模型介紹為中英雙語，中文使用者可以關閉翻譯。
--
PornMaster_SDXL_Illustrious_noob_Pony_t2i_工作流下載(Workflow Download)
PornMaster Simple 4X Magnification Workflow
--
--
Stable Diffusion WebUI Forge - Neo
--
Creating Photorealistic Images With AI: Using Stable Diffusion
SDXL 1.0 藝術研究。(SDXL 1.0 Artistic Studies)
--
v7之前的版本为sd1.5 "Amateur系列"检查站。
Versions prior to v7 were the sd1.5 "Amateur Series" checkpoints.
--
My other checkpoints and Lora.
--
我已將模型上傳至runninghub。
I have uploaded the model to runninghub.
PornMaster-Pro-SDXL-V7-VAE
https://www.runninghub.ai/model/public/1981749531459108865/?inviteCode=rh-v1304
PornMaster-Pro-SDXL-V7-VAE-inpainting
https://www.runninghub.ai/model/public/1981757608040243202/?inviteCode=rh-v1304
PornMaster-Pro-SDXL-V8-VAE
https://www.runninghub.ai/model/public/2008921333920436225/?inviteCode=rh-v1304
PornMaster-Pro-SDXL-V8-inpainting
https://www.runninghub.ai/model/public/2043676887620259841/?inviteCode=rh-v1304
--
AUTOMATIC1111:
Check "Clip skip SDXL" in the settings to enable the Clip function.
勾选设定中的"Clip skip SDXL"，可使Clip功能生效。
ENSD:31337
Eta 噪點種子偏移（ENSD）
(ENSD; does not improve anything, just produces different results for ancestral samplers - only useful for reproducing images)
--
如果你的WEBUI无法显示某些版本的LORA:请到"设定"，然后勾选"Always show all networks on the Lora page(otherwise, those detected as for incompatible version of Stable Diffusion will be hidden)"
--
不要問我如何安裝的問題與comfyui如何使用。
客製化lora:我的Telegram ID:@iamddtla0620
我的微信ID:iamddtla
我的line ID:iamddtla
時間寶貴，請直接說明來意，例如:訂製lora.
Do not ask me questions about how to install it or how to use ComfyUI.
Customized lora: My Telegram ID: @iamddtla0620
My WeChat ID: iamddtla
My Line ID: iamddtla
Time is precious, please state your intention directly, for example: customized LORA.
--
面部局部重绘可使用adetailer.
高解析度放大可使用"ultimate SD upscale".
For partial facial redrawing,use adetailer.
High resolution upscaling can be done using "ultimate SD upscale".
--
Adhere to the Fair AI Public License, ensuring that any derivative works are also open source.
遵守Fair AI Public License，確保任何衍生作品也是開源的。
Description
基於Pro-SDXL-V7制作的inpainting檢查站，功能更加靈活。
如何使用inpainting，我沒有免費教學。
此模型包含VAE，無需使用VAE。
--
The inpainting checkpoint made based on Pro-SDXL-V6-VAE has more flexible functions.
I don't have a free tutorial on how to use inpainting.
This model contains VAE, there is no need to use VAE.
FAQ
Comments (52)
whattt
In general this model is good, high quality but it can't generate two person having sex.. the output is really weird when trying to gen sex, Anyone can give me some tips?
Where has the prompt documentation gone?
The latest SDXL-V6 and V7 versions no longer require "r/ " as prompt words, so I deleted them.
Thank you for your answer. So basically you put the name of the sub in natural language?
@cmoaln1249 yes.
What sampling method and schedule type should be used for v7 inpaint?
Pros: +creative NSFW generations
+universal chars (even tho it still favors asian chars)
+the overall quality is great
Cons: -very prompt heavy even with embeddings
-needs a lot of loras to force the model to be more precise or stop generating weird poses
-not very multiple char friendly a lot of clipping and fusing which refers to my second point loras loras loras...
Do you prompt it with natural language with sentences or like il/pony/noob?
