There is no Last Unicorn Inpainting version?

finally someone who does understand and know the meaning of showcasing true potential of their checkpoint!

New update looks great! <3

Thank you for new model!

is AmateurRealLife more like Amateur photography? I've been stuck with X-plus-plus. Should I change it to AmateurRealLife what improvements to expect?

I know you're busy, but can we expect an inpainting for the newest 'amateur' model?

Hey guys, if anyone is using Y photoreal model can you tell which inpaint model should we use for images generated with Y? amazing work with this model!

The negative is essential if you want to use the amateur model.

For some reason all my outputs with this model are very bad. But all your other models create very perfect results. What am i doing wrong? Pics are blurry and look just bad. Why?

Hi guys :) I'm using epiCPhotoGasm Z-Inpaiting but my stable diffusion diffusers pipeline is not using the right VAE I think (I get desaturated inpainted areas).

A solution is to specify the right VAE in my code, what VAE should I use?

Additional details:

Right now I'm using diffusers python library and as a vae I use vae-ft-mse-840000-ema-pruned.safetensors

If i don't specify any VAE, the library will use the default one specified in diffusers library for stable diffusion and not the baked-in VAE

Here my code:

vae = AutoencoderKL.from_single_file( "./stabilityai/sd-vae-ft-mse-original/vae-ft-mse-840000-ema-pruned.safetensors", torch_dtype=torch.float16, ) pipe = pipe_class.from_single_file( path, load_safety_checker=False, local_files_only=True, torch_dtype=torch.float16, vae=vae, ).to("cuda")

Any idea?

i was using Z-inpainting model , for removing objects changing clothes , also removing clothes and nudifying of existing image but the results like nipples , pu$$y are bad quality , anyone tell me why is that ? or should i download any other model ?

Beautiful checkpoint, one of my favorites for sure! Thank you!

so which is better, last unicorn or amateur?

What is the difference between x and z inpainting models?

tldr: looks like last one model forgot about colors, almost at all.

complete description:

Hi dude, wanna say thank you for all all that models that you are created. It's huge amount of work that you done....but, the last one have a serious problem with colors. I have 3 of your model at the same time in my SD GUI and when I noticed that there is something wrong with the last one iteration I decided to do an experiment and generated a X/Y/Z grid that has 31 sampler columns and 3 of iterations of your model (xPhotoreal, lastUnicorn and amaturereallife). All other settings are equal. I uploaded result to GDrive that anyone can check it:

https://drive.google.com/file/d/1ranFn0CmiWhcj4JJl8CTmCFxxDXEYIiT/view?usp=drive_link

Positive prompt is:

(close distance photo:1.3) of Beautiful female angel,focus on face,rays of light clothes,(hair made of multycolor fog, floating glowing light particles:1.3),(surrealism:1.3),shiny skin,bright multicolor backlight,looking to the side,(celestial messenger-graceful, ethereal being appearing as radiant, glowing apparitions, cosmic void:1.1),melting neon wings,godrays,(glowing neon veins on the girl's body:1.2),(volumetric fog, mist:1.3),8k,high detailed,ultra realistic,UHD,

Negative prompt: