diagnostic
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Diagnostic vctr (G5)
27Runtime (s)
9
Diagnostic normacc (G4)
16Runtime (s)
9
diagnostic (test-unseen)
85.21Accuracy@50
9
diagnostic seen (test)
84.97Acc@50
9
diagnostic (test-unseen)
80.23Acc@50
9
diagnostic (test-seen)
82.02Accuracy@50
9
Diagnostic (Avg. YouCook2, COIN, CrossTask) (test)
76.92State Accuracy
8
Diagnostic Average of YouCook2, COIN, CrossTask
53.81State Accuracy
5