Image Captioning on CHAIR (test)
52.6Cs ScoreVanilla
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| VanillaBackbone=LLaVA-13B2025.11 | 52.6 | 14.1 | 76.4 | |
| VAFBackbone=LLaVA-7B2025.11 | 52.4 | 14.6 | 75.3 | |
| VanillaBackbone=LLaVA-7B2025.11 | 52.2 | 13.9 | 76.1 | |
| VAFBackbone=LLaVA-13B2025.11 | 49.6 | 13.6 | 76.5 | |
| TARACBackbone=LLaVA-7B2025.11 | 43 | 11.2 | 75.1 | |
| TARACBackbone=LLaVA-13B2025.11 | 43 | 11.2 | 75.1 | |
| PAIBackbone=LLaVA-13B2025.11 | 40.2 | 10.8 | 76.9 | |
| TARACBackbone=Qwen2.5-VL2025.11 | 39.8 | 9.8 | 74.6 | |
| VanillaBackbone=Qwen2.5-VL2025.11 | 39.2 | 9.7 | 75 | |
| VAFBackbone=Qwen2.5-VL2025.11 | 38.2 | 9.8 | 74.8 | |
| PAICDBackbone=LLaVA-13B2025.11 | 37 | 9.9 | 77.3 | |
| PAIBackbone=Qwen2.5-VL2025.11 | 36 | 10.3 | 71.9 | |
| VGABackbone=Qwen2.5-VL2025.11 | 35.4 | 9.4 | 75.1 | |
| VanillaBackbone=LLaVA-Next2025.11 | 34.2 | 9.3 | 72.2 | |
| PAIBackbone=LLaVA-Next2025.11 | 34 | 8.9 | 73.7 | |
| PAIBackbone=LLaVA-7B2025.11 | 33.6 | 9 | 77.1 | |
| VAFBackbone=LLaVA-Next2025.11 | 32.8 | 8.8 | 72.6 | |
| PAICDBackbone=LLaVA-7B2025.11 | 30.8 | 8.1 | 76.3 | |
| VGABackbone=LLaVA-13B2025.11 | 30 | 8.2 | 77.1 | |
| VGABackbone=LLaVA-7B2025.11 | 29.8 | 8.2 | 76.3 | |
| TARACBackbone=LLaVA-Next2025.11 | 29.8 | 8 | 71.5 | |
| VGABackbone=LLaVA-Next2025.11 | 28.6 | 8 | 71.6 |