Mask and Compress: Efficient Skeleton-based Action Recognition in Continual Learning

Open in new window