New top story on Hacker News: Vid2Seq: A pretrained visual language model for describing multi-event videos

Vid2Seq: A pretrained visual language model for describing multi-event videos
12 by og_kalu | 2 comments on Hacker News.