Description
Version 2 available at https://youtu.be/NqnveRSDLRM
FAQ
Comments (26)
Hello,
Thank you for the workflow; it's working great on my RTX 3080 10GB. ❤️
I have a question: how do I choose the starting point for the reference video? I can't find the option to adjust that.
thank you for the WF.
the WanAnimateToVideo node is always missing no matter what I do, is there any solution please ?
Edit: I switched to the nighty version of confyui from the manager and it worked
How did you manage to change the version? I did it from the administrator, but when I reopen ComfyUI, that node still doesn't appear. Can you help me, please?
@CrishxExtrem go to comfyui manager, on the left side of the window you will see comfyui stable version, change it to nightly, then go to custom nodes manager in comfyui, find wanvideowrapper, click change version and choose nightly, restart comfyui
Imagine if they focused on a useful model.
A damn shame on the team for coming up with this.
Traceback (most recent call last):
File "C:\Users\crz\Downloads\ComfyUI_windows_portable_NEW\ComfyUI\nodes.py", line 2133, in load_custom_node
module_spec.loader.exec_module(module)
File "<frozen importlib._bootstrap_external>", line 999, in exec_module
File "<frozen importlib._bootstrap>", line 488, in callwith_frames_removed
File "C:\Users\crz\Downloads\ComfyUI_windows_portable_NEW\ComfyUI\custom_nodes\comfyui-apex-artist\__init__.py", line 2, in <module>
from .apex_depth_to_normal import ApexDepthToNormal
File "C:\Users\crz\Downloads\ComfyUI_windows_portable_NEW\ComfyUI\custom_nodes\comfyui-apex-artist\apex_depth_to_normal.py", line 75
}), "required": {
^
SyntaxError: closing parenthesis ')' does not match opening parenthesis '{' on line 65
Yo this whole py file is a mess, go look at it
try update
TypeError: WanAnimateToVideo.execute() got an unexpected keyword argument 'background_video'
KSampler
The size of tensor a (21) must match the size of tensor b (20) at non-singleton dimension 2
u need to use pixel perfect resolution
How do you instruct this workflow to render the rest of a video instead of just 81 frames? it doesnt move to the next sets of 81 frames like the Kijais workflow
Yeah.... just what i thought this workflow is bogus, it can only do 81 frames or whatever your vmem can handle, it hasn't the infinite length capability and the image resize node isn't recognized in the custom nodes as the Youtube vid promises.
Also the Wan Animate Lora doesnt load without errors.
If you want to drop a zeroday bombshell do it properly..
the native node supports extended frame i thot it was too easy to include
Does not work
why "Given groups=1, weight of size [5120, 36, 1, 2, 2], expected input[1, 16, 22, 80, 80] to have 36 channels, but got 16 channels instead" and this is after wananimate embeds and on wanvideo ksampler, literally the last step
i change dimensions nothing works
i had similar problem, if i remeber ok, it was vae file, try chage it for different
need help with lip sync, also I cant find the 2.2 fb16 models, only 8
dont work
The workflow completely ignores the reference image. So it's a great workflow for anyone who wants to use a video for input, and get the exact same video as output.
You probably miss a step .. it should work with the reference
This is also my experience. Reference image ignored.
And I don't know what step I could have missed. Where are the steps to follow?
I agree, where's the thumbs down on this clown.
Why is this WF such a disorganized mess ?
because you're dumb and should learn how to use it
A workflow that does NOTHING except render same video out. Why do you morons do this? Waste our time?