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

NLVR

Benchmarks

Task NameDataset NameSOTA ResultTrend
Visual ReasoningNLVR v1 (Test-U)
Accuracy73
8
Visual Question AnsweringNLVR
Accuracy (NLVR)74
5
Semantic ParsingNLVR (test-P)
Accuracy86.3
5
Semantic ParsingNLVR (dev)
Accuracy89.6
5
GeneralNLVR2
Score83.2
3
Semantic ParsingNLVR (test-h)
Accuracy89.5
3
Visual ReasoningNLVR (val)
Accuracy75.06
2
Visual ReasoningNLVR (test)
Accuracy75.33
2
Showing 8 of 8 rows