Few tips: 1- if you encounter this error: [Errno 2] No such file or directory....Download these models bit.ly/3LrGFVx and place under ComfyUI\models\liveportrait\base_models (you'll need to create the base_models folder) 2- You can actually use a non-square video, I recommend you use an image with the same aspect ratio if you come across other issues, please comment below👇or head over to my discord for more help: bit.ly/4bRF0DB
Hi there, I fixed several errors but in the end after installing inswapper_128.onnx manually I still got this error AttributeError: 'INSwapper' object has no attribute 'taskname' Cannot move beyond that, please help
@@MDMZ in insightface troubleshooting it says to ake it if you have another error (that I also had) so I downloaded it and was able to move to next node where I got this mistake( I went through 3 different tutorials to no avail
@@MDMZ Indeed, these 4 minute videos are the best content. Can you make any comfy UI video about doing pixar-like animation with consistent character for children stories?
it's installed and all but when I queue prompt it says It works! Thanks man! For some reason tho I only get 3 second videos now.... It started off good exporting the same duration but now it has this bug... hm....
@ its working now! Sometimes you just gotta reset everything and queue a prompt with the default setup (default images INCLUDED) to sort of reset the whole thing. Then it magically works fine again. Funny. But it works! Thank you so much. 😊
Nice video. Question: If I want to stop the outputted video from following the eyes from my driving video. How to I do it? The guy in my video is looking up all of the time reading lyrics. Thanks.
This is a great tutorial! Thank you so much! But I have one question: What is the advantage of using Comfy UI with Life Portrait? Why not just Life Portrait itself?
it's really just a choice, there are many people like me who have comfyui and like using it, you can also customize the workflow to your liking and get much more creative
@@MDMZ Yeah thx^^ but I will wait until there is a way to get it to run in realtime with webcam. The auto queing already works good but it's not usable for streams etc. At least for now :D
Hey, in CMD, after I downloaded the python version and put it in the root folder, I run the "python -m pip install -U pip", but it can't find the file.
Hey Brother... I did the same but I'm facing this error - 'NoneType' object is not subscriptable I dont know how to counter it... tried reinstalling older transformers (someone on reddit advised ) but that didn't help
Hello! Please tell me how to fix the error. (Error occurred when executing LivePortraitProcess: LivePortraitProcess.process() missing 1 required positional argument: 'crop_info') Thank you
Why is there always one node missing :( the Liveportraitnode stays red and when i install and update that node with the manager and restart it still comes out red. I followed the directions. idk what i did :(
So it looks as if the repo is down, or have a python conflict in local machine. But this module in particular cannot be installed properly. If anyone finds a solution please share
Error occurred when executing LivePortraitNode: No face detected in the source image! File "C:\Users\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\execution.py", line 151, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\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\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\execution.py", line 65, in map_node_over_list results.append(getattr(obj, func)(**input_data_all)) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-liveportrait odes\live_portrait.py", line 382, in run live_portrait_pipeline.execute(args) File "C:\Users\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-liveportrait odes\LivePortrait\src\live_portrait_pipeline.py", line 83, in execute crop_info = self.cropper.crop_single_image(img_rgb,src_face=crop_info) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-liveportrait odes\LivePortrait\src\utils\cropper.py", line 137, in crop_single_image raise Exception("No face detected in the source image!")
Yes I watched, Currently, when you input a face photo and a video with different facial expressions, the output is a video, right? What I'm asking is if you can input a face photo and another photo with a different facial expression, can the output be a photo with the changed facial expression?
