Narrative Action Evaluation on FineGym NAE re-annotated (test)
16.2mAPPrompt-Guided Multimodal Interaction
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Prompt-Guided Multimodal InteractionBackbone=Video Swin Transformer, Vision-Language Model=CLIP2024.04 | 16.2 | 0.749 | 1.55 | 28.9 | 23.7 | 20.7 | 93 | |
| SwinBert2024.04 | 11.8 | 0.656 | 2.13 | 27.4 | 21 | 17.8 | 85 | |
| VLTinT2024.04 | 9.4 | 0.64 | 2.33 | 15.4 | 17.6 | 14.2 | 84 | |
| VLCap2024.04 | 8.6 | 0.627 | 3.07 | 13.6 | 12.1 | 13.5 | 81 | |
| UniVL2024.04 | 5.7 | 0.604 | 3.81 | 11 | 10.7 | 13.3 | 79 | |
| C3D-AVG2024.04 | 5.1 | 0.606 | 3.76 | 10 | 9.8 | 11.6 | 80 | |
| MSCADC2024.04 | 2.5 | 0.583 | 4.42 | 10.2 | 10.3 | 12.6 | 76 |