Mastering ComfyUI: How to use ReActor for Face Swap - TUTORIAL
HTML-код
- Опубликовано: 7 сен 2024
- In this video, I'll walk you through the process of creating flawless faceswaps using the ReActor node. We'll cover installation, model selection, and how to achieve professional-quality results. So, grab your favorite images and let's dive into the future of AI-powered faceswaps together! Don't forget to like, subscribe, and leave your questions in the comments!
** Links from the Video Tutorial **
- ComfyUI ReActor Node: github.com/Gou...
- Video by Freepik: www.videvo.net...
- ComfyUI-N-Suite (for the video nodes): github.com/Nuk...
- Workflow** : www.patreon.co...
** Let me be EXTREMELY clear: I don't want you to feel obligated to join my Patreon just to access this workflow. My Patreon is there for those who genuinely want to support my work. If you're interested in the workflow, feel free to watch the video - it's not that long, I promise! 🙏
❤️❤️❤️Support Links❤️❤️❤️
Patreon: / dreamingaichannel
Buy Me a Coffee ☕: ko-fi.com/C0C0...
Hey guys, only a little update: The portable enviroment of ComfyUI has been updated to python 3.11 so make sure ,if you have this new enviroment, to select the correct version of Insightface!
hai insightface problem...!!
I wasted 3 hours for reactor with no results, thanks to you i managed to do the job in minutes.Thanks a lot man.
I. Am. Gobsmacked. It actually worked. This is like the backdoor to the matrix. Thank you.
After checking 20+ of ReActor guids this is the first one that actaully work, Great Job!
Of all tutorials, this was the only one that worked. everyone else skipped over the most important which was installing IN the python_embeded folder. They didn't even specify to download it to THAT specific path. that's literally all it took
This is the first guide that works for me, I was using python 3.11 so if your also using 3.11 do everything as described but, use the other prebuilt Insightface package
you dont believe it, i try to fix this problem all day, your video saved me, thanks a lot, love you!
I loved the AI distinguished gentleman voice saying, "I'm just sorry that this type of operation is not usable for anime style models, which is my FAVORITE style."
This is the best and most detailed tutorial I've found about installing ReActor, thank you!
Thank you so much, for helping me find the light
You always provide the most accurate and relevant information. Thank you!
Thank you so much for making this video! After following several videos of installing Insightface, only yours works. Seems like it has to be installed directly into python_embeded folder instead of the root folder as many instructed.
THANKS!!!! I've been racking my brains for almost a month on this!!!! subscribed :)
thx, your Step by Step-Explanation is awesome !
simplicity and clarity
Thank you!!! I'v been trying to install it for a long time and got stuck with the errors.... Also, thanks for explaining the nodes so clearly, really amazing.
Man you're AWESOME! Thanks a million!
Asolute fantastic video... simple, understandable and friendly
this is the only place fixed my problem !! thanks
such a good video, simple and informative. Thanks!
Very good! Thanks a lot. You actually helped me a lot with the face swap I tried to do by myself. Keep going with these tutorials. Cool!
thank You for that video , I use it and final video looks so good
like the concept and the style! keep em comming :)
This video has been super useful, thank you so much!
Thanks for your sharing.
Could you create a video about outpaint on comfyui using sdxl?
I need to try that first! If everything goes well then I will make a video for sure!
awesome video, do you mind sharing what AI voice you are using? It sounds great
Greetings, I am a newbie to this type of program, I have some problems installing ComfyUI-N-Nodes, although due to the lack of nodes it has the name SaveVideo.
I will continue watching your videos, maybe there is one that will help me, thank you
Hello and thanks so much for the video. There is a way to improve the face AFTER the Swap? I mean, in A1111 the swapping is the last thing then you can't do anything else. The result is ok but the face is often poor realistic as in the example here in your video. Is there a way in ComfyUi to correct this? Thanks!
Hi! You can try using the facedetailer node but I wouldn't want that to distort the swapped face too much!
Where do we get the LoadVideo node?!
Thank you for your tutorial, but how can i make videos with AUDIO ?
I am getting a bunch of errors when installing ReActor custom node. It fails to get requirements to build wheel. It says "ModuleNotFoundError: No module named 'Cython'" Subprocess exited with errors.