Installing comfyui and getting it to work on mac is very tricky, that being said some people do manage to get it to work, but i heard it's very slow, if you plan to use it much, u almost definitely need and nvidia gpu
![Generate ENTIRE 3D SETS For Your Projects FAST! [FREE Blender + SDXL/FLUX Workflow]](http://i.ytimg.com/vi/U_NHijmIF3E/mqdefault.jpg)
![BOWKYLION - วิงวอน (ex-change) [Official MV]](http://i.ytimg.com/vi/cLdnZWDaO-w/mqdefault.jpg)
Few tips:
1- if you encounter this error: [Errno 2] No such file or directory....Download these models bit.ly/3LrGFVx
and place under ComfyUI\models\liveportrait\base_models (you'll need to create the base_models folder)
2- You can actually use a non-square video, I recommend you use an image with the same aspect ratio
if you come across other issues, please comment below👇or head over to my discord for more help: bit.ly/4bRF0DB
Brother, thank you for this. Grateful I found this comment
Hi there, I fixed several errors but in the end after installing inswapper_128.onnx manually I still got this error AttributeError: 'INSwapper' object has no attribute 'taskname'
Cannot move beyond that, please help
@@borisrin3775 hey, at which part did you decide you needed to install inswapper_128.onnx manually ? I don't remember having to do that
@@MDMZ in insightface troubleshooting it says to ake it if you have another error (that I also had) so I downloaded it and was able to move to next node where I got this mistake( I went through 3 different tutorials to no avail
@@borisrin3775 can you share your workflow and the input files on discord? i will look into it
Thanks, this does look like the best right now.
there's also Act one by RunwayML
Wow man, I learn with you soooooo many things, the best channel of youtube, thank you a lot, hugs from Brazil!!
Wow, thanks!
@@MDMZ Indeed, these 4 minute videos are the best content. Can you make any comfy UI video about doing pixar-like animation with consistent character for children stories?
Please make more tutorials on this for videos as well, this is truly amazing
More to come!
This is pure awesome. Thank you!
You're very welcome!
Man this is awesome, drop more sauce like this.
Thanks! Will do!
it's installed and all but when I queue prompt it says
It works! Thanks man! For some reason tho I only get 3 second videos now.... It started off good exporting the same duration but now it has this bug... hm....
that's weird, I will re-test this
@ its working now! Sometimes you just gotta reset everything and queue a prompt with the default setup (default images INCLUDED) to sort of reset the whole thing. Then it magically works fine again. Funny. But it works! Thank you so much. 😊
@@luci0c4s3 haha yep! that happens
Nice video. Question: If I want to stop the outputted video from following the eyes from my driving video. How to I do it? The guy in my video is looking up all of the time reading lyrics. Thanks.
I dont think u are able to do it with this workflow, following the facial expressions, lips and eye movements is exactly what makes this a good tool
This is a great tutorial! Thank you so much! But I have one question: What is the advantage of using Comfy UI with Life Portrait? Why not just Life Portrait itself?
it's really just a choice, there are many people like me who have comfyui and like using it, you can also customize the workflow to your liking and get much more creative
great video,, for AMD GPU? would be possible
yes, there's a guide on how to install comfy on AMD, check the comfyui installation page
Please please please do a video to video workflow!
noted!
Seems amazing bro. I'll definitely try it
Thanks for great video tutorial!!! But, i have some errors...
hi, check the pinned comment
are you doing this on mac ? or windows machine ? what pc spec is best/enough for AI image/videos to run smoothly? mac pro m1 16 GB is slow.
I'm doing it on windows: i9 14900, 4090 RTX
Hey I saw a lot of people having the settings for eyeretargeting etc in the LIVEPortrait Process tab. But I don't have this setting with webcam^^
they were probably using a different workflow, will see if i can make videos on more advanced use cases of LivePortrait
@@MDMZ Yeah thx^^ but I will wait until there is a way to get it to run in realtime with webcam. The auto queing already works good but it's not usable for streams etc. At least for now :D
thanks, bro. Make vid2vid version, please, it will be good continue
MDMZ Thank you! Is it possible to film yourself with the webcam and move the avatar in real time?
I haven't explored that yet, i'm sure something like that will be available soon, if not already
I’m looking for video clips with different facial expressions from your videos, please.
