Atomix Pony XL
This is a mixed model based on Pony Diffusion V6 XL, which merged a variety of other models and Lora. Finally, I have achieved a style that I feel good about, I hope you will also like it.
Tips
This model is base on pony,so any pony Lora is working for this model.
This model has merged common Lora, which you can see from the Suggested Resources. So you do not need to import these Lora again.
Each image I upload contains a complete workflow, which you can download the image and drag to the ComfyUI interface to obtain.
Suggested Setting
Positive prompt:
score_9, score_8_up, score_7_up,Nagative prompt:
score_4, score_5, score_6, Sampler:(Turbo acceleration only work with these settings):
cfg: 1~2
steps: 6~10
sampler_name: dpmpp_sde_gpu
scheduler: sgm_uniformDescription
Just adding some styles that I like
FAQ
Comments (20)
Real MVP of pony models,v1 best, v2 gone somewhat worse and pushes out more generic stuff.
I'm glad that you like my model and thank you for your feedback. yep, V2 just merged some style Lora.
wow, great work, I have some anime characters in mind that I want to merge/bake into a pony base checkpoint as default characters instead of using their loras. Would you please teach me how to merge/bake anime characters to another checkpoint (e.g. Pony Diffusion V6 XL)?
If you are using ComfyUI like me, you can refer to this article Model Merging Examples | ComfyUI_examples (comfyanonymous.github.io)
@AlexLai so sad, I dont use ComfyUI, I use A1111 or Forge, and Kohya_SS for training
I wonder what cool creations people will ma--- oh nevermind
The desire for food and sex is part of human nature --Mencius
sorry to bother, im a total newbie with ai art generation and im in the process of reading and watching some tutorials, i wanted to replicate one of your images, using the same references and resources and i made sure everything was the same, but faces keep coming out as blurry messes, could i be pointed in the right direction? like a resource where i can do some reading? im eager to start using your models at their full potential. btw im using automatic1111 branch of sd.
Hey,you need to upscale original image, it means Hires.fix in automatic1111,in txt2img interface.
I highly recommendyou use ComfyUI,which faster and More convenient than automatic1111.and if you do,you can easily get others's workflow.
@AlexLai Thank you very much for your answer! I’ve been slowly improving my skills, and I can already see better results. I’ve been trying to used to ComfyUI, but I just can’t at the moment. I’ll surely move to that UI once I improve a bit more.
@blackk_magic good for you👍
Hi, thanks for this Pony going straight forward into the style I'm looking for. Got a question though... you say it is very important clip skip value is -2. But in WebUI settings it is only possible to set positive values. Does it need to be used with an other GUI ? Or is it a typo ?
In ComfyUI it is -2, in WebUI it is+2. I found that now it can be unset without affecting the image,so you don't have to pay attention to this.and i will update the Description.
@AlexLai Thanks for your answer :)
Hello the model is great and I like it very much. Sorry for the stupid question but this is my first pony model and whenever I use the adetailer the face is distorted or blurred. I have tried yolo8n and yolo9c, same thing with both. Is there a setting I need to change? Thank you
Thank you,I'm glad you like my work.
I have also been troubled by this issue before, which only appeared on the Pony model and was difficult to fix with face detailer. My current solution is to use an increase in upscale resolution . When the resolution is increased by 3 times or more, facial problems can be greatly improved, and there is basically no need to use face detailer.
This is the workflow I am currently using. default 3x upscale:Atomix Txt2Img Workflow - SDXL | Stable Diffusion Workflows | Civitai
make shore you are 1024 x 1024 not 512 x 512
I knew that I had seen this message already! xD
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.