Multimodal Question Answering on Recap-COCO
65.94CPTCAP (Ours)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TCAP (Ours)Model=Qwen3-VL2026.01 | 65.94 | 0 | |
| Vanilla FTModel=InternVL2.52026.01 | 65.35 | 90.17 | |
| BYEModel=Qwen3-VL2026.01 | 65.07 | 87.4 | |
| Random DropModel=InternVL2.52026.01 | 64.59 | 89.94 | |
| TCAP (Ours)Model=LLaVA-NeXT2026.01 | 64.46 | 3.4 | |
| Vanilla FTModel=LLaVA-NeXT2026.01 | 64.02 | 92.83 | |
| SampDetoxModel=Qwen3-VL2026.01 | 63.47 | 75.36 | |
| BYEModel=LLaVA-NeXT2026.01 | 63.38 | 94.13 | |
| Random DropModel=LLaVA-NeXT2026.01 | 62.61 | 94.42 | |
| Vanilla FTModel=Qwen3-VL2026.01 | 62.55 | 93.5 | |
| TCAP (Ours)Model=InternVL2.52026.01 | 61.19 | 0.17 | |
| Random DropModel=Qwen3-VL2026.01 | 61.1 | 91.36 | |
| BYEModel=InternVL2.52026.01 | 60.91 | 54.13 | |
| SampDetoxModel=InternVL2.52026.01 | 60.25 | 79.16 | |
| SampDetoxModel=LLaVA-NeXT2026.01 | 54.39 | 77.11 |