Video Question Answering on TGIF-QA-R (test)
0.665Action AccuracyMAMA-VIOLET
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MAMA-VIOLETadaptation=VIOLET2026.01 | 0.665 | 0.717 | 0.895 | |
| LaViLa-VIOLETadaptation=VIOLET2026.01 | 0.643 | 0.707 | 0.87 | |
| Vid2Seq-VIOLETadaptation=VIOLET2026.01 | 0.642 | 0.701 | 0.866 | |
| VIOLET2026.01 | 0.626 | 0.7 | 0.863 | |
| OmniVL2026.01 | 0.62 | 0.695 | 0.855 | |
| MAMA-VGTadaptation=VGT2026.01 | 0.617 | 0.627 | 0.74 | |
| MERLOT2026.01 | 0.614 | 0.693 | 0.84 | |
| VGT2026.01 | 0.61 | 0.617 | 0.732 | |
| VindLU2026.01 | 0.595 | 0.658 | 0.872 | |
| LAVENDER2026.01 | 0.571 | 0.669 | 0.84 | |
| Singularity2026.01 | 0.531 | 0.651 | 0.815 | |
| All-in-one2026.01 | 0.349 | 0.65 | 0.538 |