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

Quilt

Benchmarks

Task NameDataset NameSOTA ResultTrend
Text-to-Image RetrievalQUILT-1M (holdout set)
R@16.27
14
Medical Visual Question AnsweringQuilt-1M
Accuracy73.74
11
Text-Image RetrievalQuilt
Recall@126.88
10
Image-Text RetrievalQuilt
R@132.46
10
Metric Sensitivity AnalysisQuilt-1M Visual Hallucination
Performance Score0.9
5
Showing 5 of 5 rows