Catplayinginthesnow: Impact of Prior Segmentation on a Model of Visually Grounded Speech

41605

Knowledge Triples