Image Classification on MNIST-10 (test)
92.9Test AccuracyNFNNP
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| NFNNPType=Equivariant NFN2023.02 | 92.9 | — | — | — | |
| NFNHNPType=Equivariant NFN2023.02 | 92.5 | — | — | — | |
| GIW2023.05 | 90.2 | — | — | — | |
| QML Circuit (VAE)encoding=VAE2024.10 | 89.8 | — | — | — | |
| Pretrain-val2023.05 | 83.44 | — | — | — | |
| Val-only2023.05 | 79.86 | — | — | — | |
| Reweight2023.05 | 76.41 | — | — | — | |
| R-DIW2023.05 | 74.94 | — | — | — | |
| DIW2023.05 | 74.75 | — | — | — | |
| MW-Net2023.05 | 74.67 | — | — | — | |
| IMPLICIT solversolver_type=implicit2024.10 | 73.68 | 27.3 | 7,301.74 | 0.0056 | |
| IMPLICIT+WARMUP solversolver_type=implicit, warmup=true2024.10 | 73.33 | 27.3 | 7,017.38 | 0.0016 | |
| DIRECT solversolver_type=direct, layers=12024.10 | 72.78 | 14.5 | 3,343.4 | 4.026 | |
| DIRECT solversolver_type=direct, layers=52024.10 | 72.46 | 62.9 | 1,737.82 | 0.0021 | |
| DIRECT solversolver_type=direct, layers=22024.10 | 72.07 | 26.6 | 4,304.54 | 0.596 | |
| QML Circuit (PCA)encoding=PCA2024.10 | 71.75 | — | — | — | |
| DIRECT solversolver_type=direct, layers=102024.10 | 71.18 | 1,233.3 | 2,856.94 | 0.0002 | |
| MLPAugLayers=3-layer MLP, Hidden units=1000, Permutation augmentation=true2023.02 | 21 | — | — | — | |
| MLPLayers=3-layer MLP, Hidden units=1000, Permutation augmentation=false2023.02 | 14.5 | — | — | — |