Video Reconstruction on Webvid (val)
34.75PSNRARVAE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ARVAE# Data=0.1M, Param.=5.9M, Down.=8×8, # Ch.=42025.12 | 34.75 | 0.944 | 0.036 | |
| VideoVAE+# Data=10M, Param.=500M, Down.=8×8, # Ch.=162025.12 | 34.42 | 0.94 | 0.037 | |
| VAE-SD3Params=34M + 49M, FCR=1x2024.05 | 33.4 | 0.91 | 0.064 | |
| ARVAE# Data=0.1M, Param.=6.2M, Down.=16×16, # Ch.=162025.12 | 33.4 | 0.91 | 0.053 | |
| VidTok-KL# Data=16M, Param.=174M, Down.=8×8, # Ch.=162025.12 | 33.24 | 0.91 | 0.036 | |
| Step-Video# Data=670M, Param.=499M, Down.=16×16, # Ch.=642025.12 | 32.5 | 0.905 | 0.054 | |
| CVVAE-SD3Params=68M + 114M, FCR=4x, Comp.=true2024.05 | 32 | 0.899 | 0.087 | |
| CogVideoX# Data=35M, Param.=206M, Down.=8×8, # Ch.=162025.12 | 31.9 | 0.901 | 0.08 | |
| Cosmos-CV# Data=10M, Param.=101M, Down.=8×8, # Ch.=162025.12 | 31.25 | 0.886 | 0.103 | |
| ARVAE# Data=0.1M, Param.=6.4M, Down.=32×32, # Ch.=162025.12 | 30.55 | 0.873 | 0.084 | |
| LTX-Video# Data=N/A, Param.=419M, Down.=32×32, # Ch.=1282025.12 | 30.25 | 0.853 | 0.093 | |
| VAE-SD2.1Params=34M + 49M, Frame Compression Rate (FCR)=1x, Compatibility with generative models=N/A2024.05 | 28.9 | 0.81 | 0.145 | |
| CV-VAE (3D)Params=100M + 156M, Frame Compression Rate (FCR)=4x, Compatibility with generative models=true2024.05 | 28.6 | 0.819 | 0.145 | |
| CV-VAE (2D+3D)Params=68M + 114M, Frame Compression Rate (FCR)=4x, Compatibility with generative models=true2024.05 | 28.5 | 0.817 | 0.143 | |
| VAE-OSPParams=94M + 135M, Frame Compression Rate (FCR)=4x, Compatibility with generative models=false2024.05 | 26.7 | 0.781 | 0.166 | |
| TATSParams=7M + 16M, Frame Compression Rate (FCR)=4x, Compatibility with generative models=false2024.05 | 24.1 | 0.729 | 0.31 |