You can use the online workflow below. If the results are good, you can then try setting it up and running it locally.
👉 Click Here For Online WorkFlow
As a fan benefit, you’ll receive 1,000 points upon registration, plus 100 points for daily logins. Enjoy a smooth experience with an RTX 4090 and 48GB of VRAM.
Model Description
I've tested several LoRAs trained for "Anime -> Realistic" style conversion, but wasn't fully satisfied with their stability and aesthetics.
That's why I trained this new style-enhancement LoRA to achieve more balanced results.
Recommended Settings
Sampler: Euler
LoRA Weight: 1
Pair with Face-Model LoRA:
I recommend my other face-model LoRA 👉 VirtualIdol01 by aldniki
Or replace it with any face-model LoRA you like
Suggested Weight: 0.65 (going higher is not recommended)
Use Cases
✅ Convert anime-style images into realistic style
✅ Repainting real-life photos
✅ Simple rendering & coloring for B/W sketches, clay renders, or depth maps
✅ Mix and match with other LoRAs for diverse style explorations
Feel free to experiment and discover more possibilities!
模型介绍
此前我测试过一些其他作者训练的「二次元 -> 写实」风格强化 LoRA, 但在稳定性和美感方面并不完全满意。
因此我训练了这一版风格强化 LoRA, 以实现更好的画面表现力和稳定度。
推荐设置
采样器: Euler
LoRA 权重: 1
搭配脸模 LoRA:
推荐我发布的另一个脸模 LoRA 👉 VirtualIdol01 by aldniki
当然也可以换成自己喜欢的脸模 LoRA
权重建议: 0.65, 不宜过高
使用场景
✅ 二次元风格 -> 写实风格转换
✅ 现实照片的重绘
✅ 黑白图 / 白模 / 深度图的简易渲染与上色
✅ 与其他 LoRA 搭配实现多样化风格探索
欢迎大家多多尝试, 发挥更多可能性!
Description
FAQ
Comments (19)
Add prompt to adjust facial expressions. or stiff face
: : Workflow 🔀🔜 20sec convert
https://civitai.com/models/1831687
hey man, if you want to do it even better- convert anime to 3d render first, then try to convert this 3d render to more realistic but keep the same face , this lora changes face too much, but when you convert to 3d - the face will be very similar and you only have to make it more realistic, even with other model and face mask
Thanks, this approach sounds good. Do you have an example I can look at to see the effect?
Works great but is there any way to have non-asian faces? Even if i use loras of western girls they still look asian. Prompts are not helping either.
Version 1.0 mainly used a dataset of Asian faces, so the issues you’ve noticed are likely due to the limitations of the dataset. This might be improved in future updates.
I see you are promoting Runnig Hub, I'm testing it now...
Just a question for my curiosity: Are you a chinese?
And is yes, are you in the team of Running Hub or are you just a user?
Generally, just a user
This lora creates an error in cmd;
CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16
clip missing: ['text_projection.weight']
model weight dtype torch.bfloat16, manual cast: None
model_type FLUX
Requested to load FluxClipModel_
loaded completely 13505.926642227172 9319.23095703125 True
Requested to load Flux
ERROR lora diffusion_model.final_layer.adaLN_modulation.1.weight shape '[6144, 3072]' is invalid for input of size 196608
I’ve tested it both locally and on the cloud, and other users haven’t had any issues generating images. What are you using to generate the images? If it’s ComfyUI, which version is it?
@aldniki217 It generates images allright, it just creates the error in cmd. According to Googles AI it is probably because the lora is not trained properly on the Flux Kontext model.
BTW; ComfyUI latest update & ComfyUI default Flux Kontext workflow.
your workflow doesnt work in comfyui
what's the issue
@aldniki217 TypeError: Cannot read properties of undefined (reading 'type')
@sergiomlbb21248 The snippet you gave is a pretty vague and generic Python error, so it’s hard to tell where the issue is coming from. Can you provide a more detailed log? It’d help if you could include where the error is originating.
@aldniki217 When i import the workflow to comfyui that error appears. # ComfyUI Error Report ## Error Details - Node ID: N/A - Node Type: N/A - Exception Type: La carga se interrumpió debido a un error al recargar los datos del flujo de trabajo - Exception Message: TypeError: Cannot read properties of undefined (reading 'type') ## Stack Trace ``` TypeError: Cannot read properties of undefined (reading 'type') at nodeType.onConnectionsChange (http://127.0.0.1:8188/extensions/ComfyUI-Impact-Pack/impact-pack.js:327:57) at ComfyNode.configure (http://127.0.0.1:8188/assets/index-CWzmkThr.js:108699:33) at ComfyNode.configure (http://127.0.0.1:8188/assets/index-CWzmkThr.js:268034:15) at LGraph.configure (http://127.0.0.1:8188/assets/index-CWzmkThr.js:118365:16) at LGraph.configure (http://127.0.0.1:8188/assets/index-CWzmkThr.js:275567:26) at LGraph.configure (http://127.0.0.1:8188/extensions/comfyui-custom-scripts/js/reroutePrimitive.js:14:29) at ComfyApp.loadGraphData (http://127.0.0.1:8188/assets/index-CWzmkThr.js:275819:18) at async app.loadGraphData (http://127.0.0.1:8188/extensions/comfyui-manager/components-manager.js:774:9)
@sergiomlbb21248 I can't pinpoint exactly what the problem is, but you can try updating the Impact Pack. Or just delete the Impact Pack and reinstall it. I’ve also encountered errors caused by the Impact Pack before. Also, make sure your ComfyUI version is up to date.
Hey, amazing lora, if possible can you make a new version of this Lora where the face is also remains the same, not human but as close as the anime character?
求fp8版本的
I love this, but is it possible to make the conversion without it cropping or changing the size of my reference images?
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.