Wan 2.2 Temporal Context

[solss-bit]

With the addition of Context Options allowing for longer t2v and i2v video generations, It would be nice to be able to prompt for different segments of the video, even if it isn't exact. I know there are ways to extract last frames, continue from where you left off, or through a series of spaghetti layouts, it could probably be done.

I prefer using native nodes to more experimental WanVideoWrapper -- although I appreciate the experimentation. I also realize context options are in beta at the moment. Kijai has added to his clip text encoder the option to use | pipe symbol to use multiple prompts for a single generation. I hope native ComfyUI nodes can implement a similar feature in the future as it would help to make use of longer video generations. At the present, Most of my experiments with context options have resulted in longer videos that don't deviate but ever so slightly throughout the generation.

Thank you!