Image Classification on CIFAR-10 Standard data augmentation (test)
3.38Test Error RateCoPaNet-R
Evaluation Results
| Method | Links | |
|---|---|---|
| CoPaNet-RDropout=true, Depth=164, Params=15.7 M, Pathways (k)=2, Width (m)=32017.09 | 3.38 | |
| DenseNet-BCDropout=false, Depth=190, Params=25.6 M, Growth Rate=402017.09 | 3.46 | |
| DenseNet-BCDepth=190, Params=25.6M, config=L = 190, k = 402017.07 | 3.46 | |
| CoPaNet-RDropout=true, Depth=164, Params=7.00 M, Pathways (k)=2, Width (m)=22017.09 | 3.55 | |
| NASDepth=39, Params=37.4M, config=post-processing, Hardware=800 GPUs2017.07 | 3.65 | |
| DenseNetDropout=false, Depth=100, Params=27.2 M, Growth Rate=242017.09 | 3.74 | |
| CoPaNetDropout=true, Depth=164, Params=27.9 M, Pathways (k)=2, Width (m)=42017.09 | 3.74 | |
| Wide ResNetDropout=true, Depth=28, Params=36.5 M, Width (m)=102017.09 | 3.89 | |
| CoPaNetDropout=true, Depth=164, Params=6.98 M, Pathways (k)=2, Width (m)=22017.09 | 4.1 | |
| Wide ResNetDepth=28, Params=36.5M2017.07 | 4.17 | |
| EASDepth=20, Params=23.4M, config=plain CNN, Hardware=5 GPUs2017.07 | 4.23 | |
| Wide ResNetDropout=true, Depth=16, Params=11.0 M, Width (m)=82017.09 | 4.27 | |
| NASDepth=39, Params=7.1M, config=max pooling, Hardware=800 GPUs2017.07 | 4.47 | |
| CoPaNetDropout=true, Depth=164, Params=1.75 M, Pathways (k)=2, Width (m)=12017.09 | 4.5 | |
| DenseNet-BCDepth=100, Params=0.8M, config=L = 100, k = 122017.07 | 4.51 | |
| pre-activation ResNetDropout=false, Depth=1001, Params=10.2 M2017.09 | 4.62 | |
| ResNet (pre-activation)Depth=1001, Params=10.2M2017.07 | 4.62 | |
| Wide ResNetDepth=16, Params=11.0M2017.07 | 4.81 | |
| Stochastic DepthDropout=false, Depth=1202, Params=19.4 M2017.09 | 4.91 | |
| ResNet (stochastic depth)Depth=1202, Params=10.2M2017.07 | 4.91 | |
| Stochastic DepthDropout=false, Depth=110, Params=1.7 M2017.09 | 5.23 | |
| DenseNetDepth=40, Params=1.0M, config=L = 40, k = 122017.07 | 5.24 | |
| ResNet with stochastic depthData Augmentation=Standard2016.03 | 5.25 | |
| Large-Scale EvolutionHardware=250 GPUs, Params=5.4M2017.07 | 5.4 | |
| pre-activation ResNetDropout=false, Depth=164, Params=1.7 M2017.09 | 5.46 | |
| NASDepth=20, Params=2.5M, config=predicting strides, Hardware=800 GPUs2017.07 | 6.01 | |
| Gen. PoolData Augmentation=Standard2016.03 | 6.05 | |
| Scalable BOData Augmentation=Standard2016.03 | 6.37 | |
| ResNet with constant depthData Augmentation=Standard2016.03 | 6.41 | |
| ResNetDropout=false, Depth=110, Params=1.7 M2017.09 | 6.43 | |
| ResNetDepth=110, Params=1.7M2017.07 | 6.61 | |
| Maxout Network In NetworkDropout=true, Params=1.6 M, Pathways (k)=52017.09 | 6.75 | |
| R-CNNData Augmentation=Standard2016.03 | 7.09 | |
| All-CNNData Augmentation=Standard2016.03 | 7.25 | |
| Learning ActivationData Augmentation=Standard2016.03 | 7.51 | |
| Highway NetworkData Augmentation=Standard2016.03 | 7.6 | |
| Highway NetworkDropout=false2017.09 | 7.6 | |
| Deeply SupervisedData Augmentation=Standard2016.03 | 7.97 | |
| Net in NetData Augmentation=Standard2016.03 | 8.81 | |
| Network In NetworkDropout=true, Params=0.98 M2017.09 | 8.81 | |
| DropConnectData Augmentation=Standard2016.03 | 9.32 | |
| MaxoutData Augmentation=Standard2016.03 | 9.38 | |
| Maxout NetworkDropout=true, Pathways (k)=22017.09 | 9.38 |