CRRL Implementation for Contrast-Reconstruction Representation Learning for Self-Supervised Skeleton-Based Action Recognition. Network Architecture Visualization Requirements Python 3.6+ Pytorch 1.0+ Torchvision 0.4+