Fetching the paper…

Weakly-Supervised Video Object Grounding from Text by Loss Weighting and Object Interaction · Around