Long-Context Video Prediction on Minecraft 128x128 (test)
0.448SSIMFAR-M-Long
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| FAR-M-LongParams=280 M2025.03 | 0.448 | 16.9 | 0.251 | 39 | |
| TECOParams=274 M2025.03 | 0.381 | 15.4 | 0.34 | 116 | |
| Latent FDMParams=33 M2025.03 | 0.349 | 13.4 | 0.429 | 167 | |
| FitVidParams=176 M2025.03 | 0.343 | 13 | 0.519 | 956 | |
| CW-VAEParams=140 M2025.03 | 0.338 | 13.4 | 0.441 | 397 | |
| Perceiver ARParams=166 M2025.03 | 0.323 | 13.2 | 0.441 | 76 |