanks. Mind me asking which lora it was?
My images don't have workflow metadata in them unfortunately. Sorry.
@KinkMaster Sure, it was my own lora at 1.0 strength, double blocks. I've set yours to 0.5 and now it seems to work. Could be mine is overtrained too though.
Your example videos do contain workflow metadata. You can see it when you open it with something like Mediainfo. Here's an excerpt from the first video.
comment: {"prompt": "{\"1\": {\"inputs\": {\"unet_name\": \"hunyuan_video_720_cfgdistill_fp8_e4m3fn.safetensors\", \"weight_dtype\": \"default\"}, \"class_type\": \"UNETLoader\", \"_meta\": {\"title\": \"Load Diffusion Model\"}}, \"2\": {\"inputs\": {\"lora_name\": \"hunyuan\\\\test\\\\plastered_epoch6.safetensors\", \"strength\": 0.6, \"blocks_type\": \"double_blocks\", \"model\": [\"1\", 0]}, \"class_type\": \"HunyuanVideoLoraLoader\", \"_meta\": {\"title\": \"Hunyuan Video LoRA Loader\"}, \"is_changed\": [\"hunyuan\\\\test\\\\plastered_epoch6.safetensors_0.6_double_blocks\"]}, \"5\": {\"inputs\": {\"shift\": 7.0, \"model\": [\"6\", 0]}, \"class_type\": \"ModelSamplingSD3\", \"_meta\":,
@funscripter627 oh wow, very useful to know on the metadata! Thanks! I'm honestly a bit surprised there is any metadata at all as many of my main lora showcase pieces get a finetune pass through topaz prior to posting. that would explain why the comfy workflow check doesn't work.
@KinkMaster Posting without meta data is rude and antisocial. Be a better man!
You're doing god's work! Thanky you! <3
cmon not the god work
OKKKKKK..... you and I are apparently cut from the same cloth when it comes to likes!!!!!!!
TIP TIME BABY!!!!
Haha. Great tastes, for sure!
hell yeah brother!
AT LAST , Thank you
I can't stop creating clips. I have no idea why, but the Lora pushes the style in a very special direction.
Do keep sharing and tagging the resource in the clips if you can. The last one you made was fantastic!
@KinkMaster I always add the main Lora I use to support the creators and maybe to make others aware of the Lora.
Some times this lora give me a penis....
Can't imagine any possible way that originates from this lora. This was trained purely on innie pussies with no data (image, video, or otherwise) containing a penis nor any captioning mentioning a penis in any way. If it is generating with a penis, it must be tied to the prompting and hunyuan base in some way.
@KinkMaster its not a true penus, but just long part of body
Please port/train on WAN!
Prompts / workflows please. Thank you!
Check my other posts and comments. Many others have already asked. Also check my posted workflow. It's called the goonmaker. All the settings are there.
@KinkMaster Thank you!