Support other T2V models?

[burnquiet]

Thanks for your greak work.
I tried to use an AnimateDiff model, but i failed to generate meaningful contents. Have you ever tried to apply your method to other models? (The model i use do not have temporal conv layer)

[Yuanshi9815]

Thank you for your feedback and for trying out the AnimateDiff model with our Video-Infinity method.

Unfortunately, due to the design of the AnimateDiff model, particularly thetemporal position embeddings, we haven't been successful in applying our Video-Infinity to it.

However, we are actively considering adapting Video-Infinity for use with other video generation models. If we manage to adapt it for AnimateDiff, we will update you immediately.