PornMaster_SDXL_Illustrious_noob_Pony_t2i_工作流下載(Workflow Download)
PornMaster Simple 4X Magnification Workflow
PornMaster Z-image-turbo Controlnet: Cosplay - Anime to Realistic Style
PornMaster [Flux.2 Klein 9B] - Anime to Realism Workflow
--
Stable Diffusion WebUI Forge - Neo
--
模型介绍为中英双语，中文用户可以关闭翻译。
此版本可以在写实风格与动漫风格之间切换。
可使用写实风格或动漫风格的lora。
可使用sdxl和Illustrious的lora。也可以使用部分的"pony"lora。
--
This version can switch between realistic style and anime style.
This version can use realistic or anime style lora.
Can use sdxl and lora of Illustrious.This version can also use parts of the "pony" lora.
---
我已將模型上傳至runninghub。
I have uploaded the model to runninghub.
PornMaster-Pro-noob-V6-VAE:
https://www.runninghub.ai/model/public/2004812200420274178/?inviteCode=rh-v1304
PornMaster-Realism-IL-V4-VAE:
https://www.runninghub.ai/model/public/2043980764240420866/?inviteCode=rh-v1304
--
如果在我的检查站训练某些lora效果不好时，可以在” noobaiXLNAIXL_epsilonPred11Version”或” illustriousXL_v01训练lora，这样lora在我的检查站依然可用。
If the effect of training some lora in my checkpoint is not good, you can train lora in "noobaiXLNAIXL_epsilonPred11Version" or “illustriousXL_v01", so that lora is still available in my checkpoint.
--
Sliders lora for color temperature, brightness, saturation, and contrast.
--
Add details and improve the accuracy of prompt words lora.
--
--
NoobAI-XL ControlNet-OPENPOSE.
--
My other checkpoints and Lora.
--
--
Light and color control: sd-channel-offset
--
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。
--
--
Preview images & prompts for 20,000+ Danbooru Characters:
https://www.downloadmost.com/NoobAI-XL/danbooru-character/
Preview images & tags for 6,000+ Danbooru Artists' styles:
https://www.downloadmost.com/NoobAI-XL/danbooru-artist/
Preview images & tags for 4,000+ e261 Artists' styles:
https://www.downloadmost.com/NoobAI-XL/e621-artist/
Preview images & tags for 3,000+ e621 Characters:
https://www.downloadmost.com/NoobAI-XL/e621-character/
--
简易使用方式:
至” Danbooru"或"e621”选择一张图片，将图片中的提示词全部复制，然后输入提示词即可，或是對空格添加逗號並且移除下畫線。
Easy to use:
Go to "Danbooru" or "e621" to select a picture, copy all the prompt words in the picture, and then enter the prompt words, or add commas to spaces and remove underscores.
--
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)
--
使用自然语言并且混合"danbooru"标签时，能获得更好的效果。
Better results are achieved when using natural language and mixing in "danbooru" tags.
--
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)"
--
SDXL-加速器(SDXL-accelerator):
lora權重(lora Weight):0.6~1
取样方法(Sampler)=LCM
Schedule type: SGM Uniform
CFG:1
8~14步(steps)
--
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".
Description
FAQ
Comments (28)
v2 deserves to be so popular that it can be run on Civitai. Really good example images!
Probably the highest adaptability model that tried recently, it can nearly 95% up correspond to the prompt words.
Easily to switch either animation style or realistic style if you add lora correctly
近期嘗試過最好用的模型,能充分對應提示詞
動漫風跟真實風也可以完美切換
Ironically, on Civitai, this generates deep fried images if you use PornMaster LoRAs on them. How to fix this?
I don't know how Civitai sets the parameters for online image generation, but when I use the lora I made on my local computer, I don't have the problem you mentioned.I use automatic1111's webui to run on my local computer.So the lora I made can only be tested on the automatic1111 webui.
hey love your models and loras man keep it up! but i wonder why dont you make them on noobai 1.0 Vpred basis? its much better in many aspects than epsilon version
I don't know how to use Vpred, I don't even know how to install it.
What does "PornMaster-Pro" actually mean? Is it more capable of generating porn stuff?
no it produces mostly flowers and unicorns - clearly - based on the pictures displayed on the page we're commenting on. seriously what kinda question is this..?
Is anyone else having troubles recreating the images? I'm using Forge & noob-V2-VAE and matching all settings but the faces always come out looking funky, either in the eyes, proportions, or face overall.
I've tried all combinations of samplers/schedulers/cfg/steps and it never comes out right.
Maybe it's a forge issue?
I don't know how to use "Forge", my checkpoint is in "AUTOMATIC1111/stable-diffusion-webui'" for testing.
hey can you pls fix the V1 pornmaster? for some reason it gives out nan error while genning and img just break to pitch black one, and even merging this model with some other model result in same broken state, can you like convert it or idk make it again?
I don't know which version you are talking about, but it works normally when I use it in auto11111's webui.
@iamddtla i meant noob v1 model, idk if your pytorch is outdated or something, but on current fresh setup it gives out NaNs error which means its corrupted somewhere and divides by zero its calculation breaking the generation
@makotoyuki I don't know if other people have the same problem as you, and I don't know how to fix the problem you mentioned, maybe you can use my other version checkpoint.
@iamddtla i mean i am using v2 primary tyvm for making it, its just when i tried to remake gens i made with v2 on v1 ive been so amazed and suddenly it got black mid gen with this error /usr/local/lib/python3.11/dist-packages/torchvision/transforms/functional.py:282: RuntimeWarning: invalid value encountered in cast npimg = (npimg * 255).astype(np.uint8)
i just hope mb theres some way to maybe remake it or something, that will fix models corrupt state
@makotoyuki I can't help you with this technical issue.
@iamddtla ok sry for bothering bro, btw where is pornmaster noob v0.1? i rly liked your gens made on it. can you pls upload it?
@makotoyuki It was removed for violating regulations.
@iamddtla can you pls upload it somewhere else and mb share the link to me in dms? id love that, and tyvm again for your loras and model, they are so amazing, so good, literally 0 coherence and concepts lost even tho it was trained to be realistic
@makotoyuki I will not release deleted versions again.
Definitely one of the best checkpoints for prompt following and anatomy. Only tried v15, trying the new one now. My only wish is that the faces are more expressive, eyes more realistic, less "dead stare" look! Also @iamddtla this is very compatible with clip_l (not clip g) from animagine4. Although animagine gives lots of style and distortion it actually has amazing knowledge and variety when you tame it a little with some gradA mixing of other checkpoints or or DMD2 lora.
hey can you help me pls, ive noticed if i try to use many of anime chars with realistic setup their faces get all creasy and wrinkly and weird, this happens both with style loras added slightly into the prompt or overral heavily anime looking chars in the model, how do you counter this? 0.8 weight on char tag often doesnt change much
From what you described, I really can't understand what happened to you. Maybe other users can help you solve your problem.
Is Noob-V2-VAE epsilon or v-pred?
epsilon.
This model gave me the most intense orgasm I ever felt since Nov 2022.
Thank you.
Proud of you, man.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.