Great video, I get the following error while installing the troubleshooting file for python .10 "ERROR: insightface-0.7.3-cp310-cp310-win_amd64.whl is not a supported wheel on this platform." Any idea how to fix that?
Me to. The installation file doesn't appear after pip install in the comfyui directory and while exhausting all options try the webui directory, get the same error as you.
Cancel that - I re did the first step and it worked .. in under 30 seconds lol
Any idea how to PADDING at output face ? i want to swap on the eye and nose area and leave mouth area, i try figured out how to do that without masking cause i have batch process image...
example i have image man eating big bread in front his mouth, but Reactor always change that bread into mouth shape
Thank you so much, finally got Reactor to run😀. Had to use Python 3.11 , 3.10 wouldn't load for me.
Is there a way to connect it to the promp to create characters with the face of someone? i can do it on the normal stable diffusion....but in comfyUI??
Thank you! how to swap complet face with hair?
the following error appears with the Load Video node: Error occurred when running LoadVideo [n-suite]:
'NoneType' object has no 'shape' attribute
Would you help me?
after installing I could not find the facedetection folder inside the models folder, any ideas why?
Yes, you need to create i!
Amazing tutorial! Thank you very much! Is there a way to further perfect the face swapping? Some results I get turn out like bad photoshop and in the good cases the end result looks more like a merger of the faces, not really a swap. :( Also, when I tried the video one my results didn't have the initial audio. Will this always be the case? Thank you again for your amazing content!
Hi! Perhaps you can refine the face after the swap with ImpactPack's Facedetailer custom node.Also the audio is only removed when rendering part of the video, but if you are rendering the entire video the audio is preserved.
I cant get through 1:47, the error that comes up is: ERROR: insightface-0.7.3-cp312-cp312-win_amd64.whl is not a supported wheel on this platform
im trying to use a windows machine with very limited knowledge of windows since I use mac all the time
i tried again with python 3.10.11 installed instead of 3.12. and i got to the 1:41 part of your video and when I entered that command to install that whl file I was told by terminal: "ERROR: insightface-0.7.3-cp310-cp310-win_amd64.whl is not a supported wheel on this platform."
Thank you very much for this. I am unable to find load video when searching as you did in this tutorial. Do I need to install other components? **Update: I should of read your description, all good now :)
GREAT TUTORIALA . BUT FROM WHERE CAN I FIND THE FACE RESTORER MODEL LINK . I CANT SEE IT ON THE GIT HUB PAGE
Hello! Just press on the text with the arrow that says "Standalone (Portable) ComfyUI for Windows" under the "Installation" section
thanks . you are a life saver !
@@DreamingAIChannel
ModuleNotFoundError: No module named 'onnx'
I receive this message despite having insightface installed properly
Anyone else having issues with face reactor recently ?
Mine stopped working, Ive reinstalled it and even gotten an updated version of comfyui but it still dose not load.
It use to work fine but just stopped working recently.
in SD 1111 there is an option "restore face mask". is it in Comfy ui?
summary of errors and fixes in comments: Seems like it has to be installed directly into python_embeded folder instead of the root folder as many instructed. -- The portable enviroment of ComfyUI has been updated to python 3.11 so make sure ,if you have this new enviroment, to select the correct version of Insightface! ---- from my iphone (jpg) .I'll try to convert it to png then ------ there is no FACERESTOR_MODEL folder.where can i find that? if you dont have that folder then just create one ----- finally got Reactor to run. Had to use Python 3.11 , 3.10 wouldn't load for me. --- If you swap the model to "inswapping128", as default, it works ---
One of latest ComfyUI updates stopped Reactor from working. Follow the instructions under this heading " If you see errors with building "future-0.18.3" package, do the following: " (( whether you get this error or not!!! ( This is were I wasted lots of time)) The Reactor should work again.
I keep on getting this error: Error occurred when executing ReActorFaceSwap:
'NoneType' object has no attribute 'get'
Like this one? github.com/Gourieff/comfyui-reactor-node/issues/11
Thank you for this wonderful tutorial. You made it all sound so easy. I am facing this error in console when I have "save video" option as TRUE under SaveVideo node. Error: Video Path already deleted
Nbm, the above error was not an issue. I have a different question. I couldn't find any help on "frame interpolation". I am wondering how does the multiplier work? Could you please explain? Thanks in advance.
So I've explained it a little bit in the "Video to Video"... Video 🤣 Essentially it recreates a number of frames equals to the multiplier per the number of frames per second, for example if you have 4 fps, with a multiplier of 4 you will bring the video to 16 fps.All thanks to a technique ai based called RIFLE. Be aware that the more frames you will try to recreate the more artifacts will appear in the final video.
Thank you :)@@DreamingAIChannel
why I can't find python_embeded folder
great tuto thanks but few question : what are the custom models for the video nodes plz ? And is 12 gb Vram enough for that ?
Hi, is in the description! And yes, i have 10 GB of VRAM and it run perfectly!
@@DreamingAIChannel excellent news thanks
hey my queue gets stuck on the LoadVideo node after it extracts the frames to the temp folder if its plugged in to the ReActor node. it seems to work fine if i do LoadFramesFromFolder to ReActor. any idea on how to fix this?
i don't think that is stuck, maybe you are sending too many images at a time, did you put an images_limit or at least a batch_size?
@@DreamingAIChannel thanks! i stupidly left the image limit at 0
@@badterry 👍
@@DreamingAIChannel hey another question. do you know how i can do longer videos for faceswap? lets say i have 300 frames but i have to image_limit of 30 for it ot work, is there a way i can queue it up in 30 frame chunks?
Yes, just use batch_size and put 30 there!
hello can anyone help me with this Error " D:\a\_work\1\s\onnxruntime\python\onnxruntime_pybind_state.cc:891 onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasnt able to be loaded. Please install the correct version of CUDA andcuDNN as mentioned in the GPU requirements page"..????
Hello, does anybody knows how to run this same functionality in Python or Diffusers? Thanks!
2:04 I don't have that facerestore_models folder. I did all the previous steps and fixed the errors too.
So you need to manually create it!
Embedded python not found. Please install manually.
Pressione qualquer tecla para continuar. . .
?
i got this too@@DreamingAIChannel
Great video, thanks! How would you batch process all the images in a given folder to faceswap them, without having to select them one by one?
Uhm you can use LoadFramesFromFolder node, it's in my suite
Great stuff as usual. Do you think this will work with 4GB VRAM? I can get AnimateDiff to work if I lower the res and stick with 24 frames. I suppose if it works on ONE frame at a time it could work 🤞🏻
Yes, I think it will work!
"ERROR: Could not install packages due to an OSError: [WinError 206] The filename or extension is too long:" Any idea on how to solve this?
To fix this error on your Windows machine on regedit and navigate to Computer\HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\FileSystem and edit LongPathsEnabled and set value from 0 to 1
Great tutorial. I always get "this video has X frames per second, reducing to 12 frames per second", though. Do you know if there's a way to do a full fps convert? I guess you could do some interpolation, save the images and then mix a final video on your computer or something. But maybe there's an easier way?
If I understand correctly you want to use all the frame of the original video right? If so just put "Original" in the framerate of the videoload node
@@DreamingAIChannel Thanks! I did and it looks a bit weird and blocky when playing it back. Might just be my settings, will test it further. Thank you for the amazing tutorials!
very best..
Hello!! Thanks for this tutorial. I was able to install everything with no issues. I was able to make a face swap. But now when I load the workflow, SaveVideo is missing. I'm not sure what could've caused this. I've done some other installs, but didn't know I'd erase a node.
Hi! That's because I did an update like 1 weeks ago and I had to change the name of the node, so you should search it again in the menu, is called SaveVideo [n-suite]
@@DreamingAIChannel ok. now I get a 'tuple index out of range' error
In which node? Because it is possible that ReActor is simply not recognizing any face in the image
@@DreamingAIChannel i was missing one of your nodes. I got it to work. Thanks!!!
@@carlankrum7137 👍
thanks for nice tutorial, it work great, i only have 1 issue, the video can't play on "load video" and "save video", but i can found it on my out folder, anyone know what plugin or custom node is missing?
Hi! Update the Animatediff node to the last version
@@DreamingAIChannel i did updated, looks link ComfyUI-N-Nodes Conflicted Nodes:
LoadVideo [comfyui-animatediff]
sir can you do tutorial how to face swap batch of images ? grate video thanks sir
having this error
Error occurred when executing ReActorFaceSwap:
File "D:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 155, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 85, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 78, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-reactor-node
odes.py", line 112, in execute
script.process(
File "D:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-reactor-node\scripts
eactor_faceswap.py", line 96, in process
result = swap_face(
^^^^^^^^^^
File "D:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-reactor-node\scripts
eactor_swapper.py", line 195, in swap_face
source_faces = analyze_faces(source_img)
^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-reactor-node\scripts
eactor_swapper.py", line 114, in analyze_faces
face_analyser = copy.deepcopy(getAnalysisModel())
^^^^^^^^^^^^^^^^^^
File "D:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-reactor-node\scripts
eactor_swapper.py", line 69, in getAnalysisModel
ANALYSIS_MODEL = insightface.app.FaceAnalysis(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-reactor-node
eactor_log_patch.py", line 59, in patched_faceanalysis_init
assert 'detection' in self.models
^^^^^^^^^^^^^^^^^^^^^^^^^^
Hi, this is really an advanced error and is pretty impossible to me tell you exactly why is this happening, i can say that maybe there are some problems with face detection in the input image or in the one where the face should be placed but i'm not totally sure.Try to use some other images and see if it is keep happening, otherwise there are some problem with the model used... are you using the inswapper_128 or some other models? Tf you still can't solve it go ahead and post in the official reactor repository here: github.com/Gourieff/comfyui-reactor-node/issues
I do not have a facestore_models folder...I made one so fingers crossed...and it didn't work. I don't see the option for Reactor. I think where I went wrong, I used download manager to install it prior to all this.
there should be an error in the terminal when you start ComfyUI, what does it say?
@@DreamingAIChannel in the command window it says "[ReActor] - STATUS - Running v0.4.1-b8 in ComfyUI" which show it is okay. Inside comfy download manager it is red saying import failed. I had it working prior. Then closed comfyui. I got the video loader. I noticed my vid 2 vid failer in the download manger is red out then I noticed Reactor missing. I tried reinstalling too. Not sure what made it go away.
Update I got it working. I decided to get the Visual C++ build tools and open comfy ui and hit install. bat on the on the reactor file and hit refresh inside the download manager and it works again!
Where are the face models though? I cannot find the links
there is no FACERESTOR_MODEL folder.where can i find that?
Hi! If you don't have it just create it ant it should be fine
These video tutorials are great. Love the generated narration, his voice sounds irritated sometimes when explaining some stuff. What tool did you use to generate the voice narration?
Hi! I'm using Bark + RVC + A Lot of patience! xD I will make a video sooner or later!
Thanks for letting me know, I'll have to check it out@@DreamingAIChannel
Please help!! When I clicked "install.bat", it said "Embedded python not found. Please install manually". But when I tried to install "insightface-0.7.3-cp310-cp310-win_amd64.whl" into "ComfyUI_windows_portable\python_embeded", it said "ERROR: insightface-0.7.3-cp310-cp310-win_amd64.whl is not a supported wheel on this platform". What should I do to fix this?
Hi! If you are using the latest portable edition of comfyui you should use the cp311 and not the 310 of insightface
When trying to face swap two faces simultaneously, it now creates 2 separate videos with 1 different swapped face each. Any idea why that is?
What did you put in the source_faces_index field? If I'm remembering correctly you should put something like 0,1 if for example you have two faces to swap , "0,1,2" if you have three of them, and so on!
I put "0,1". So what it will do is faceswap face 0 alone and ignore face 1, then do the entire video again with face 1 and ignore face 0. Both videos will be output in a single file, and the 2nd will not have audio. The only solution I found is to make batch=1. Weird
I get a very weird python error:
ComfyUI\execution.py
line 154, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
TypeError: '>=' not supported between instances of 'int' and 'str'
During Interpolation step.
Tried to remove the interpolator and same error spawns on SaveVideo
Hi! I need the full log!
love your videos and explanation.
sometimes there’s a glitch in your voice. It kind of add a scifi touch as if you were in another dimension .... okay keep dreaming 😊
Can't get insight face to work anymore, I've installed visual studio 2022 and its up to date, I've installed VS C++ Build Tools and its up to date, I've also updated python and tried to use both versions of insightface.
