Innovate Futures @ Benji
patreon
[Quick Update Workflow] CogVideoX Video Extend - Ver.20241121 (For Wrapper New Nodes Update)
🕑 Added 2024-11-20 21:33:33 +0000 UTC23 Dec, 2024
***Workflow update for CogVideoX node connection fixed.
ComfyUI CogVideoX Wrapper node just got an update, and I did a quick change after the last video talk about Video Extend : https://www.patreon.com/posts/comfyui-and-ai-116364425
(update: 2024-11-25)
Download the latest model, VAE from here :
https://huggingface.co/Kijai/CogVideoX-comfy/tree/main
If you have just check out that video and you have downloaded the ComfyUI-CogVideoXWrapper latest Update8.
Then use this version of workflow.
This version also works for model 5B 1.5 ( the latest version of CogVideoX model)
New video talk about 5B 1.5 will be post later this week.
See yall!
Comments
Benjamin Law
Wheres your CogvideoX VAE? Anyway, I will make a video about new Cog 1.5 and Wrapper update8 version later.
Jonathan Oh
File "D:\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-CogVideoXWrapper\pipeline_cogvideox.py", line 500, in __call__ latent_padding = torch.zeros(padding_shape, device=device, dtype=self.vae_dtype) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ RuntimeError: Trying to create tensor with negative dimension -1: [1, -1, 16, 60, 90] I changed model loader as you mentioned and this is my first time to run. Do you have any idea for this? Thank you in advance:)
