ShadowMix

I started combining a few models and as I added more, came up with this one. It has a lot of models I've really enjoyed such as revAnimated and cartoonish. So far, I've liked what I've created through it and I hope you do too. Is it perfect, well.....no. Hands are still a gamble, sometimes they come out great, othertimes, no. It's created some shockingly good and bad renders....it all depends on the prompts and Loras used. Works well with some and not with others so you'll have to experiment. It has a tendency to smooth out surfaces and/or objects that should be more distinct but it still creates some pretty amazing pictures. I haven't played around with it extensively so put your creative prompts out there and see what it does. I've tried simple prompts and detailed ones, both have created some great renders.

Here are some examples, some pulled from other creative geniuses:

candid photo of a captivating, alluring woman with long hair and in vintage attire, reminiscent of a bygone era, elegantly poised at a brutalist train station, waiting for transportation.
blurry, dawn, realistic lighting,

Negative prompt: cartoon, painting, illustration, (worst quality, low quality, normal quality:2), grayscale

Or:

(masterpiece, top quality, best quality, official art, beautiful and aesthetic:1.2), ultra realistic, 32k, extremely detailed CG unity 8k wallpaper, detailed background, 1girl, full body, <lora:Tingyun_v16:0.8>, tingyun, green eyes, ponytail, hair ribbon large white collar, (small fox ears:1.1), seductive smile, growling, waiting impatiently for the viewer to hurry up, (random waiting poses:1.4), (nude:1.5), medium-large breasts, thigh gap, cameltoe, cleft of venus, ass, looking at viewer, <lora:hair_length_slider_v1:1.8> <lora:CutePussy:0.6> <lora:Cameltoe:0.4> <lora:GoodHands-vanilla:1>

Negative prompt: (worst quality, bad quality, low quality:2), monochrome, zombie,overexposure, watermark,text,bad anatomy,bad hand,extra hands,extra fingers,too (bad-artist-anime), bad-artist, bad hand, extra legs ,(ng_deepnegative_v1_75t)

Or:

1girl, young girl sitting on a park bench outside in the rain, best quality, ultra-detailed, intricate, detailed face, detailed eyes, beautiful eyes, (complex lighting), cinematic lightning, solo, perfect skin, (blue eyes, round pupils), (medium-large breasts, cleavage:1.4), wet clothes, (strapless white see through dress, transparent dress:1.1), bracelets, jewelery, necklace, earrings, lakeside restaurant, lamps, (dynamic pose:1.2), (very long hair, random hairstyle), wet hair, looking at viewer, night, rain, soft lighting, glowing lights, blurry background, crowd, <lora:GodPussy1 v3:0.3> <lora:Cameltoe:0.4>  <lyco:ShinyOiledSkin_v20-LyCORIS:0.4> <lora:pokies_9:0.5> <lora:transparent_dress:0.8> 

Negative prompt:  EasyNegative,worst quality, low quality, badhandv4,bad anatomy, text, logo,

I usually use DPM++2M Karras for the sampler. Steps between 15-25. Hires Upscaler R-ESRGAN 4x+ on with a Denoising strength anywhere between 0.25-0.45. Clip Skip at 2 (1 also works depending on what you want). CFG scale 5.5-10. Feel free to play around and see what you get though.

For faces, I HIGHLY recommend adetailer. I will usually put something like "detailed face, long eyelashes, blush, lipstick, perfecteyes eyes, <lora:perfecteyes:1>"

If you've enjoyed this checkpoint, please leave a comment, a render you've created with it, and/or favorite it. Thanks!