It just doesn't show up in Comfy as a node.
without an error to analyze is really hard to help you!
@@DreamingAIChannel I managed to fix it by putting the insightface model into the python embeded folder and then opening the console and using this command
python -m pip install insightface-0.7.3-cp311-cp311-win_amd64.whl
I'm not sure why because I had everything installed but I kept getting an error saying no insightface found and I couldn't get reactor to load in Comfy but now I can.
Now I'm trying to do video to video but I just get 4 second videos.
Your guide was great, its the only one I've found so far that works other than the one step I had to change.
In the same way you upscale the output images couldn't you use a IPAdapter to apply a anime style then recombine to final video
Yes I think it will work, but it is not really a faceswap in that way. I need to know the starting anime face instead the only thing I can do is apply the IP adapter at the starting image hoping that the ipadapter will work in the same way even after the faceswap!
Hi, can you elaborate on ComfyUI N Nodes?
Yes, you need to watch my previous video: the video to video one and the GPT one :)
hi my comfyui missing "farmelnterpolator" what can i do?🙏
hi! you should look at the console, you probably have some errors there. If you want you can open an issue on GitHub with the full log that you have in your console
Is it possible to install ReActor on Linux?
I think that this will answer to your question: github.com/Gourieff/comfyui-reactor-node/issues/107
Comfyui-N-Nodes do not work together with Reactor-Nodes. For some reason when I put Reactor, the N-Nodes disappear
Uhm, weird... Do you see some errors in the console?
@@DreamingAIChannel I'm sorry to bother you, I somehow managed to fix it by downloading an old version of opencv-python, I honestly don't know how I did it as I'm very bad at it.
@@What_the_f4ck_is_a_kilometer don't worry! The important thing is that now is working for you!
i can't find image resize and frameinterpolator in the search bar. how should I proceed?
Hi! Image resize is from the WAS nodes, and the frame interpolator is from my custom nodes togheter with the video node (link in the description)
ty @@DreamingAIChannel
@@DreamingAIChannel Can we apply it to 2 people in a video with 2 pictures? Is this possible?
@@natura.teraphy.8d yes is possible! Forr the details you can refer to the reActor GitHub
can you make tutorial how to instal comfyui?
Hi! Here ruclips.net/video/ikYtP0ThN5A/видео.html
having this error
Error occurred when executing SaveVideo:
list index out of range
File "E:\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\execution.py", line 153, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "E:\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\execution.py", line 83, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "E:\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\execution.py", line 76, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "E:\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\execution.py", line 60, in slice_dict
d_new[k] = v[i if len(v) > i else -1]
~^^^^^^^^^^^^^^^^^^^^^^^^^
Hi! I need to see the rest of the log
@@DreamingAIChannel I've found the solution. By the way great video
Reactor in Comfyui now has a "face model in", it doesn't work without. Does anyone know how tocreate this model that I can load into the model loader?
why do you say " it doesn't work without"? i don't think is mandatory
@@DreamingAIChannel The plugin has changed since the video. The default model doesn't work without a "load face model" connected that you need with a model as well. If you swap the model to "inswapping128", it works even without connecting a model.
Hi, I know that the plugin has changed but it didn't break any compatibility since the input variable "face_model" is placed as optional field in the source code of the node .
I've just installed ReActor from zero to verify that, i don't know which other model do you have besize inswapping128 but know that inswapping128 is the default one.
I'm thinking that maybe was the other model that you had that was causing the issue, but i'm not sure 😵💫
@@DreamingAIChannel Hey that's great to hear. For me the default on different machines I installed it on was never inswapping128, so in case anyone has the same problem, this might be the solution. Keep up the great work!
Am I doing something wrong or does it just not work great with SDXL? My problem is, when I'm generating an image in 1024×1024 amd use the face swap, the replaced image is relatively blurry, or not only the face but also everything within the bounding box of the replaced area. Are there any settings to make it as sharp as the original image? I'm also experimenting with the Insight FaceSwap Bot on Discord, this produces much sharper results although it should be the same model...
nope, it doesn't matter the model you use since is acting after the generation of the image, did you enabled the facerestore model?
