Image Captioning on OmniDrive
86.57CIDErVGGDrive
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| VGGDriveEvaluation Protocol=Fine-tuned2026.02 | 86.57 | 37.58 | 34.4 | 52.85 | |
| VGGT-DistEvaluation Protocol=Fine-tuned2026.02 | 86.41 | 37.28 | 34.32 | 52.67 | |
| VGGT-AddEvaluation Protocol=Fine-tuned2026.02 | 86.38 | 37.23 | 34.26 | 52.62 | |
| BaselineEvaluation Protocol=Fine-tuned2026.02 | 86.29 | 37.29 | 34.33 | 52.64 | |
| HERMESEvaluation Protocol=Fine-tuned2026.02 | 74.1 | — | 32.7 | — | |
| OmniDriveEvaluation Protocol=Zero-shot2026.02 | 68.6 | 38 | 32.6 | 46.4 | |
| RoboTronEvaluation Protocol=Zero-shot2026.02 | 34.33 | 20.3 | 23.67 | 26.1 | |
| LLAVA-OVEvaluation Protocol=Zero-shot2026.02 | 28.41 | 16.14 | 22.14 | 22.23 | |
| GPT-4oEvaluation Protocol=Zero-shot2026.02 | 24.42 | 10.91 | 22.34 | 19.22 |