ML Research Wiki / Benchmarks / Image-to-Text Retrieval / WHOOPS!

WHOOPS!

Image-to-Text Retrieval Benchmark

Performance Over Time

đź“Š Showing 7 results | đź“Ź Metric: Specificity

Top Performing Models

Rank Model Paper Specificity Date Code
1 BLIP2 FlanT5-XXL (Text-only FT) đź“š Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 94.00 2023-03-13 -
2 BLIP2 FlanT5-XXL (Fine-tuned) đź“š Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 84.00 2023-03-13 -
3 BLIP2 FlanT5-XL (Fine-tuned) đź“š Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 81.00 2023-03-13 -
4 BLIP Large Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 77.00 2023-03-13 -
5 CoCa ViT-L-14 MSCOCO Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 72.00 2023-03-13 -
6 BLIP2 FlanT5-XXL (Zero-shot) Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 71.00 2023-03-13 -
7 CLIP ViT-L/14 Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 70.00 2023-03-13 -

All Papers (7)