@@DreamingAIChannel Yes, I tried with and without. Of course I know the model doesn't actually matter, it's just that nowadays with SDXL most people are generating larger images than with the older models. And I've read somewhere this inswapper checkpoint or whatever this is (don't know all the technical terms) was trained on 128×128 pixels... which surely wouldn'tbe great for a standard 1024×1024 SDXL image I assume... unless the face is just in a very small area of the image, which is when I experienced less blurriness as most other times.
@@gordonbrinkmann i dont know the images that i've used in this video are 1024x1024 for the face and 1024x1600 for the output, i've made a mesurament now of the face of the output and is like 259x349, so maybe it could be that, as you said, if the face is too big the model is not acting correctly!
@@DreamingAIChannel So okay, it doesn't seem I could do anything about it apart from accepting it's not working as I hoped for. I have an image where the face area is about 140×230, there it's not noticeable, but I had a series of images with face being more like 600×750 or so (not the face alone, the rectangular part replaced by face swap), then it's very visible.
The worst part is not the face itself - I can use another face restore afterwards - but the face swap just replaces a rectangular part of the image around the face, which the face restore will of course not repair - because it's no face.
I found it very comfortable to put the face swap dafter the k sampler to directly swap the face in one go, seems like I have to go back to saving two images, one before swapping and one afterwards and replace the face manually to retain the unblurred surrounding of the face.
damn, i'm sorry man! I hope you'll find a different solution!
i cant find the code former file
You need to look in the "installation" section of the reActor github readme and click on the arrow where it says "Standalone (Portable) ComfyUI for Windows"
Hard to judge, if this is actually a "flawless faceswap" process, because you are using the face of a general model as a source that does not really have recognisable features.
reactor is a fu**ing nightmare to install (mac and windows), tried and retried but nothing...
Hi NUKED, How are you!
I am really exciting that your video can help to install ReActor Node.
But after I follow up all your instruction
When install batch file appear；
note: This error originates from a subprocess, and is likely not a problem with pip.
error: subprocess-exited-with-error
× Getting requirements to build wheel did not run successfully.
After restart ComfyUI appear:
ModuleNotFoundError: No module named 'insightface'
Cannot import F:\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-reactor-node module for custom nodes: No module named 'insightface'
(IMPORT FAILED): F:\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-reactor-node
Over try one week. Still can’t fix. Do know where to go! Cheers!
Hi,I'm fine, busy but fine! And you? I think you just need to install the right insightface because you probably have python version 3.11!
Yes, I have python version 3.11! Where to get right insightface ? Thanks for your help! Cheers!
Also I did this "python_embeded\python.exe -m pip install insightface-0.7.3-cp311-cp311-win_amd64.whl"
This is correct, it's giving you some errors?
@@DreamingAIChannel Hi NUKED,Thanks for your reply.
After I hit “install batch file” appear:
Installing...
Status: The 'insightface==0.7.3' distribution was not found and is required by the application
Collecting insightface==0.7.3
Using cached insightface-0.7.3.tar.gz (439 kB)
Building wheel for insightface (pyproject.toml) did not run successfully.
│ exit code: 1
╰─> [225 lines of output]
error: command 'C:\\Program Files (x86)\\Microsoft Visual Studio\\2022\\BuildTools\\VC\\Tools\\MSVC\\14.38.33130\\bin\\HostX86\\x64\\cl.exe' failed with exit code 2 [end of output]
After I restart “run_nvidia_gpu” appear:
0.0 seconds (IMPORT FAILED): C:\DISK D\AI\Stable Diffusion\ComfyUI\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-reactor-node
THANKS FOR YOUR HELP AGAIN!
crazy ai
Das Programm ist noch zu kompliziert für die Mehrheit.
cython error...
ERROR: insightface-0.7.3-cp310-cp310-win_amd64.whl is not a supported wheel on this platform. Edit: like stated in the update he pinned, used python 3.11 and it fixed my issues (also make sure you got the 3.12 pre built (works on 3.11)
Hi...
I got error when the proses on save the video.
there is message like:
Frames have been successfully reassembled into /Users/iccangninol/ComfyUI/temp/video.mp4
!!! Exception during processing!!! MoviePy error: the file /Users/iccangninol/ComfyUI/temp/video.mp4 could not be found!
Please check that you entered the correct path.
