RH Invite Code:
https://www.runninghub.ai/?inviteCode=rh-v1279
New user with 1000 free RH credit
All my workflow and models are avaible on RunningHub
F2K 9B Anything2Real A
New version of Anything2Real on F2K model.
Based on previous version, adjusted the style to more realistic and less 3D
Flux2 Klein 9B
Based on strong transformation ablitiy on Flux2 Klein 9B
The new anything2real lora adjusted aesthetic and image fidelity
Prompt:
transform the image into high quality realistic photograph. {male/female}
The model mostly trained on female images. Please add male if the input image is male.
Lora strength:
0.8-1
The lora tends to be a little bit more "3D" use less strength to gain more realistic from base model.
RH App One Click Transform Image:
https://www.runninghub.ai/ai-detail/2016070515554783233/?inviteCode=rh-v1279
RH Workflow:
https://www.runninghub.ai/post/2016060164108980226/?inviteCode=rh-v1279
RH Comparer with base:
https://www.runninghub.ai/ai-detail/2016075121722658818/?inviteCode=rh-v1279
RH Workflow:
https://www.runninghub.ai/post/2016072419668140033/?inviteCode=rh-v1279
Hugging Face:
https://huggingface.co/lrzjason/Anything2Real/tree/main
f2k_anything2real.safetensors
Qwen Edit 2511
RH APP:
https://www.runninghub.ai/ai-detail/2007507982543757313/?inviteCode=rh-v1279
RH online workflow:
https://www.runninghub.ai/post/2007432627858444289/?inviteCode=rh-v1279
Anything2Real 2601 A is an adjusted version which mixed alpha more and with extra weight adjustment based on 2601. It produces stronger transformation between image and realistic.
This LoRA, built on Qwen-Edit 2511, incrementally maps images of any visual style into the photographic domain. The training objective is to minimize perceptual distance rather than to perform pixel-level reconstruction.
Training Protocol
Dataset
7 style-paired subsets, 200 source images ≥ 1024 px, covering portraits, landscapes, and objects.
Training Pipeline
Two-stage tuning: (a) trigger-word + long-caption pairs, (b) trigger-word only.
Layer-wise merge with Anything2Real α; weights manually tuned to preserve low-frequency structure while replacing high-frequency texture.
Recommended Positive Prompt
transform the image to realistic photograph. {detailed description}
Note: {detailed description} is optional. Style-related tokens such as “anime”, “cyber”, or “oil-painting” are strictly forbidden; their inclusion will inject out-of-domain noise.
Strength Tuning
Default range: 0.8–1.0.
If over-smoothing or texture loss occurs, decrease in 0.1 steps down to 0.5.
Should realism remain insufficient, blend with an additional photorealistic LoRA and adjust to taste.
Because “realism” is inherently subjective, first modulate strength or switch base models rather than further increasing the LoRA weight.
Contact
Feel free to reach out via any of the following channels:
Twitter: @Lrzjason
Email: [email protected]
QQ Group: 866612947
WeChat ID:
fkdeaiCivitAI: xiaozhijason
Description
FAQ
Comments (20)
Hopefully this one works better. Qwen usually adjusts proportion and anatomy to be more human-like, Klein just makes big headed characters instead and that was pissing me off. The previous version didn't fix the big head problem on more stylized references for me.
- After some testing, it does work better, get more realistic results, but the bobble-head problem remains. I'm pretty sure it's a model limitation, since that's how it works without lora also. If the reference is semi-realistic or realistically proportioned it works the best, but anything more stylized turns into a bobble-head
No space on drive. Can someone run a picture of Jinx and/ or Jasmine from Aladdin through this for me? Will buzz
Here you go:
@ultimo_intento Two-for-one, thank you kind sir. I deposited 20 buzz into your account for your efforts
Works pretty good to reign in how much Klein likes to change too much of the source image. Still doesn't retain enough to my liking but i'll be using this for every one of my images anyway!
Damn, klein version is amazing refiner for "ai slop to actual photo"
Holy moly! Using this as a refiner with Klein is OP. You just changed my whole edit model workflow!! Thank you!
感谢更新！flux的版本很快，但flux的底模会偏向于生成欧美、日本老写真集素颜审美的娃娃脸，肉肉的脸颊和法令纹、过于真人的无妆容无美颜感（和另一个叫Qwen Edit Omni Cosplay By Aldniki的lora的风格好像），以及不太现实的大眼睛，需要加不少提示词来修正。
我认为qwen2511配合alpha和2601a的效果是最棒最接近cosplay妆容的！眼睛很好，我会经常用这两个来roll图，就是脸边缘有一些模糊感和遍布各处的焦外虚化，且对让动漫脸型变瘦这一点不管哪一款的提示词都显得比较无力，以及张嘴时牙齿、龅牙太明显了(尽管嘴开的很小)。
2601a确实还可以，klein 9B练得比较一般，有时间再更新。
Really liked the QWEN version. This one is my favorite. I use an optional prompt that helps gets some good results.
"transform the image into high quality movie still. color balance. maintain background."
It's great that you can still use Klein9B's prompts to edit in the perfect Output.
Kudos and great work
Thanks for comment. My favourite is Qwen Edit 2601A version. 9B version seems under grade for me but I was quite busy at that time. Might train a better version if I have time
@xiaozhijason Both versions are great. The underlying models are what make each unique.
This version picks up on Klein9b's ability to really fine tune edit via prompts. Klein more effective with human based imagery.
The QWEN version does a great job with non human/fantast rpg images without special prompting.
Again. Thanks for this LORA. Bringing many old (ugly) images back to life.
I created a workflow that compared the lora effects of multiple anime-to-real human conversions, including this lora（anything2real_2601_A & f2k_anything2real_a_patched）.You can click the link to give it a try.
Thanks to the author for providing such a powerful tool for our community!
link returned 404
@wallstreetnoobs410 Try again. My clicks are normal
how to preserve expression? the end result always has the character looking back at me
Good job . Its realy work perfect .
Can u make it without asian faces ?
Is there a Flux2 Kelin 4b version for it? thx
No