Audio-visual QA on MUSIC-AVQA
84.3Overall AccuracyCAT-7B-FT
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| CAT-7B-FTEvaluation Protocol=Task-specific, Model Scale=≥ 7B2026.05 | 84.3 | 84.9 | 86.1 | 83.2 | |
| V-LynX-7BEvaluation Protocol=Task-specific, Model Scale=≥ 7B, ∆Params.=195.0M2026.05 | 83 | 80.8 | 93.5 | 78.8 | |
| PAVE-7BEvaluation Protocol=Task-specific, Model Scale=≥ 7B, ∆Params.=256.7M2026.05 | 82.3 | 79.7 | 93 | 78 | |
| V-LynX-0.5BEvaluation Protocol=Task-specific, Model Scale=< 7B, ∆Params.=68.7M2026.05 | 81.1 | 78.9 | 92.2 | 76.5 | |
| VASTEvaluation Protocol=Task-specific, Model Scale=< 7B2026.05 | 80.7 | — | — | — | |
| PAVE-0.5BEvaluation Protocol=Task-specific, Model Scale=< 7B, ∆Params.=127.6M2026.05 | 78.8 | 77.3 | 89.8 | 74.1 | |
| LLaVA-OV-7B-FTEvaluation Protocol=Task-specific, Model Scale=≥ 7B, ∆Params.=161.5M2026.05 | 77.4 | 75.4 | 89.3 | 72.3 | |
| LLaVA-OV-0.5B-FTEvaluation Protocol=Task-specific, Model Scale=< 7B, ∆Params.=35.2M2026.05 | 67.6 | 69.6 | 76.3 | 62.8 | |
| LLaVA-OV-7BEvaluation Protocol=Zero-shot2026.05 | 60.4 | 68.8 | 70.6 | 52.8 | |
| LLaVA-OV-0.5BEvaluation Protocol=Zero-shot2026.05 | 52.8 | 60 | 57.1 | 48.5 | |
| CAT-7BEvaluation Protocol=Zero-shot2026.05 | 48.6 | — | — | — |