Can this be used with just a text input instead of an image input?
#9
by mpeabody - opened
it seems like it might just be using CLIP as input, since the output video doesn't start with the image as a frame, so could we just pass a prompt as an input?
This model is not supported yet, but soon a 720P test-to-video model like this will be released. Please stay tuned for updates from us.
Is that raw 720p output, or 720p after an upscaling video2video step?