Text-to-video on Wan 2.1
0LPIPSFull Inference
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Full InferenceSteps=25, TFLOPs=4139, Speedup=×1.002026.06 | 0 | 0.234 | 77.717 | |
| ReCache TaylorSeerSteps=13, O=1, TFLOPs=2169, Speedup=×1.902026.06 | 0.073 | 0.233 | 77.708 | |
| DPCacheSteps=13, O=1, TFLOPs=2168, Speedup=×1.902026.06 | 0.074 | 0.233 | 77.816 | |
| ReCache HiCacheSteps=13, O=1, TFLOPs=2169, Speedup=×1.902026.06 | 0.084 | 0.232 | 77.677 | |
| ReCache Taylor-DPSteps=13, O=1, TFLOPs=2168, Speedup=×1.902026.06 | 0.085 | 0.234 | 77.893 | |
| ReCache TaylorSeerSteps=7, O=1, TFLOPs=1184, Speedup=×3.502026.06 | 0.263 | 0.228 | 76.041 | |
| ReCache HiCacheSteps=7, O=1, TFLOPs=1184, Speedup=×3.502026.06 | 0.266 | 0.223 | 75.247 | |
| ReCache Taylor-DPSteps=7, O=1, TFLOPs=1183, Speedup=×3.502026.06 | 0.287 | 0.229 | 75.541 | |
| uniform TaylorSeerSteps=13, O=1, TFLOPs=2169, Speedup=×1.902026.06 | 0.406 | 0.216 | 75.069 | |
| uniform Taylor-DPSteps=13, O=1, TFLOPs=2168, Speedup=×1.902026.06 | 0.408 | 0.216 | 74.641 | |
| Full InferenceSteps=13, TFLOPs=2168, Speedup=×1.902026.06 | 0.41 | 0.217 | 74.974 | |
| uniform HiCacheSteps=13, O=1, TFLOPs=2169, Speedup=×1.902026.06 | 0.413 | 0.214 | 74.612 | |
| DPCacheSteps=7, O=1, TFLOPs=1183, Speedup=×3.502026.06 | 0.42 | 0.201 | 69.643 | |
| uniform TaylorSeerSteps=7, O=1, TFLOPs=1184, Speedup=×3.502026.06 | 0.499 | 0.185 | 70.43 | |
| uniform Taylor-DPSteps=7, O=1, TFLOPs=1183, Speedup=×3.502026.06 | 0.503 | 0.184 | 70.005 | |
| Full InferenceSteps=7, TFLOPs=1182, Speedup=×3.502026.06 | 0.506 | 0.188 | 70.652 | |
| uniform HiCacheSteps=7, O=1, TFLOPs=1184, Speedup=×3.502026.06 | 0.514 | 0.181 | 69.996 |