Reference-to-Video Generation on OpenVid-HD R2V benchmark (test)
0.3107CLIP-I Video QualityContextAnyone
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| ContextAnyoneModel Scale (Parameters)=1.3B, Base Model=Wan 2.1 T2V 1.3B2025.12 | 0.3107 | 0.9831 | 0.6003 | 0.4824 | 0.5943 | 0.479 | 0.9457 | |
| PhantomModel Scale (Parameters)=1.3B2025.12 | 0.3095 | 0.9802 | 0.5636 | 0.4719 | 0.5412 | 0.4419 | 0.849 | |
| VACEModel Scale (Parameters)=1.3B2025.12 | 0.3012 | 0.9903 | 0.5489 | 0.4468 | 0.5394 | 0.4103 | 0.8123 |