Open Yangfan-96 opened 7 months ago
They are typically fixed for all these diffusion models trained on videos with specific resolutions and durations. Although they can be modified during inference, the performance cannot be guaranteed in this case.
They are typically fixed for all these diffusion models trained on videos with specific resolutions and durations. Although they can be modified during inference, the performance cannot be guaranteed in this case.