Hi @LinB203 , just want to bring VSTAR: Generative Temporal Nursing for Longer Dynamic Video Synthesis to your attention, where the temporal attention mechanism is thoroughly examined. Thus, some regularization loss might be helpful to learn longer video synthesis. Looking forward to your thoughts on this :)
Hi @LinB203 , just want to bring VSTAR: Generative Temporal Nursing for Longer Dynamic Video Synthesis to your attention, where the temporal attention mechanism is thoroughly examined. Thus, some regularization loss might be helpful to learn longer video synthesis. Looking forward to your thoughts on this :)