Traceback (most recent call last):
File "/Users/iccangninol/ComfyUI/execution.py", line 151, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
File "/Users/iccangninol/ComfyUI/execution.py", line 81, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
File "/Users/iccangninol/ComfyUI/execution.py", line 74, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
File "/Users/iccangninol/ComfyUI/custom_nodes/ComfyUI-N-Nodes/py/video_node_advanced.py", line 616, in save_video
video_clip = VideoFileClip(videos_output_temp_dir)
File "/Users/iccangninol/miniconda3/envs/Comfy2/lib/python3.10/site-packages/moviepy/video/io/VideoFileClip.py", line 88, in __init__
self.reader = FFMPEG_VideoReader(filename, pix_fmt=pix_fmt,
File "/Users/iccangninol/miniconda3/envs/Comfy2/lib/python3.10/site-packages/moviepy/video/io/ffmpeg_reader.py", line 35, in __init__
infos = ffmpeg_parse_infos(filename, print_infos, check_duration,
File "/Users/iccangninol/miniconda3/envs/Comfy2/lib/python3.10/site-packages/moviepy/video/io/ffmpeg_reader.py", line 270, in ffmpeg_parse_infos
raise IOError(("MoviePy error: the file %s could not be found!
"
OSError: MoviePy error: the file /Users/iccangninol/ComfyUI/temp/video.mp4 could not be found!
Please check that you entered the correct path.
Can you tell me, where wrong with my case?
showing this error ERROR: insightface-0.7.3-cp310-cp310-win_amd64 (1).whl is not a valid wheel filename.
Error occurred when executing LoadVideo [n-suite]:
'NoneType' object has no attribute 'shape'
File "C:\Users\Usuário\ComfyUI\execution.py", line 151, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Usuário\ComfyUI\execution.py", line 81, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Usuário\ComfyUI\execution.py", line 74, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Usuário\ComfyUI\custom_nodes\ComfyUI-N-Nodes\py\video_node_advanced.py", line 474, in encode
width, height = calc_resize_image(FRAMES[0], size, resize_by)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\Usuário\ComfyUI\custom_nodes\ComfyUI-N-Nodes\py\video_node_advanced.py", line 159, in calc_resize_image
height, width = image.shape[:2]
^^^^^^^^^^^
uhm, it's like it can't break down the video into pictures, can you try with another video?
nice very good! btw how to search for nodes in comfyui than the slow manual way? thx
update: i found out now, just double-click on an empty space of comfy! :)
@DreamingAI hi bro , by any chance u can take a look at my log?
Error occurred when executing LoadImage:
not a TIFF file (header b'n' not valid)
File "H:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 152, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "H:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 82, in get_output_data
return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "H:\AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 75, in map_node_over_list
results.append(getattr(obj, func)(**slice_dict(input_data_all, i)))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "H:\AI\ComfyUI_windows_portable\ComfyUI
odes.py", line 1486, in load_image
for i in ImageSequence.Iterator(img):
File "H:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\PIL\ImageSequence.py", line 56, in __next__
self.im.seek(self.position)
File "H:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\PIL\MpoImagePlugin.py", line 149, in seek
self._reload_exif()
File "H:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\PIL\Image.py", line 1459, in _reload_exif
self.getexif()
File "H:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\PIL\Image.py", line 1443, in getexif
self._exif.load(exif_info)
File "H:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\PIL\Image.py", line 3734, in load
self._info = TiffImagePlugin.ImageFileDirectory_v2(self.head)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "H:\AI\ComfyUI_windows_portable\python_embeded\Lib\site-packages\PIL\TiffImagePlugin.py", line 513, in __init__
raise SyntaxError(msg)
Hi! Well I think that your image is just invalid, use another image or convert this one in another format!
@@DreamingAIChannel tks so much for your help bro. the image went straight from my iphone (jpg) .I'll try to convert it to png then , tks again.
@@PHUONGNGUYEN-vp3lt 👍
Amazing video, thank you for the lessons. I am trying to install reactor node in Colab/ComfyUI, would you be able to write the prompt to donwload it automatically. e.g to download a lora you use
" !wget - c [link to download from civitai] -o ./models/ loras.... Thanks!
I have managed to get this done, but I can not see the option of "LOADVIDEO" any help with this? @dreamingAIChannel? Thank you