ComfyUI ReActor Face Swap Image & Video
Vložit
- čas přidán 15. 06. 2024
- Workflow: github.com/dataleveling/Comfy...
Custom Nodes
ReActor: github.com/Gourieff/comfyui-r...
Video Helper Suite: github.com/Kosinkadink/ComfyU...
WAS Node Suite: github.com/WASasquatch/was-no...
Comfy Frame Interpolation: github.com/Fannovel16/ComfyUI...
Models
CodeFormer: github.com/sczhou/CodeFormer/...
GFPGAN: github.com/TencentARC/GFPGAN/...
In this captivating video, embark on a journey through the innovative realm of face swapping with ReActor Node on ComfyUI. Delve into the intricate process of seamlessly swapping faces in both images and videos, utilizing cutting-edge insightface models.
Discover the power of ReActor, a bespoke node meticulously designed to revolutionize face swapping techniques. Learn how to harness its capabilities to effortlessly transform faces with precision and finesse.
Elevate your editing prowess with the Video Helper Suite, a comprehensive toolkit featuring VID2IMG functionality and advanced frame interpolation techniques. Witness how these tools synergize to enhance video quality and ensure seamless transitions.Navigate the challenges of handling videos with shifting facial positions, gaining valuable insights into effective problem-solving strategies.
Above all, remember the importance of ethical usage and responsible application of these powerful tools. Let's embark on this transformative journey together, embracing innovation while upholding ethical standards.
📈 Timestamps:
0:00 - Introduction
0:23 - Requirements
1:22 - Face Swap IMG
4:04 - Face Swap VID 1
6:51 - Frame Interpolation
8:58 - Face Swap VID 2
11:57 - Outro
👍 If this video helped you, do hit that like button, subscribe for more content, and let us know your thoughts in the comments below. - Věda a technologie
jesus man, you are just soo good. i understood everything.
Error occurred when executing ReActorFaceSwap:
Landmark.get() takes 3 positional arguments but 4 were given - can you help ?
Bro, you just saved my assignment
Very good tutorial, thanks
Molto interessante, grazie!
Thank you very much for such a detailed step by step tutorial, even a child can follow this. I am curious to know how did you understand and learn all this ?
Is there a specific place or forum where one can get a formal training to learn all these type of softwares ?
Great stuff as usual mate im having a hard time with hair over the face eg witcher scenes and arty styles
So awesome if only i had a gpu to do any of this. I might be able to use reactor on cpu for faceswap though.
❤thank you
Love your vids! Can you make a tutorial for AnimateDiff?
Will make one on that... soon!
great tuto but i have one question : When I go through Reactor, I first create an image on which I apply Reactor and Face Detailer to make the whole thing coherent. However, I lose the details of the face from the first generated image: tattoos, scars, war paint, etc. If there's a way to solve this problem, could you please make a tutorial about it?
I get this error Error occurred when executing ReActorFaceSwap:
This ORT build has ['TensorrtExecutionProvider', 'CUDAExecutionProvider', 'CPUExecutionProvider'] enabled. Since ORT 1.9, you are required to explicitly set the providers parameter when instantiating InferenceSession.
anyone know how to fix this??
I can not install Reactor node in comfyUI it says (IMPORT FAILED) ReActor Node for ComfyUI
Hi, did you follow the steps correctly as I just updated the ReActor custom node to latest and faced no issues. Reinstalling the custom node after installing insightface might help!
thx for friendly answer. is there any other swap_model or new except inswapper_128? it works for not bad. but i'm looking for better model :)
Hey sorry for the late response, sadly the insightface developers only provide the inswapper_128 opensourced... However you can use their discord / midjourney's faceswap where they have they license for higher resolution face swap
Is there any way to reproduce the Lil Yatchy video replaced by the Joker using ComfyUI? I can't get pass the face swap
Is there a way to feed it multiple images for better quality?
facerestore_model folder is not there in models .......
what should i do ?
can i suggest a video idea? - where you show what what you ahve created using all these techniques and methods. like to show what what is possible
Great idea, will keep that in mind!
I thank you for ur great work.
I get a error working with ReActorFast Face Swap .
I can mention that i downloaded insightface and a workflow with ipadapter using insightface worked without error.
But with ReActor i get this:
Error occured when executing ReActorFaceSwap: This ORT build has ['TensorrtExecutionProvider', 'CUDAExecutionProvider', 'CPUExecutionProvider'] enabled. Since ORT 1.9, you are required to explicitly set the providers parametee when instantiating InferenceSession. For example, onnxrunyime.InferenceSession(..., providers=['TensorrtExecutionProvider', CUDAExecutionProvider', ' CPUExutionProvider'],...)
i got some error like 130563 mb more needed when i use face restore model codeformer or any model
When using the repair model, the message ‘Error occurred when executing ReActorFaceSwap’, what's wrong with it, it's been bothering me for a long time. 1：Weights only load failed. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution.Do it only if you get the file from a trusted source. WeightsUnpickler error: unexpected EOF, expected 101452458 more bytes. The file might be corrupte
i did it all following your video, but my video has changed only 1 second among 5 second. how can i fix it?
Hihi do you mean your 5 second video was reduced to 1 second?
A few things to check, make sure
1. select_every_nth is corresponding to the frame interpolation multiplier (if multiplier is higher than select_every_nth, frame_rate in video combine node has to reduce)
2. frame_load_cap to be set to 0 to load all the frames
3. Default frame_rate to be same as original video's fps before making adjustments.
what is the extension that tags what extension each node comes from?
Hi, that is not really an extension. It is a setting under ComfyUI Manager -> Badge: Nickname
how to add codeformer and yolo_face model
Prompt outputs failed validation: Required input is missing: images
VHS_VideoCombine:
- Required input is missing: images
ReActorFaceSwap:
- Value not in list: face_restore_model: 'codeformer.pth' not in ['none', 'codeformer-v0.1.0.pth', 'GFPGANv1.3.pth', 'GFPGANv1.4.pth', 'GPEN-BFR-512.onnx']
why?
i got error no module cyhton
iam getting error while generating image the error is ""landmark.get() takes 3 positional arguments but 4 were given"" pls suggest what i am missing
Hi, I have not encountered that error before but others have from here: github.com/Gourieff/comfyui-reactor-node/issues/136
The solution seems to be along the line of replacing the inswapper_128.onnx file in ComfyUI\models\insightface
inswapper_128: github.com/facefusion/facefusion-assets/releases/download/models/inswapper_128.onnx
i installed but i don't see the ReActor folder. Can you guide me?
Ahhh I see. The 'facerestore_models' folder should be created under
ew_ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\models after you install the nodes on the manager, close the ComfyUI terminal and start the server again.
This is because when the git pull happens only after restart and all the actions related to ReActor will start to execute.
After that step, check if ReActor is installed correctly through ComfyUI Manager filter installed. Then check the folder again, if its still not there, create the 'facerestore_models' manually and place the models inside it.
Hello, can you guide me to export video, high quality
(IMPORT FAILED) toooo
yo bro, i got this error while using queue prompting with Film VFI:
Error occurred when executing FILM VFI:
The following operation failed in the TorchScript interpreter.
Traceback of TorchScript, serialized code (most recent call last):
File "code/__torch__/interpolator.py", line 15, in forward
x1: Tensor,
batch_dt: Tensor) -> Tensor:
_0 = (self).debug_forward(x0, x1, batch_dt, )
~~~~~~~~~~~~~~~~~~~
After updating PyTorch, this node stopped working. I created an issue on the GitHub repository for this node