Going Beyond Nouns With Vision & Language Models Using Synthetic Data

Benchmark Model Rank Results
visual-reasoning-on-winogroundsyn-CLIP#48Text Score: 30.00Image Score: 11.50Group Score: 9.50
visual-reasoning-on-winogroundsyn-CyCLIP#49Text Score: 30.00Image Score: 10.75Group Score: 8.25
visual-reasoning-on-winogroundCyCLIP#51Text Score: 28.50Image Score: 9.50Group Score: 7.25