Need the v2v tutorial....that stuff is juicy
Noted!
Hey, in CMD, after I downloaded the python version and put it in the root folder, I run the "python -m pip install -U pip", but it can't find the file.
Forget this one. Fuck all that. Thank him later
th-cam.com/video/KP6kDbe_-e8/w-d-xo.html
Can I do this but with a live video camera?
there are few solutions that you can use for that, maybe i will cover it on my channel
Top
🙏
Could make a tutorial for Mac users?
I avoided covering this on the channel because it doesn't always work for everyone, i dont recommend running comfyui on a mac
Can it sync with live camera?
I believe it's possible with additional nodes
We need tutorial Bro❤❤❤❤
You're watching it 😂
I pressed Clear button and everything is different now! how do i get this setup back?
simply re-import the workflow
Hey Brother... I did the same but I'm facing this error - 'NoneType' object is not subscriptable
I dont know how to counter it... tried reinstalling older transformers (someone on reddit advised ) but that didn't help
can you share the video and image you're trying to use as input? drop them on discord
Hello! Please tell me how to fix the error.
(Error occurred when executing LivePortraitProcess:
LivePortraitProcess.process() missing 1 required positional argument: 'crop_info')
Thank you
hi, I didnt come across this one yet, check the pinned comment
@@MDMZ Thanks for your reply. Your scheme works.
Have a nice day.Thanks for your reply. Your scheme works.
Have a nice day.
Why is there always one node missing :(
the Liveportraitnode stays red and when i install and update that node with the manager and restart it still comes out red. I followed the directions. idk what i did :(
same here
Same here, tried to update, install again the module, but doesn't work... Unsure what to do.
So it looks as if the repo is down, or have a python conflict in local machine. But this module in particular cannot be installed properly. If anyone finds a solution please share
Error occurred when executing LivePortraitNode:
No face detected in the source image!
File "C:\Users\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\execution.py", line 151, in recursive_execute
output_data, output_ui = get_output_data(obj, input_data_all)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\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\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\execution.py", line 65, in map_node_over_list
results.append(getattr(obj, func)(**input_data_all))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-liveportrait
odes\live_portrait.py", line 382, in run
live_portrait_pipeline.execute(args)
File "C:\Users\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-liveportrait
odes\LivePortrait\src\live_portrait_pipeline.py", line 83, in execute
crop_info = self.cropper.crop_single_image(img_rgb,src_face=crop_info)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\User\Downloads\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-liveportrait
odes\LivePortrait\src\utils\cropper.py", line 137, in crop_single_image
raise Exception("No face detected in the source image!")
the error says "No face detected in the source image!"
velmi komplikovane
I don't need a video with changing expressions; I need a photo with a changed expression. Is this also possible?
I think you're describing what the video is about, did you watch ?
Yes I watched, Currently, when you input a face photo and a video with different facial expressions, the output is a video, right?
What I'm asking is if you can input a face photo and another photo with a different facial expression, can the output be a photo with the changed facial expression?
the missing nodes doesnt show in the manager, how i fix this?
make sure your comfyui is up to date
Bro, Change your youtube cover to CR7 screenshot and you sure get more views!
U mean the thumbnail ? I'm running an A/B test, the cr7 version isn't really winning 😅
Anyone else getting
Error occurred when executing LivePortraitNode:
Torch not compiled with CUDA enabled
Can Mac do this? Or what spec pc can make it work?
Installing comfyui and getting it to work on mac is very tricky, that being said some people do manage to get it to work, but i heard it's very slow, if you plan to use it much, u almost definitely need and nvidia gpu
How to make in mobile please one more video
Can you share workflow file json
hey, I explain where to find it in the video
Buddy, you know your links don't work?
they work for me, plz try another browser and lmk
NoneType' object is not subscriptable , Any one?
can u share more info on discord? including your workflow and input files
To complicated
one of the simplest comfyUI workflows out there