Text-to-Video Generation on VideoJAM-Bench
4.126T2V Alignment ScoreSpecLoR-Adapter
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| SpecLoR-AdapterNFE=44, Time=1.1×2026.06 | 4.126 | 3.441 | 3.927 | 94 | 97.7 | |
| SpecLoR-ZeroNFE=44, Time=1.1×2026.06 | 4.118 | 3.448 | 3.939 | 93.9 | 97.7 | |
| SRVSNFE=60, Time=1.5×2026.06 | 4.067 | 3.341 | 3.926 | 94.3 | 97.9 | |
| Wan2.2 T2V (UniPC)NFE=402026.06 | 4.062 | 3.237 | 3.873 | 93.5 | 97.2 | |
| Wan2.2 T2V (UniPC) (+ NFE×1.5)NFE=60, Time=1.5×2026.06 | 4.046 | 3.26 | 3.834 | 93.5 | 97.4 | |
| FlowMoNFE=40, Time=1.7×2026.06 | 4.044 | 3.179 | 3.805 | 93.9 | 97.4 | |
| HunyuanVideo-1.5 + SpecLoR-ZeroNFE=57, Time=1.14×2026.06 | 3.89 | 3.26 | 3.736 | 94.3 | 99.2 | |
| HunyuanVideo-1.5NFE=502026.06 | 3.833 | 2.924 | 3.486 | 93.8 | 98.5 |