Video Frame Interpolation on X (test)
30.81PSNRM2M-PWC
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| M2M-PWCPretrained flow=true, # Param. (Mill.)=7.6, Memory (for 4K)=10 GB, Training dataset=Vimeo90K, Inference (in s/f)=0.212022.11 | 30.81 | — | — | |
| fLDR-NetPretrained flow=false, # Param. (Mill.)=0.9, Memory (for 4K)=4.6 GB, Training dataset=X-Train, Inference (in s/f)=0.512022.11 | 30.45 | — | — | |
| XVFIPretrained flow=false, # Param. (Mill.)=5.5, Memory (for 4K)=>12 GB, Training dataset=X-Train2022.11 | 30.34 | — | — | |
| M2M-DISPretrained flow=true, Training dataset=Vimeo90K2022.11 | 30.18 | — | — | |
| ABMEPretrained flow=false, # Param. (Mill.)=18.1, Training dataset=Vimeo90K2022.11 | 30.16 | — | — | |
| RIFEmPretrained flow=false, # Param. (Mill.)=9.8, Memory (for 4K)=6.8 GB, Training dataset=X-Train, Inference (in s/f)=0.402022.11 | 28.06 | — | — | |
| RIFEmPretrained flow=false, # Param. (Mill.)=9.8, Memory (for 4K)=6.8 GB, Training dataset=Vimeo90K, Inference (in s/f)=0.402022.11 | 26.8 | — | — | |
| SoftSplatPretrained flow=true, # Param. (Mill.)=7.7, Training dataset=Vimeo90K2022.11 | 25.48 | — | — | |
| fLDR-NetInference (in s/f)=0.50*/0.512022.11 | — | 0.871 | 0.099 | |
| M2M-PWCInference (in s/f)=0.212022.11 | — | 0.914 | 0.086 | |
| RIFElInference (in s/f)=0.402022.11 | — | 0.793 | 0.227 | |
| RIFEmInference (in s/f)=0.402022.11 | — | 0.751 | 0.26 | |
| XVFIInference (in s/f)=0.66*/N/A2022.11 | — | 0.874 | 0.085 |