Visual Question Answering on Conflict-VQA
79.34AccuracyCALRD
Evaluation Results
| Method | Links | |
|---|---|---|
| CALRDDecoding Strategy=Beam, Model=Qwen3-VL2026.06 | 79.34 | |
| CALRDDecoding Strategy=Greedy, Model=Qwen3-VL2026.06 | 77.42 | |
| CALRDDecoding Strategy=Nucleus, Model=Qwen3-VL2026.06 | 76.55 | |
| OPERADecoding Strategy=Beam, Model=Qwen3-VL2026.06 | 76.43 | |
| DoLaDecoding Strategy=Greedy, Model=Qwen3-VL2026.06 | 76.06 | |
| VanillaDecoding Strategy=Beam, Model=Qwen3-VL2026.06 | 75.94 | |
| VanillaDecoding Strategy=Greedy, Model=Qwen3-VL2026.06 | 75.58 | |
| VCDDecoding Strategy=Nucleus, Model=Qwen3-VL2026.06 | 75.5 | |
| DeCoDecoding Strategy=Beam, Model=Qwen3-VL2026.06 | 75.46 | |
| DeCoDecoding Strategy=Greedy, Model=Qwen3-VL2026.06 | 75.33 | |
| DeCoDecoding Strategy=Nucleus, Model=Qwen3-VL2026.06 | 74.79 | |
| VanillaDecoding Strategy=Nucleus, Model=Qwen3-VL2026.06 | 73.39 | |
| CALRDDecoding Strategy=Beam, Model=Qwen2.5-VL2026.06 | 70.25 | |
| CALRDDecoding Strategy=Greedy, Model=Qwen2.5-VL2026.06 | 70.15 | |
| CALRDDecoding Strategy=Nucleus, Model=Qwen2.5-VL2026.06 | 68.91 | |
| OPERADecoding Strategy=Beam, Model=Qwen2.5-VL2026.06 | 68.49 | |
| DoLaDecoding Strategy=Greedy, Model=Qwen2.5-VL2026.06 | 66.22 | |
| VanillaDecoding Strategy=Greedy, Model=Qwen2.5-VL2026.06 | 65.61 | |
| VanillaDecoding Strategy=Beam, Model=Qwen2.5-VL2026.06 | 65.49 | |
| VCDDecoding Strategy=Nucleus, Model=Qwen2.5-VL2026.06 | 64.04 | |
| DeCoDecoding Strategy=Greedy, Model=Qwen2.5-VL2026.06 | 63.79 | |
| DeCoDecoding Strategy=Beam, Model=Qwen2.5-VL2026.06 | 63.55 | |
| DeCoDecoding Strategy=Nucleus, Model=Qwen2.5-VL2026.06 | 63.55 | |
| VanillaDecoding Strategy=Nucleus, Model=Qwen2.5-VL2026.06 | 61.48 | |
| CALRDDecoding Strategy=Beam, Model=LLaVA-1.62026.06 | 50.79 | |
| CALRDDecoding Strategy=Greedy, Model=LLaVA-1.62026.06 | 49.51 | |
| CALRDDecoding Strategy=Nucleus, Model=LLaVA-1.62026.06 | 48.72 | |
| VCDDecoding Strategy=Nucleus, Model=LLaVA-1.62026.06 | 46.35 | |
| DeCoDecoding Strategy=Nucleus, Model=LLaVA-1.62026.06 | 45.44 | |
| DoLaDecoding Strategy=Greedy, Model=LLaVA-1.62026.06 | 44.59 | |
| CALRDDecoding Strategy=Beam, Model=InstructBLIP2026.06 | 44.49 | |
| CALRDDecoding Strategy=Beam, Model=LLaVA-1.52026.06 | 44.47 | |
| CALRDDecoding Strategy=Greedy, Model=InstructBLIP2026.06 | 44.12 | |
| VanillaDecoding Strategy=Beam, Model=LLaVA-1.62026.06 | 43.99 | |
| OPERADecoding Strategy=Beam, Model=LLaVA-1.62026.06 | 43.74 | |
| CALRDDecoding Strategy=Nucleus, Model=LLaVA-1.52026.06 | 43.4 | |
| DeCoDecoding Strategy=Greedy, Model=LLaVA-1.62026.06 | 43.38 | |
| DeCoDecoding Strategy=Beam, Model=LLaVA-1.62026.06 | 43.01 | |
| VCDDecoding Strategy=Nucleus, Model=LLaVA-1.52026.06 | 42.4 | |
| VanillaDecoding Strategy=Greedy, Model=LLaVA-1.62026.06 | 42.35 | |
| VanillaDecoding Strategy=Nucleus, Model=LLaVA-1.62026.06 | 41.92 | |
| CALRDDecoding Strategy=Greedy, Model=LLaVA-1.52026.06 | 41.9 | |
| DeCoDecoding Strategy=Greedy, Model=InstructBLIP2026.06 | 41.31 | |
| DeCoDecoding Strategy=Beam, Model=InstructBLIP2026.06 | 40.83 | |
| OPERADecoding Strategy=Beam, Model=InstructBLIP2026.06 | 40.12 | |
| DoLaDecoding Strategy=Greedy, Model=LLaVA-1.52026.06 | 40.1 | |
| VanillaDecoding Strategy=Beam, Model=InstructBLIP2026.06 | 39.98 | |
| VanillaDecoding Strategy=Greedy, Model=InstructBLIP2026.06 | 39.61 | |
| OPERADecoding Strategy=Beam, Model=LLaVA-1.52026.06 | 39 | |
| VanillaDecoding Strategy=Beam, Model=LLaVA-1.52026.06 | 38.88 | |
| VanillaDecoding Strategy=Nucleus, Model=LLaVA-1.52026.06 | 38.76 | |
| DeCoDecoding Strategy=Nucleus, Model=LLaVA-1.52026.06 | 38.45 | |
| DeCoDecoding Strategy=Beam, Model=LLaVA-1.52026.06 | 38.15 | |
| DeCoDecoding Strategy=Greedy, Model=LLaVA-1.52026.06 | 38.03 | |
| VanillaDecoding Strategy=Greedy, Model=LLaVA-1.52026.06 | 36.2 | |
| DoLaDecoding Strategy=Greedy, Model=InstructBLIP2026.06 | 32.72 | |
| CALRDDecoding Strategy=Nucleus, Model=InstructBLIP2026.06 | 28.07 | |
| DeCoDecoding Strategy=Nucleus, Model=InstructBLIP2026.06 | 24.91 | |
| VanillaDecoding Strategy=Nucleus, Model=InstructBLIP2026.06 | 23.94 | |
| VCDDecoding Strategy=Nucleus, Model=InstructBLIP2026.06 | 23.05 |