Audio-Video Question Answering on MUSIC-AVQA
51.77AV Temporal AccMMER
Evaluation Results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MMERMethod Type=Training-free Model Merging2025.05 | 51.77 | — | — | — | — | — | — | — | — | — | — | — | — | 47.25 | 51.27 | |
| PCB-MergingMethod Type=Training-free Model Merging2025.05 | 49.67 | — | — | — | — | — | — | — | — | — | — | — | — | 44.32 | 50.24 | |
| TIESMethod Type=Training-free Model Merging, DARE=true2025.05 | 48.98 | — | — | — | — | — | — | — | — | — | — | — | — | 43.78 | 49.54 | |
| TIESMethod Type=Training-free Model Merging2025.05 | 48.6 | — | — | — | — | — | — | — | — | — | — | — | — | 43.27 | 49.27 | |
| X-InstructBLIPMethod Type=Training-based Multimodal Baseline2025.05 | 48.34 | — | — | — | — | — | — | — | — | — | — | — | — | 40.71 | 41.23 | |
| NaiveMCMethod Type=Training-free Model Merging, DARE=true2025.05 | 47.67 | — | — | — | — | — | — | — | — | — | — | — | — | 39.78 | 47.98 | |
| NaiveMCMethod Type=Training-free Model Merging2025.05 | 47.4 | — | — | — | — | — | — | — | — | — | — | — | — | 39.31 | 47.65 | |
| TAMethod Type=Training-free Model Merging, DARE=true2025.05 | 46.92 | — | — | — | — | — | — | — | — | — | — | — | — | 40.46 | 47.98 | |
| TAMethod Type=Training-free Model Merging2025.05 | 46.7 | — | — | — | — | — | — | — | — | — | — | — | — | 40.22 | 47.97 | |
| ImageBind-LLMMethod Type=Training-based Multimodal Baseline2025.05 | 39.72 | — | — | — | — | — | — | — | — | — | — | — | — | 36.54 | 38.76 | |
| CREMAModality=V, A, F, D, N, Trainable Params.=38M, GFLOPS=1.84K2024.02 | 0.735 | 0.883 | 0.647 | 0.832 | 0.862 | 0.913 | 0.887 | 0.849 | 0.749 | 0.852 | 0.686 | 0.777 | 0.817 | — | — | |
| CREMAModality=V, A, F, Trainable Params.=21M, GFLOPS=1.81K2024.02 | 0.717 | 0.871 | 0.61 | 0.815 | 0.842 | 0.903 | 0.872 | 0.834 | 0.742 | 0.826 | 0.687 | 0.764 | 0.805 | — | — | |
| CREMAModality=V, A, Trainable Params.=9M, GFLOPS=1.78K2024.02 | 0.713 | 0.89 | 0.614 | 0.83 | 0.847 | 0.85 | 0.848 | 0.844 | 0.732 | 0.848 | 0.632 | 0.756 | 0.794 | — | — | |
| CREMAModality=V, Trainable Params.=4M, GFLOPS=1.30K2024.02 | 0.706 | 0.883 | 0.606 | 0.823 | 0.844 | 0.852 | 0.848 | 0.848 | 0.717 | 0.808 | 0.638 | 0.746 | 0.787 | — | — | |
| BLIP-2Modality=V, Trainable Params.=108M, GFLOPS=1.30K2024.02 | 0.701 | 0.867 | 0.585 | 0.763 | 0.872 | 0.937 | 0.905 | 0.815 | 0.72 | 0.813 | 0.643 | 0.742 | 0.789 | — | — | |
| BLIP-2Modality=V, A, Trainable Params.=216M, GFLOPS=1.78K2024.02 | 0.698 | 0.863 | 0.584 | 0.76 | 0.876 | 0.93 | 0.903 | 0.804 | 0.683 | 0.826 | 0.638 | 0.735 | 0.784 | — | — | |
| LAVISHModality=V, A, Trainable Params.=21M2024.02 | 0.691 | 0.856 | 0.659 | 0.814 | 0.802 | 0.811 | 0.806 | 0.846 | 0.692 | 0.788 | 0.656 | 0.738 | 0.769 | — | — | |
| BLIP-2Modality=V, A, F, Trainable Params.=324M, GFLOPS=3.21K2024.02 | 0.684 | 0.86 | 0.599 | 0.764 | 0.849 | 0.925 | 0.888 | 0.818 | 0.714 | 0.797 | 0.651 | 0.736 | 0.781 | — | — | |
| AVQAModality=V, A, Trainable Params.=18M2024.02 | 0.671 | 0.803 | 0.6 | 0.773 | 0.745 | 0.778 | 0.761 | 0.814 | 0.687 | 0.699 | 0.646 | 0.709 | 0.735 | — | — |