Image Classification on CIFAR-100 Standard data augmentation (test)
17.18Test ErrorDenseNet-BC
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DenseNet-BCDropout=false, Depth=190, Params=25.6 M, Growth Rate=402017.09 | 17.18 | — | |
| CoPaNetDropout=true, Depth=164, Params=27.9 M, Pathways (k)=2, Width (m)=42017.09 | 18.67 | — | |
| Wide ResNetDropout=true, Depth=28, Params=36.5 M, Width (m)=102017.09 | 18.85 | — | |
| CoPaNet-RDropout=true, Depth=164, Params=15.7 M, Pathways (k)=2, Width (m)=32017.09 | 18.9 | — | |
| DenseNetDropout=false, Depth=100, Params=27.2 M, Growth Rate=242017.09 | 19.25 | — | |
| CoPaNet-RDropout=true, Depth=164, Params=7.00 M, Pathways (k)=2, Width (m)=22017.09 | 20.29 | — | |
| Wide ResNetDropout=true, Depth=16, Params=11.0 M, Width (m)=82017.09 | 20.43 | — | |
| CoPaNetDropout=true, Depth=164, Params=6.98 M, Pathways (k)=2, Width (m)=22017.09 | 20.48 | — | |
| pre-activation ResNetDropout=false, Depth=1001, Params=10.2 M2017.09 | 22.71 | — | |
| CoPaNetDropout=true, Depth=164, Params=1.75 M, Pathways (k)=2, Width (m)=12017.09 | 22.86 | — | |
| pre-activation ResNetDropout=false, Depth=164, Params=1.7 M2017.09 | 24.33 | — | |
| Stochastic DepthDropout=false, Depth=110, Params=1.7 M2017.09 | 24.58 | — | |
| ResNet with stochastic depthData Augmentation=Standard2016.03 | 24.98 | — | |
| Scalable BOData Augmentation=Standard2016.03 | 27.4 | — | |
| Frac. PoolData Augmentation=Standard2016.03 | 27.62 | — | |
| ResNet with constant depthData Augmentation=Standard2016.03 | 27.76 | — | |
| Maxout Network In NetworkDropout=true, Params=1.6 M, Pathways (k)=52017.09 | 28.86 | — | |
| Learning ActivationData Augmentation=Standard2016.03 | 30.83 | — | |
| Highway NetworkData Augmentation=Standard2016.03 | 32.24 | — | |
| Highway NetworkDropout=false2017.09 | 32.34 | — | |
| Network In NetworkDropout=true, Params=0.98 M2017.09 | 35.68 | — | |
| Maxout NetworkDropout=true, Pathways (k)=22017.09 | 38.57 | — | |
| EvolutionStarting Point=SINGLE LAYER, ZERO CONVS., Constraints=POWER-OF-2 STRIDES, Post-Processing=NONE, Params=40.4 M2017.03 | — | 77 | |
| Q-LearningConstraints=DISCRETE PARAMS., MAX. NUM. LAYERS, NO SKIPS, Post-Processing=TUNE, RETRAIN, Params=11.2 M2017.03 | — | 72.9 |