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

SciVidEval

Benchmarks

Task NameDataset NameSOTA ResultTrend
Video-Quiz EvaluationSciVidEval
VLM-as-Judge Score99.5
10
Visual Quality EvaluationSciVidEval
VLM-as-Judge Score10
9
Narration Quality EvaluationSciVidEval
Perplexity (PPL)16.67
8
Synchronization EvaluationSciVidEval
CLIP Score0.643
7
Showing 4 of 4 rows