ClipSitu: Effectively Leveraging CLIP for Conditional Predictions in Situation Recognition

Benchmark Model Rank Results
grounded-situation-recognition-on-swigClipSitu#1Top-1 Verb: 58.19Top-1 Verb & Grounded-Value: 40.01
situation-recognition-on-imsituClipSitu#1Top-1 Verb: 47.23Top-1 Verb & Value: 29.73Top-5 Verbs: 85.69