Audio-Visual Question Answering on VGGSound
69.8AccuracyMirasol3B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Mirasol3BModality=V, A, Trainable Params.=3B, Evaluation Protocol=Fine-tuning2024.02 | 69.8 | — | |
| Mirasol3B TTMModality=V, A, Trainable Params.=3B, Evaluation Protocol=Fine-tuning2024.02 | 66.4 | — | |
| CAV-MAEModality=V, A, Trainable Params.=324M, Evaluation Protocol=Fine-tuning2024.02 | 65.5 | — | |
| CREMAModality=V, A, Trainable Params.=9M, Evaluation Protocol=Fine-tuning2024.02 | 62.4 | 67 | |
| CREMAModality=V, Trainable Params.=1M, Evaluation Protocol=Fine-tuning2024.02 | 51.5 | 53.1 | |
| CAV-MAEModality=V, Trainable Params.=324M, Evaluation Protocol=Fine-tuning2024.02 | 47 | — |