This is a basic model characterized by good realism of human subjects but with less expressive rigidity than fully photorealistic models. Designed for long and complex prompts. It supports high weights, but without exaggerating. It is also able to create NSFW content but without necessarily having to worry about preventing characters from walking around naked, thus allowing you to create discrete content more easily. It is recommended to moderate the use of LoRA: too many can degenerate the generation. This is still an experimental model and although tested for a long time it could give unexpected results.
Description
FAQ
Comments (2)
My new favorite. I can't believe it took me this long to discover your awesome model. Keep up the excellent work
____________________________________
ComfyUI (locally)
token_normalization - length+mean
weight_inerpertation - A1111
steps 30
cfg 6.3 to 6.9
sampler euler_dy_negative*
scheduler sgm_uniform
denoise 6.5 to 6.9
*euler_dy_negative is by far the best sampler (that perhaps, you've never tried)
Thank you, I'm so glad you like my work, and I wanted to let you know how much I appreciate your support. I hope you have a good day.



