Better Face Swap = FaceDetailer + InstantID + IP-Adapter (ComfyUI Tutorial)

preview_player
Показать описание
Discover how to use FaceDetailer, InstantID, and IP-Adapter in ComfyUI for high-quality face swaps. Achieve flawless results with our expert guide.
Note: If you encounter difficulties locating the IP-Adapter Unified Loader node, please ensure that you upgrade ComfyUI to the latest version. Additionally, you should download the "IP-Adapter Plus" node available in the ComfyUI Manager, developed by cubiq. This step is crucial for accessing the complete set of features needed for the workflow.

🖹 Article Tutorial:

🔗 Links:

Комментарии
Автор

I made a mistake in this video: If you encounter difficulties locating the IP-Adapter Unified Loader node, please ensure that you upgrade ComfyUI to the latest version. Additionally, you should download the "IP-Adapter Plus" node available in the ComfyUI Manager, developed by cubiq. This step is crucial for accessing the complete set of features needed for the workflow.

my-ai-force
Автор

Using this method perfectly solved the problem of repairing the face in a full-body frame. Thank you very much!

derekdu
Автор

Thank you for the tutorial! this is a great tutorial for beginners fellas.

anseltrust
Автор

Could you help me with the installation, get several videos that are 3 months old but they don't solve how to install it in the current portable version of comfyIU

sandocarlos
Автор

Thanks for the great tutorial. Exactly what I was looking for. But I have a question. How long time it take for the full process for you? In my case RTX 4070 takes about 265 seconds. It's too much. Is where any case to speed up the generating process with similar result?

membranobruno
Автор

How can I make sure it also keeps the hair and body structure of the reference image like it does if i only use a instantid face model?

familienieuwenburg
Автор

hello, thank you for this good tutorial, i was wondering you can take this workflow further but adding the inpaint-anything model for further background processing

godly_wisdom
Автор

Everything works perfectly for me BUT I don't want to create a whole new image to change it's face. I want to import my own image and then change the face. Is this possible in this workflow?

GoodCharless
Автор

and if it is there missing nodes, not working nodes, that's because some nodes are can't able to work properly with eachother, also try to download it manually with command prompt if it is there missing nodes, there are youtube videos for this:)

anseltrust
Автор

Looks really good, how does it perform against pulid or faceid ?

nguwusic
Автор

Really great video!! I am interested in creative upscaling, like turn a simple into a hyper realistic prompt following generation. Something like Magnific ai and others. Any Comfy ideas?

WhySoBroke
Автор

Hello, Thanks for the video. Just a thing, is there a FaceSwap Workflow for img2img?
I use to do FaceSwap with A1111 and it works good but the face result is often not realistic, I mean, I'd like to have a face with realistic skin texture. There is a way with ConfyUi to process the Image after swapping? Thanks!

giusparsifal
Автор

Through IPAdapter Unified Loader FaceID report an Error when the Error occurred when executing IPAdapterUnifiedLoaderFaceID:

ClipVision model not found.

junchien
Автор

好主意, 相当于先让提示词直接生成画面, 再用facedetailer修改画面

fukong