Quick Generate
Model Description by Creator
This is kinda an experimental merge with my vibe. It is intended for fashion shoots, film and artistic photography
Explore the generation data under the photo in this post. This will give you a clue on how to use this model and what it is intended for
May conflict with other loras, as it already contains my lora in it! You can try to lower the weight of lora (<7) and CFG scale (5.5)
V1 Included model themes down below. You can use the trigger words below in the prompt:
fashion - modern fashion shoots, including creative ones
film grain - cosy film photography
like a painting - photographs in the style of Renaissance paintings, with a touch of romanticism and femininity. You can add clouds, muted colors, flowers to the prompt
eboy - fashionable guys in black clothes and lots of jewelry, idols
couple of lovers - lovestory cosy photoshoots, especially film photography
psychedelic - can give interesting and creative results in the style of photobashes and photo collages
flowers - photographs with people and flowers
makeup - mostly portrait shots, more modern creative makeup
bodyart and statue - can be mixed to get fantasy characters on the verge of realism
There may be a bias towards an asian type of face.
V1 is not for drawings, nsfw, anime, non-human, landscape prompts. V2 works better but is not specified for these topics
V1 Based on Realistic Vision, V2 - including ChillOutMix model
Not for commercial purposes
Tips:
22-30 sampling steps (more detailed composition - more steps)
Samplers: DPM++ 2M Karras, DPM++ SDE Karras
CFG scale: 5,5(!)-7
VAE: vae-ft-mse-840000-ema-pruned for some generations, but not necessary
Resolution: 512x768(!) + hires fix using 4x_Struzan or 8x_NMKD-Superscale_150000_G upscaler, ~0.4 Denoising strength, 10-15 Hires steps. Or upscale it in img2img
I use Adetailer extension for face and hands, but some generation can be fixed just using Hires fix
