Frame-wise Action Representations for Long Videos via Sequence Contrastive Learning - Citation Graph | Papersgraph