spatial-temporal-action-detection
OpenMMLab's Next Generation Video Understanding Toolbox and Benchmark