version 003:
This version has undergone some fine-tuning, enhancing the expressiveness of pictures.
version 002:
Some fine-tuning from version 001,mproved some clarity.
The suggestion for use are still the same as in version 001.
This is SD1.5 model to produce some realistic images.
The LCM model produces images much faster.
First of all, thank you very much for the work of the LCM research team, so that I can produce high quality images at such a fast speed, just 8 step.
The ADetailer must be used because the face has a high probability of breaking in a full-body or half-body photo.
I hope you like to use it.
The biggest support for me is posting more pictures, :)
Suggestion (LCM Series):
Sampling method: LCM
Sampling steps: 8~12
CFG Scale: 1.5~2
If you want better image whit SOTG series, my suggestion is,
Sampling method: DPM++ 3M SDE Karras
Sampling steps: 40~60
CFG Scale: 7~12
Attention:
Please don't do anything illegal with this model. Thanks!!!
Description
FAQ
Comments (3)
your checkpoint is very specific on very small details. i didn't even need to prompt it and got those super tiny details very exaggerated detailed. check my upload to see what i mean...
EDIT: another very positive plus: prompting anything body-based (bombshell-body in my case) does generate what i expect and not naked characters the whole time. i love that. knowing that this checkpoint doesn't take everything insanely serious, i can continue my journey on creating prompts without being too careful every time.
My suggestion would be to use an LCM sampler to get a sharper picture!
@Hustmox thanks for your suggestion, but i really am just a prompt guy. i'm not interested in anything else.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.