paper-with-me

Image Captioning 벤치마크

Image Captioning on WHOOPS!

6개 결과 · ⬇ CSV · JSON

BLEU-4

0 10.5 21 31.5 42 2023-03 2026-11 BLIP2 FlanT5-XXL (Fine-tuned) — 42.0 (2023-03-13) BLIP2 FlanT5-XL (Fine-tuned) — 41.0 (2023-03-13) BLIP2 FlanT5-XXL (Zero-Shot) — 31.0 (2023-03-13) CoCa ViT-L-14 MSCOCO — 25.0 (2023-03-13) BLIP Large — 13.0 (2023-03-13) OFA Large — 0.0 (2023-03-13) BLIP2 FlanT5-XXL (Fine-tuned) — 42.0 (2023-03-13)
RankModel BLEU-4CIDEr Extra Training Data PaperCodeYear
1 BLIP2 FlanT5-XXL (Fine-tuned) 42177 ✓ Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 2023
2 BLIP2 FlanT5-XL (Fine-tuned) 41174 ✓ Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 2023
3 BLIP2 FlanT5-XXL (Zero-Shot) 31120 Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 2023
4 CoCa ViT-L-14 MSCOCO 25102 Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 2023
5 BLIP Large 1365 Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 2023
6 OFA Large 00 Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images 2023
1–6 / 6 페이지당 10 20 50 100