somuchtome

Hi~ Is there AnyV2V solution based on "animatediff"? I am not sure which layers' features should be extracted and at which timesteps the feature should be injected