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

private CCTV

Benchmarks

Task NameDataset NameSOTA ResultTrend
Open-ended Reasoningprivate CCTV 80 videos (test)
Rescaled BertScore F139.47
6
Verificationprivate CCTV 80 videos (test)
Accuracy85
6
Multiple Choice Questioning (MCQ)private CCTV 80 videos (test)
Accuracy (MCQ Test)88.75
6
Showing 3 of 3 rows