My first attempt to create a photorealistic SDXL-Model.
I hope, you like it.
I merged it on base of the default SD-XL model with several different models. (Around 40 merges)
SD-XL VAE is embedded.
I strongly recommend ADetailer.
You should set "CFG Scale" to something around 4-5 to get the most realistic results.
v1.2 is an updated version with some added elements with a slight tendency to be a miss of the dark. 馃槉
v1.3 has new improved details, better nsfw results, I also refined the output for better photo realism, the missus is still here 馃槈
v1.4 has better results for NSFW images, I also refined some details, less missus but improvements in base model 馃槉
v1.5 is better fine tuned, improved details, refined output
v1.6 is an improved model with better details and better NSFW compatibility.
A music video, I made using this model and Udio and wav2.lip and KlingAi and much more is here:
Description
Improved details, refined model
FAQ
Comments (8)
This model seems to get less realistic with each iteration. You should really quit doing the same images with the same seed each time for example images. And Euler A is one of the worst samplers if you want realism. The woman on the motorcycle is starting to look like a mutant. You should at least randomize the seed and do a few attempts until you get a good one.
It's personal opinion, which version is "Better" for you. Just because it's a new version, it doesn't mean, this is obligatory to be better for you.
It depends on your own like and what you want to achieve.
I of course do sample images with same settings and seed so you can quickly change between the versions and see, which matches your preferences.
@Devilmarkus聽Every update done to a model makes it a different model. Sticking to the same seed all the time doesn't accurately represent what the new model can do. Using the same prompt is one thing, but reusing a seed that worked well in the past on a different model but gets a crappy result on the new model, just makes the new model look bad, when in fact it might be better. If you keep the prompt and randomize the seed, you can generate a few options and get the one that works better for this model.
I really hope you will make a Flux version of this model !
I just moved on sdxl 2 week ago. I tried at least 80-100 checkpoint, and finally it seems i found the best.
I was looking realistic and EMOTIONAL capable checkpoint. Interesting fact, most realistic checkpoint cant do any emotions, besides smile, sad. With this is can do naughty, slutty, seductive look.
Thanx
I noticed an interesting failure in this Lora, it absolutely refuses to do a male penis :-) Men come out dickless no matter what I do.
I'm using this model; the generation is fine up to 98% then at the last the image degrades terribly and the result is terrible. What could it depend on? I hope someone can help me because I really like the model
If the faces are degraded, check face restoration.
If the whole scene is degraded, perhaps check your upscaler or the sampler you are using.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.