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

AVQA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Audio-Visual Question AnsweringAVQA
Accuracy87.3
85
Audio Visual Question AnsweringAVQA (test)
Total Accuracy94.3
36
Audio-Visual Question AnsweringAVQA (val)
Existence Accuracy88.24
9
Audio-Visual Question AnsweringAVQA (subset 2000 samples)
ASR Accuracy96.03
7
Audio Visual Question AnsweringAVQA
AVQA Clean Accuracy95.6
7
Audio-Visual Question AnsweringAVQA 69 (test)
Accuracy93.8
5
Showing 6 of 6 rows