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

SQA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Visual Question AnsweringSQA
Accuracy93.42
41
Sequential Question AnsweringSQA (test)
Accuracy (All)74.5
33
Visual Question AnsweringSQA-Image
Accuracy70.2
25
Question AnsweringSQA
Accuracy79.62
24
Science Question AnsweringSQA-I
Score79
24
ReasoningSQA
Accuracy85
23
Science Question AnsweringSQA IMG
Score97.67
23
Image-Language UnderstandingSQA
EM71.6
21
Science Question AnsweringSQA
Exact Match98.76
14
Table Question AnsweringSQA (test)
Accuracy (All)72.4
11
Table Question AnsweringSQA Perturbed (test)
Overall Accuracy0.723
8
Science Question AnsweringSQA
Accuracy (SQA)70.1
7
Science Question AnsweringSQA IMG
Accuracy70.7
7
Scholarly Question AnsweringSQA CS V2
Overall Score89.7
6
3D Visual Question AnsweringSQA (test)
EM@153.32
5
Sequential Question AnsweringSQA
Overall Accuracy74.5
5
Sequential Question AnsweringSQA first fold (dev)
Accuracy (ALL)68
5
Question AnsweringSQA (test)
MRR0.7957
4
Visual Question AnsweringSQA Short (test)
Accuracy94.8
2
3D Visual Question AnsweringSQA (val)
EM@152.05
1
Showing 20 of 20 rows