BLIP-2 (ViT-L, fine-tuned)

Reported on 3 benchmarks across 1 task · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing3 results

Image-to-Text RetrievalonCOCO (Common Objects in Context)
Recall@1· 2023-01-30
83.5
best: 85.4 (BLIP-2 (ViT-G, fine-tuned))
BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models arXiv:2301.12597
Image-to-Text RetrievalonCOCO (Common Objects in Context)
Recall@10· 2023-01-30
98
best: 99.8 (Oscar)
BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models arXiv:2301.12597
Image-to-Text RetrievalonCOCO (Common Objects in Context)
Recall@5· 2023-01-30
96
best: 97 (BLIP-2 (ViT-G, fine-tuned))
BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models arXiv:2301.12597