Human-Object Interaction Video Generation on Self-filmed and in-the-wild video clips (test)
67.501FIDOurs
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| OursBase Model=Wan2.1-14B2026.03 | 67.501 | 0.072 | 0.547 | 560.29 | 0.995 | 0.958 | 0.987 | 0.891 | 0.89 | 0.832 | |
| HunyuanCustom2026.03 | 72.185 | 0.06 | 0.541 | 742.34 | 0.994 | 0.956 | 0.975 | 0.832 | 0.77 | 0.603 | |
| VACE-14BModel Scale=14B2026.03 | 105.708 | 0.134 | 0.524 | 862.65 | 0.995 | 0.953 | 0.972 | 0.871 | 0.802 | 0.677 | |
| Nano+WanAnimateImage Editing=Nano, Animation=WanAnimate2026.03 | 119.89 | 0.295 | 0.537 | 948.61 | 0.995 | 0.953 | 0.989 | 0.771 | 0.806 | 0.643 | |
| HuMotype=multi-subject video generation2026.03 | 254.62 | 0.552 | 0.464 | 1,265.26 | 0.994 | 0.977 | 0.982 | — | 0.81 | 0.709 |