Questions about random_ref_frame and local_image

[linjing7]

Hi, thanks for your great work! UniAnimate is the best work I know for human image animation! I have two questions regarding the random_ref_frame, vit_frame, and local_image:

1. In unet_unianimate.py, I notice that you have performed the dropout operation for local_image and dwpose, but do not dropout random_ref_frame. But during inference, we can choose to set random_ref_frame=None. Does it mean you randomly set random_ref_frame to None during training?
2. In UniAnimate_infer_long.yaml, I notice that you did not include local_image in 'partial_keys'. Does it mean that the first-frame strategy would not make a significant difference in this setting?