I am running LTX2.3 Image-to-Video workflow in ComfyUI.
The workflow comes with a pre-loaded Reference Image node that has a child's drawing of a woman. There is no way to turn this off. There is even a baleen that enables "text to video" and is set as true. Every prompt I have tried results in some sort of woman sitting on a couch in a business suit. The output is wildly different from my prompt.
Workflow seems broken.
I am running LTX2.3 Image-to-Video workflow in ComfyUI.
The workflow comes with a pre-loaded Reference Image node that has a child's drawing of a woman. There is no way to turn this off. There is even a baleen that enables "text to video" and is set as true. Every prompt I have tried results in some sort of woman sitting on a couch in a business suit. The output is wildly different from my prompt.
Workflow seems broken.