Update: Comfyui Workflow has been officially released!
This model is an inpaint model for realistic images.
-This model allows you to naturally modify parts of the generated images by prompting.Improve the quality of your inpainting with the rich knowledge of pony models and the flexibility of prompts to meet the needs of more creative purposes.
-V4 version
V4 was created by adding NoobAI to merge list and then remerging the illustrious models to restore realism.Non-remerging model also uploaded as ALT version.
The difference between the two is the internal ratio of NoobAI. The normal version is slightly better in skin tones, etc. In most cases, there's not much difference, but in N-situations or complex compositions, the ALT version will produce better output.
How to use:
Sampler
V3 and V4 = DPM++ 2M SDE (karras for comfyui) or DPM++ 2S a
V1 and V2 = Euler a
Soft inpainting check
Mask bluer 4~8
CFG Scale 4~7
Denoising strength 0.9~1
steps 20~50
-Soft inpainting settingchange "Transition contrast boost" to 1 (default 4)change "Difference threshold" to 0.1~0.25 (default 0.5)
Don't use "Apply color correction to img2img results to match original colors".
Refiner:
Using normal model as refiner (eg. pony realism) at 0.65~0.95 (recommend 0.8) can produce even more great results. I will pick out models that can be used as a refiner, so please try it out.
Good Refiner model list
Real Wives Pony v_1.1 [believe me, best skin texture ever. (need neg fat)]
https://civarchive.com/models/709934?modelVersionId=794078PornMaster-Pro Realism-IL-V3 [prompt variety, danbooru tag]
https://civarchive.com/models/1045588?modelVersionId=2099739RedCraft Illustrious3 Relustion [skin tone, prompt variety, liquid]
https://civarchive.com/models/958009?modelVersionId=1517097ULTRA v9 [soft, dark atmosphire]
https://civarchive.com/models/228525?modelVersionId=2098608Ponyvinculum Exotic_Mix [clear, soft, lighting, animal, nature]
https://civarchive.com/models/1424028?modelVersionId=1809768
Tips:
Only masked is best for quality
Start with Masked content "original", if not worked change to "latent noise" or "fill".
Use image size with height + width>2000~2500 (Recommended size for SDXL).
or use resized by scale around recommended size (aspect-ratio-helper).
If you mask a small part, it is a good idea to mask the top and bottom of the main mask so that the model can see wide area
Your image works as quality prompt, You don't need a lot of prompts, keep them to a minimum and let the model do work.
dog, in beach-Use this model only for AI-generated images.
-Prohibits uses that harm any person by using this model on real-world images.
Description
Improved skin contrast, added noise for reality.
Improved prompt reacting.
FAQ
Comments (37)
It's really good! Please make more pony inpainting models!
thanks, I will upload few more models.
If you tell me what you want, i may be able to reflect it.
Looks interesting. What model did you merge with the inpainting model to create this?
@habibing are you going to make a v3??
One of the pony(inpainting) models which works for real, perfect job! Honestly impressed by the output quality, so sharp and so seamless.
Thank you, The mask boundary was the next most important part after not ruining pony model.
Bravo, honestly, the way your model understands surroundings and the environment is stunning! It really mimics the noise, color temperature of the base image.
Very impressed with this, like wow! Although I did have an issue.
I tried to use this in forge with the “fooocus inpaint patch” controlnet. my output was just noise. Is there something about this model that causes that problem? Other SDXL and Pony checkpoints work fine.
It's Good.
Just follow the owner guide, and the rest is just simply experimenting.
If you using inpaint upload like me, make sure to create proper masking. White for what you want to change, and black is for the stuff you want to keep. And careful with the mask blur value, because you need to adjust your masking image and that value.
Also have this as additional reference is helpful, at least for my case -> https://civitai.com/articles/3990/inpaint-sdxl-for-the-manipulation-of-your-images
Forgot to ask should i use sdxl vae or vae is inside model?
sorry to late answer,sdxl vae fix is included
How many steps, do you prefer to use? I always get noisy or blurry output
The optimum step is 20-30, other possible causes include Check resolution, adjust height + width to be 2000, set sampler eura a, this is important SDE is not currently supported. denoising strength to 0.9, CFG scale 4~5 is good
I do everything as indicated in the instructions and I get a soapy image, with SD 1.5 there are no such problems, what am I doing wrong? vae I use either sdxl or remove it altogether, I can’t figure it out. I tried both through Forge and Comfi
Plz try webui AUTOMATIC1111.
It looks a little blurry and low-resolution, so if you could improve that, it would be perfect!
If Eura a doesn't solve the problem, try lowering the CFG.
The clarity of the V1 version is higher than that of the V2 version, but the rendering of clothing gloss is worse. An excellent model!!! Looking forward to the author's future updates!
Thank you bro, If there is a next update I think it will be either Illustrious alone or a mix with pony.
Does it work in comfyui and if there is any workflow? Thank you!
Sorry, I've never use comfyui. I think you'll need a workflow that replicates the inpainting of webui to make it work.
It's awesome,thank u so much!
Thank you. I have some ideas for updates, but I was working in cloud platform for model, so I will develop the updates when I subscribe next time.
I still think this is the best inpainting model out there
whle I have so far genuinely njoyed this model I have however run into a problem. Most photorealistic models tend to have some trouble with shadows but this one REALLY struggles. any shade darker than lit by a 40w bulb and it instantly pancakes into flat, washed out grays. what is worse is that any lora or prompt I try to use to remedy it only exacerbates the problem, especially with the bacckground.
any tips on working around this, because I really want to love this model.
Is there a good way to reproduce the problem?
habibing oh it happens all the time. any time I prompt for dark, shadows, low light, low-key or night time. all the colors get washed out and any blacks or darks immediately turn grey. i still haven't found a way to work around this.
I tried but couldn't reproduce it. Is it possible that the noise settings in the webui have been changed from the default? If you could send me some row pics where the problem occurred via chat, I might be able to find out what's causing it.
Hey can you share how made this model ?
What merge components is this made of ?
Thank you a lot
v2 base model are
getphatPonyReality_v40.safetensors
fennfotoPONY_v4.safetensors
indecent_v10.safetensors
and other model is
sd_xl_base_1.0_fp16_vae.safetensors
sd_xl_base_1.0_inpainting_0.1.safetensors
ponyDiffusionV6XL_v6StartWithThisOne.safetensors
Amazing. Best inpainter i’ve tried so far. Thank you.
I am able to release V3 because of people like you, so please give it a try.