PDE solving on Shallow Water 3D (test)
0.0046Relative MSE LossDynFormer
Evaluation results
| Method | Links | |
|---|---|---|
| DynFormerModel Variant=Large, Peak GPU memory (GB)=11.9, Number of Parameters=82.9K, Averaged training time (s/epoch)=18.92026.03 | 0.0046 | |
| DynFormerModel Variant=Medium, Peak GPU memory (GB)=7.1, Number of Parameters=39.7K, Averaged training time (s/epoch)=14.92026.03 | 0.0062 | |
| FactFormerModel Variant=Large, Peak GPU memory (GB)=12.3, Number of Parameters=203.3K, Averaged training time (s/epoch)=15.12026.03 | 0.0266 | |
| DynFormerModel Variant=Tiny, Peak GPU memory (GB)=3.2, Number of Parameters=7.0K, Averaged training time (s/epoch)=13.82026.03 | 0.0323 | |
| FactFormerModel Variant=Medium, Peak GPU memory (GB)=7.7, Number of Parameters=99.9K, Averaged training time (s/epoch)=13.42026.03 | 0.0558 | |
| TransolverModel Variant=Large, Peak GPU memory (GB)=12.1, Number of Parameters=96.5K, Averaged training time (s/epoch)=16.92026.03 | 0.12 | |
| FactFormerModel Variant=Tiny, Peak GPU memory (GB)=4.1, Number of Parameters=25.6K, Averaged training time (s/epoch)=13.32026.03 | 0.124 | |
| OFormerModel Variant=Large, Peak GPU memory (GB)=12.6, Number of Parameters=33.7K, Averaged training time (s/epoch)=14.62026.03 | 0.139 | |
| TransolverModel Variant=Medium, Peak GPU memory (GB)=7.2, Number of Parameters=43.4K, Averaged training time (s/epoch)=11.72026.03 | 0.173 | |
| TransolverModel Variant=Tiny, Peak GPU memory (GB)=4.7, Number of Parameters=22.4K, Averaged training time (s/epoch)=10.52026.03 | 0.212 | |
| OFormerModel Variant=Medium, Peak GPU memory (GB)=7.0, Number of Parameters=11.7K, Averaged training time (s/epoch)=13.02026.03 | 0.231 | |
| OFormerModel Variant=Tiny, Peak GPU memory (GB)=3.7, Number of Parameters=3.1K, Averaged training time (s/epoch)=12.92026.03 | 0.352 | |
| GNOTModel Variant=Medium, Peak GPU memory (GB)=7.0, Number of Parameters=7.4K, Averaged training time (s/epoch)=15.62026.03 | 0.465 | |
| GNOTModel Variant=Large, Peak GPU memory (GB)=12.3, Number of Parameters=16.5K, Averaged training time (s/epoch)=21.92026.03 | 0.475 | |
| GNOTModel Variant=Tiny, Peak GPU memory (GB)=3.9, Number of Parameters=3.9K, Averaged training time (s/epoch)=12.12026.03 | 0.594 | |
| ONOModel Variant=Medium, Peak GPU memory (GB)=7.5, Number of Parameters=17.8K, Averaged training time (s/epoch)=15.02026.03 | 0.95 | |
| ONOModel Variant=Large, Peak GPU memory (GB)=12.3, Number of Parameters=39.3K, Averaged training time (s/epoch)=17.92026.03 | 0.989 | |
| ONOModel Variant=Tiny, Peak GPU memory (GB)=4.4, Number of Parameters=8.9K, Averaged training time (s/epoch)=12.92026.03 | 0.994 |