Innovate Futures @ Benji
patreon
Hunyuan Video V2V Flow Edit Ver 20250113
🕑 Added 2025-01-13 13:00:18 +0000 UTCReferring to the video tutorial : https://www.patreon.com/posts/119902314
Video: https://youtu.be/TAdO9ZPpzww
To Patreon Supporters:
I tidied up the flow edit workflow overall for something of what I think is working better.
Also, I have add the image reference for the source and target conditions, where we are able to restyle the output image by using a reference image. Similar concept as IP adapter before.
And why I like to use the native model loader and clip loader encoder in ComfyUI to run Han Yuan video is because it just performed natively with less conflict in the Python library, and I see the generation time in the system is way shorter than using the wrapper method.
I have also demonstrated this workflow in the video tutorial in YouTube, so go check it out if you not understand how to do the setting.
Find the attachment below for the workflow zip file and enjoy :)
Comments
Benjamin Law
Should set your comfyui run with cuda 0 , which is your gpu or graphics card.
WinkJP
# ComfyUI Error Report ## Error Details - **Node ID:** 86 - **Node Type:** HyVideoFlowEditSamplerWrapper - **Exception Type:** RuntimeError - **Exception Message:** Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu! ## Stack Trace
WinkJP
HI, I got this error.
