Fetching the paper…

HACS: Human Action Clips and Segments Dataset for Recognition and Temporal Localization · Around