@delta45424155 On this model specifically i mix it.
I agree!!, this model is great but for solo character on the screen.. I still can't find a model that properly do 2 person having sex in several poses without lora 😭😭
@delta45424155 @bielzimdag346 natural language. I don’t understand why some do pony tags with this. That would be like you going up to an artist and speaking tags. They would be like “what the fuck are you saying?” until you say enough random words that they just go with it. I’ve made full body shots of 2 people fine since my 1st prompt as an amateur with no LoRAs.
This is the prompt template that I use with a LLM. I hope it help:
Photorealistic full-body shot, captured on a Canon EOS R1 with an RF 50 mm f/1.2 L USM lens at f/1.8. [Subject and Action]. The woman has [face shape], clear [eye color], [hair color], and flawless skin. The man has [face shape], clear [eye color], [hair color], with subtle body hair. Setting: [Location and Lighting]. Details: [Specific actions, fluids, and expressions].
@freestuffpl0x42069 Thanks I will test it, but the person who trained the model use tags on the example.. I guess that would be the best way to prompt haha
What is best resolution for inpaint?
我想用这个模型训练lora，能告诉我你用的是什么低模吗？我用SD-XL1.0-BASE可以训练人物lora来用在这个模型吗
直接在PRO-V7訓練就好。
@iamddtla 我电脑3070卡跑起老火啊 平时都在平台训练，这个网站我不晓得咋的 我就是充值不起
I am getting this error trying to use this - WanVideoKsampler
Expected 3D (unbatched) or 4D (batched) input to conv2d, but got input of size: [1, 4, 41, 64, 64]. Can anyone help?
This is not Wan Video!
This is the checkpoint where words become pictures.
Waiting on V8 :P saw you launched Realism-IL V4 instead. SDXL coming soon?
I have been testing for a long time, and I can't yet make Pro-V8 better than Pro-V7.
@iamddtla Same issue iv run into with my own checkpoints 🤣 months of trial and error and its never better.
@iamddtla Oh btw I just saw you are from Taiwan. I live in Taitung XD My Taiwanese friends enjoy the content I make with your models haha
My goodness. V7 is insane.
Is there a proper term or prompt to generate wispy bangs... or those bangs that doesn't cover the entire forehead?
Finding it difficult to generate that hairstyle popular in Japanese gravure. :(
Blunt bangs
After using this, I earned a master's degree in pornography. Thanks.
im new how to install that plz
i copy it on chekpoint folder but the program cant find and see it
Why isnt it available anymore?
already did thanks
how or where do you go to learn to turn your images into short video?
same, i downloaded it but the instruction are very unclear for beginners
We using I2V video model for this, try wan 2.2 there are many version of it. We generate picture and the we feed that picture into Wan 2.2 Ai model and we prompt something and done. (Check out wan 2.2 nsfw if you want to use this checkpoint)
@karollapsa165 thank you
V8 when? 🥹 V7 is amazing
I am currently unable to create a version better than v7.
Hello, can anyone help me with the ksampler settings for the SDXL v7 inpaiting?
what should I set the:
steps
sampler
scheduler
cfg
denoise.
thank you :)
I had same problem, i tried 30 steps, dpmpp2m, karras, cfg7, denoise 1 for Text2Img, or from 0.3 to 0.7, but anyway it can only give around 40% good pictrue, if you want good ones, you need a lot of detailers and lora
There’s nothing like a meeting, totally mind-blowing...
wow good job~~
is this inpainting v7 better than lustify inpainting
IMO both give good results and in my testing it was generally very similar, wondering if anyone has specifics
Hello. Could you give me tips on how to create images involving anal sex? It's been extremely difficult to generate anal images because the model has a very strong bias toward vaginal sex. For me, it's almost impossible to make the model understand that a couple is having anal sex and not vaginal.
Use anal sex Lora.
@iamddtla
could you please take this in consideration in future checkpoints? the model is heavily biased into conventional sex positions. another example is face sitting wich is hard to create.
contratulation for you work by the way.