Semantic Role Labeling on VidSitu (val)
90.12CIDErTypesDev-ucofia
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TypesDev-ucofiaBackbone=BLIP22026.04 | 90.12 | 38.36 | 44.02 | |
| CineMECBackbone=YS + SF2026.04 | 76.34 | 55.78 | 52.45 | |
| VideoWhispererBackbone=YS + SF2026.04 | 73.73 | 48.37 | 48.87 | |
| CineMEC (old backbone)Backbone=FR + SF2026.04 | 68.38 | 54.93 | 48.91 | |
| VideoWhispererBackbone=FR + SF2026.04 | 68.23 | 48.22 | 43.93 | |
| ClipSituBackbone=X-CLIP2026.04 | 61.93 | 37.77 | — | |
| HostSGBackbone=M-TDE2026.04 | 55.09 | 55.7 | 35.01 | |
| OME+OIEBackbone=I3D2026.04 | 47.16 | — | — | |
| VidSitu-SlowFastBackbone=SF2026.04 | 45.52 | 50.48 | 31.99 | |
| Human2026.04 | — | 84.85 | 72.1 |