Multimodal Action Recognition on Kinetic-Sounds ViT
74.2AccuracyDMIL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DMILBackbone=ViT2026.06 | 74.2 | 81.83 | |
| EnsembleBackbone=ViT2026.06 | 71.8 | 79.03 | |
| MLBBackbone=ViT2026.06 | 71.22 | 77.29 | |
| MCRBackbone=ViT2026.06 | 69.51 | 76.33 | |
| MMMLBackbone=ViT2026.06 | 69.14 | 75.82 | |
| I2M2Backbone=ViT2026.06 | 68.93 | 73.55 | |
| FCLBackbone=ViT2026.06 | 68.9 | 75.38 | |
| AGMBackbone=ViT2026.06 | 68.51 | 75.43 | |
| QMFBackbone=ViT2026.06 | 68.1 | 72.9 | |
| OGMBackbone=ViT2026.06 | 68.02 | 73.42 | |
| PMRBackbone=ViT2026.06 | 67.99 | 73.36 | |
| JointBackbone=ViT2026.06 | 67.81 | 73.11 | |
| MMIBBackbone=ViT2026.06 | 67.57 | 73.52 |