Event Localization on ViTT (test)
45.89RecallVid2Seq
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Vid2SeqPT=true, Backbone=CLIP, Note=reproduced from official implementation2024.12 | 45.89 | 46.53 | 46.21 | |
| Vid2SeqPre-trained=true2026.03 | 45.89 | 46.53 | 46.21 | |
| HiCM²PT=true, Backbone=CLIP2024.12 | 45 | 47 | 45.98 | |
| HiCM2Pre-trained=true2026.03 | 45 | 47 | 45.98 | |
| Sali4VidPre-trained=true2026.03 | 44.31 | 49.1 | 46.58 | |
| STaRCPre-trained=true2026.03 | 40.02 | 49.72 | 44.34 | |
| Vid2SeqBackbone=CLIP2023.02 | 0.426 | 0.462 | — | |
| E2ESGBackbone=C3D, Source=Results provided by the authors2023.02 | 0.322 | 0.321 | — | |
| Streaming V2SPT=true, Backbone=CLIP2024.12 | — | — | 35.4 | |
| Streaming V2SPre-trained=true2026.03 | — | — | 35.4 |