TraVLR
material
ID: travlr
Related Papers
TraVLR: Now You See It, Now You Don’t! A Bimodal Dataset for Evaluating Visio-Linguistic Reasoning
AbstractNumerous visio-linguistic (V+L) representation learning methods have been developed, yet existing datasets do not adequately evaluate the extent to which they represent visual and linguistic concepts in a unified space. …...