Error in CogVideoX Image-to-Video

[vladmandic]

Discussed in #3625

^{Originally posted by holgerpieta December 9, 2024}
Hello Experts

I'm trying to run CogVideoX Image-to-Video, but I seem to be doing something wrong.
Here is what I do:

• Start SD.Next
• Select any base model (not important for the problem)
• Select "Image" Tab
• Input some prompt (again not important)
• Drag and drop an image file onto the left box
• Open "Resize": "Fixed", "Nearest", "Fixed", 512x512
• Script "Video CogVideoX"
• Model: "THUDM/CogVideoX-5b-I2V"
• Generate

The result in an error message: CogVideoX: model=THUDM/CogVideoX-5b-I2V image input not supported
I've traced the message to line 187 in file scripts/cogvideo.py
Looking at that I think the problem is that the run function is called with an empty image argument.

So I presume I did something wrong, clicked the wrong buttons or the didn't click the right buttons, causing the image input to be empty.

Any hints what I did wrong?

[vladmandic]

this issue is resolved, the remaining one is tracked under #3639