Boring with flat washed skin on your image? Just try this embedding!
Use it at the end of your prompt. Weight it for different result.
Attached image show how it work in different type of checkpoint(realistic-anime-toon).
Also some image are made with KKW IDEAL to show how it work on my model.
Show your result and impression about it for make it better in future version.
Description
V1.0 initial release
FAQ
Comments (10)
I can't get it to work for the life of me. It's properly loading during startup and shows up in my embeddings, but it does not do anything in all my standard models, until I crank up the weight so high it burns the entire image...
Did you use upscale or hi.res fix?
@devilkkw Neither
@traxx2012 are you in a1111? this is really strange issue, what sampler? what model? tell me more detail on how reproduce issue, without it i'm unable to check.
@devilkkw I'm in A1111, using DPM++ 2M Karras, I've tried with ICantBelieveItsNotPhotography, revAnimated and Lyriel, RestoreFaces is on and I'm working with 40 Steps on 512x786
@traxx2012 i guess restore face is problem, it tend to washing face. let me do some test, i work at 896x1152, maybe low resolution affect result. Just for see, maybe try it with no restore face and low step.
@devilkkw that was indeed the problem. Good thinking!
I can't get this to work. It just gives me this error message:
''loading network C:\Stable Diffusion\stable-diffusion-webui\models\Lora\kkw-skin-det.pt: ValueError01:45, 2.51s/it] Traceback (most recent call last): File "C:\Stable Diffusion\stable-diffusion-webui\extensions-builtin\Lora\networks.py", line 223, in load_networks net = load_network(name, network_on_disk) File "C:\Stable Diffusion\stable-diffusion-webui\extensions-builtin\Lora\networks.py", line 148, in load_network key_network_without_network_parts, network_part = key_network.split(".", 1) ValueError: not enough values to unpack (expected 2, got 1)''
This is not a lora. You have to put in embedding folder and call it by the name of file. See sample image.
@devilkkw my bad 😇
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.

















