Classification on LRA Image N=1024 (test)
87.3AccuracyFull model
Evaluation Results
| Method | Links | |
|---|---|---|
| Full modelModel Backbone=S5, Pruning Ratio=0%, State Dimension=2562024.11 | 87.3 | |
| Global H_∞Model Backbone=S5, Pruning Ratio=30%, State Dimension=1792024.11 | 87.04 | |
| Uniform H_∞Model Backbone=S5, Pruning Ratio=30%, State Dimension=1792024.11 | 86.45 | |
| LASTModel Backbone=S5, Pruning Ratio=30%, State Dimension=1792024.11 | 86.34 | |
| Full modelModel Backbone=S4D, Pruning Ratio=0%, State Dimension=642024.11 | 77.02 | |
| Uniform H_∞Model Backbone=S4D, Pruning Ratio=0%, State Dimension=642024.11 | 77.02 | |
| Global H_∞Model Backbone=S4D, Pruning Ratio=0%, State Dimension=642024.11 | 77.02 | |
| LASTModel Backbone=S4D, Pruning Ratio=0%, State Dimension=642024.11 | 77.02 | |
| Paramixer (CDIL)Sequence Length=10242022.04 | 46.58 | |
| Paramixer (CHORD)Sequence Length=10242022.04 | 45.01 | |
| Sparse TransformerReference Source=[24], Sequence Length=10242022.04 | 44.24 | |
| ReformerReference Source=[15], Sequence Length=10242022.04 | 43.29 | |
| PerformerReference Source=[24], Sequence Length=10242022.04 | 42.77 | |
| TransformerReference Source=[24], Sequence Length=10242022.04 | 42.44 | |
| Linear TransformerReference Source=[24], Sequence Length=10242022.04 | 42.34 | |
| LongformerReference Source=[24], Sequence Length=10242022.04 | 42.22 | |
| NyströmformerReference Source=[15], Sequence Length=10242022.04 | 41.58 | |
| BigBirdReference Source=[24], Sequence Length=10242022.04 | 40.83 | |
| LinformerReference Source=[24], Sequence Length=10242022.04 | 38.56 | |
| TransformerReference Source=[15], Sequence Length=10242022.04 | 38.2 | |
| ReformerReference Source=[24], Sequence Length=10242022.04 | 38.07 | |
| LinformerReference Source=[15], Sequence Length=10242022.04 | 37.84 | |
| PerformerReference Source=[15], Sequence Length=10242022.04 | 37.07 |