Spatiotemporal Predictive Learning on Moving MNIST 10 time steps 2-digit (test)
89.8SSIMPredRNN++
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PredRNN++description=Causal LSTM + GHU (Final)2018.04 | 89.8 | 46.5 | |
| PredRNN + GHU2018.04 | 88.6 | 50.7 | |
| Causal LSTM2018.04 | 88.2 | 52.5 | |
| Causal LSTMvariant=spatial-to-temporal2018.04 | 87.5 | 54 | |
| VPNreproduced=true2018.04 | 87 | 64.1 | |
| PredRNN2018.04 | 86.7 | 56.8 | |
| DFN2018.04 | 72.6 | 89 | |
| CDNA2018.04 | 72.1 | 97.4 | |
| TrajGRU2018.04 | 71.3 | 106.9 | |
| ConvLSTM2018.04 | 70.7 | 103.3 | |
| FC-LSTM2018.04 | 69 | 118.3 |