Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Hard Positive Compositional Benchmark

Benchmarks

Task NameDataset NameSOTA ResultTrend
Visio-linguistics CompositionalityHard Positive Compositional Benchmark SWAP (test)
Test Accuracy (Original)66.1
3
Visio-linguistics CompositionalityHard Positive Compositional Benchmark (test)
Original Test Accuracy68.3
3
Showing 2 of 2 rows