Video Prediction on Kinetics-600 12 frames, 64x64
Metric: FVD (higher is better)
Metric: FVD (higher is better)
| # | Model↕ | FVD▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | LVT | 224.73 | No | Latent Video Transformer | 2020-06-18 | Code |
| 2 | OmniTokenizer-AR | 32.9 | No | OmniTokenizer: A Joint Image-Video Tokenizer for... | 2024-06-13 | Code |
| 3 | RaMViD | 16.46 | No | Diffusion Models for Video Prediction and Infill... | 2022-06-15 | Code |
| 4 | RIN (400 steps) | 11.5 | No | Scalable Adaptive Computation for Iterative Gene... | 2022-12-22 | Code |
| 5 | RIN (1000 steps) | 10.8 | No | Scalable Adaptive Computation for Iterative Gene... | 2022-12-22 | Code |
| 6 | LARP | 5.1 | No | LARP: Tokenizing Videos with a Learned Autoregre... | 2024-10-28 | Code |
| 7 | W.A.L.T.-L | 3.3 | No | Photorealistic Video Generation with Diffusion M... | 2023-12-11 | - |
| 8 | SiD2 | 2.3 | No | Simpler Diffusion (SiD2): 1.5 FID on ImageNet512... | 2024-10-25 | - |