It meant to make a detailed canine penis. Trained with highly detailed, 8k texture 3d renders and digital hand drawings.
Best strength values (go higher values if you want more realistic)
v0.1b: 0.6-0.8 | v1.0: 0.6-1.0 | v2.0: 0.6-1.2 | v3.0-3.1: 0.6-2.0 | v4.0: 0.6-1.3
For some unfortunate reason, some models will not work no matter what you do. Most of the times is because it don't understand the concept 'k9 dick', to test if it will work is simple, if you can't make the model generate k9 dick by itself, my lora won't work anyways on the model.
You can use the recommended trigger words alone or you can use it together with some extra tags below.
Negatives: holding penis, hands on penis, tucked penis, testicles.
Due to how the balls look in 3d renders "testicles" tag need to be in negatives to avoid the lora turn balls into a entra weird knot. Instead I recommend you to use "balls", "big balls", "furred balls" instead of "testicles" in the positives.
Other helpful tags down below if you want more control (v1.0-2.0 IL) \/
(Don't put together tags separetad by "/"!!! Don't mix "ventre" side with "dorso" side!!! it will conflicts and may cause glitched image)
big knot / massive knot / small knot
thick penis shaft / thin penis shaft
erect / half erect / less erect
penis ventral view / penis dorsal view
low-angle view
Description
Initial version.
It still in progress to make it better, but for now it works very good.
Using "Chroma-DC-2K-LoRA" together helps increace realism if you need it.
FAQ
Comments (10)
Is there a method for preserving shape quality in I2V? I find knots in particular try to melt away.
Not yet. I would need to make a lora for that or a whole model ans i can't, my hardware can't handle it.
@The_Doggy_Enjoyer a) so a wan lora or something like that can do it? b) I have a vague idea of how t2i loras can be trained, how does one train a i2v or t2v lora?
@yamavishnu892 a) possibly yes even tho they are more for movement than aesthetics b) you need videos to train it, that's why it's mostly hard in this case
@The_Doggy_Enjoyer How many videos, and what hardware do you need to train it?
@Lazman how many i'm not sure, maybe 20 at minimum. and hardware maybe 16GB of vram to start
@The_Doggy_Enjoyer I got the Vram, but the videos, bit more limited there..
@The_Doggy_Enjoyer you need high quality close ups?
@wedew3237 that will not help to make video lora, it need video and the motion on the videos that will be learned
This model is peak
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.