A Wide Residual Block is a type of residual block that utilises two conv 3x3 layers (with dropout). This is wider than other variants of residual blocks (for instance bottleneck residual blocks). It was proposed as part of the WideResNet CNN architecture.
Source: Wide Residual NetworksPaper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
Image Classification | 19 | 25.68% |
Out-of-Distribution Detection | 6 | 8.11% |
Object Detection | 5 | 6.76% |
General Classification | 5 | 6.76% |
Anomaly Detection | 4 | 5.41% |
Adversarial Robustness | 4 | 5.41% |
Model Compression | 3 | 4.05% |
Sparse Learning | 2 | 2.70% |
Out of Distribution (OOD) Detection | 2 | 2.70% |
Component | Type |
|
---|---|---|
Batch Normalization
|
Normalization | |
Convolution
|
Convolutions | |
Dropout
|
Regularization | |
ReLU
|
Activation Functions | |
Residual Connection
|
Skip Connections |