Metric: SPICE (higher is better)
| # | Model↕ | SPICE▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | GIT2 | 16.06 | No | GIT: A Generative Image-to-text Transformer for ... | 2022-05-27 | Code |
| 2 | GIT | 15.94 | No | GIT: A Generative Image-to-text Transformer for ... | 2022-05-27 | Code |
| 3 | Microsoft Cognitive Services team | 14.85 | Yes | Scaling Up Vision-Language Pre-training for Imag... | 2021-11-24 | - |
| 4 | VLAF2 | 14.71 | No | - | - | - |
| 5 | Human | 14.67 | No | - | - | - |
| 6 | Microsoft Cognitive Services team | 14.04 | No | VIVO: Visual Vocabulary Pre-Training for Novel O... | 2020-09-28 | - |
| 7 | test_cbs2 | 12.74 | No | - | - | - |
| 8 | icp2ssi1_coco_si_0.02_5_test | 11.84 | No | - | - | - |
| 9 | UpDown + ELMo + CBS | 11.2 | No | - | - | - |
| 10 | UpDown | 10.14 | No | - | - | - |
| 11 | Neural Baby Talk + CBS | 9.69 | No | - | - | - |
| 12 | Neural Baby Talk | 9.15 | No | - | - | - |