In-domain
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
in-domain (test)
0.11JSD
56
In-domain (test)
53.4Accuracy
50
In-domain (test)
66.6Accuracy
24
In-domain (test)
66.6Accuracy
16
In-domain Overall
0.861SRCC
15
In-Domain Multiple-Choice
92.9OCR-VQA Accuracy
15
In-domain (ID) Split
0.571mAP
14
In-domain (test)
86.37IoU
14
In-domain
14Win Rate
11
In-Domain (ID) (test)
90.71Overall Consistency Rate
10
In-domain (unseen)
57.24Exact Match (EM)
10
In-domain seen
74.05EM
10
In-Domain (ID) (val)
40.72BLEU
10
In-domain unseen
75.86EM
9
In-domain (seen)
85.95Exact Match (EM)
9
In-domain
7.354CE Score
8
In-domain
2.3Cost (Cost-first)
7
In-Domain
54Coherence
7
In-Domain
100Success Rate
6
In-Domain
96Success Rate
6
In-domain (test)
100OA
4
In-domain
64Average N
3
In-domain
0.027MSE
3
In-domain
0.005MSE
3
In-domain
0.107REL
3