Have been experimenting with merges for a long time now and have come to settle on this merge the past little while. Not exactly sure what's in there as I've made so many merges but there's probably a little bit of everything and so I would just like to thank and give credit to the entire Civitai community for this hodge podge of a model.
All previews are generated without the use of Loras, embeddings, controlnet, hires fix, face restore, etc.. I wanted to showcase the barebones model itself without any aid. a 3060 gpu and Forge were used to generate the preview samples. I typically use a CFG between 3.5 - 7.0 with eular a or dpm++2m sde samplers and haven't tested others very much yet.
I mainly use this model for semi-real and photoreal generations. Can generate both sfw and nsfw.
Enjoy!
Description
FAQ
Comments (6)
My new favorite model!
what best object i can generate using this model 'to look real' and same question again to look cg?
I'm not sure I follow. Are you asking for the best prompts to use for realism and CGI or what the best subject to generate is? ether way I'm not sure what is objectively best, the model has a bias toward realism but is influenced heavily by the prompt like any other model. Best thing is to just experiment and see what works best for what you're trying to generate.
what the best subject to generate using you model ? sorry i meant subject but i wrote object !
, not bout 'how to prompt' , just what generation can be best in your opinion , so far i found each tuned model is good at something , versatile and all around model is just a lie . that was i'm sure of .
if you don't mind then what is your model is best at as you think ?
@amazingbeauty It's pretty versatile and can generate a variety of subjects. it is good at doing photoreal, semi-real, and stylistic pictures, both sfw and nsfw. It is a pony model so although I don't personally use the scoring prompts often in my samples it is still responsive to pony style prompting. I find its strengths are humanoid characters, realism and prompt adherence.
As someone who's new to stable diffusion, I really appreciate the simple prompts in the samples. Nice to see you can get great results without overly complex prompts. Out of the dozen or so models I've tried this is my favourite so far.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.










