Text-to-Video Generation on VideoScore Predicted Failure 2 (Evaluation Set)
3.94Visual QualityCogVideoX-5B (Predicted Failure + Ours)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| CogVideoX-5B (Predicted Failure + Ours)Model=CogVideoX-5B, Setting=Predicted Failure + Ours2026.03 | 3.94 | 3.91 | 3.74 | |
| Wan2.1-1.3B (Predicted Failure + Ours)Model=Wan2.1-1.3B, Setting=Predicted Failure + Ours2026.03 | 3.84 | 3.82 | 3.67 | |
| CogVideoX-5B (Predicted Failure baseline)Model=CogVideoX-5B, Setting=Predicted Failure (wo. Intervention)2026.03 | 3.77 | 3.55 | 3.56 | |
| Wan2.1-1.3B (Predicted Failure baseline)Model=Wan2.1-1.3B, Setting=Predicted Failure (wo. Intervention)2026.03 | 3.59 | 3.54 | 3.49 |