Image Captioning on COCO (val)
78.3CLIPScoreBLIP_FUSECAP
Evaluation Results
| Method | Links | |
|---|---|---|
| BLIP_FUSECAPImages=12M, Parameters=247M2023.05 | 78.3 | |
| BLIP2-G-OPT2.7Images=129M, Parameters=3.8B2023.05 | 77.8 | |
| GITImages=800M, Parameters=700M2023.05 | 77.1 | |
| PrismerImages=13M, Parameters=1.6B2023.05 | 76.7 | |
| OFAImages=20M, Parameters=470M2023.05 | 76.6 | |
| BLIP-LImages=129M, Parameters=470M2023.05 | 76.1 | |
| BLIP†Images=12M, Parameters=247M2023.05 | 75.2 |