Convolutions
Switchable Atrous Convolution
Introduced by Qiao et al. in DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous ConvolutionSwitchable Atrous Convolution (SAC) softly switches the convolutional computation between different atrous rates and gathers the results using switch functions. The switch functions are spatially dependent, i.e., each location of the feature map might have different switches to control the outputs of SAC. To use SAC in a detector, we convert all the standard 3x3 convolutional layers in the bottom-up backbone to SAC.
Source: DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous Convolution
Papers
| Paper | Code | Results | Date | Stars |
|---|
Tasks
| Task | Papers | Share |
|---|---|---|
| Reinforcement Learning (RL) | 85 | 23.74% |
| Deep Reinforcement Learning | 59 | 16.48% |
| Reinforcement Learning | 50 | 13.97% |
| Continuous Control | 30 | 8.38% |
| OpenAI Gym | 9 | 2.51% |
| Decision Making | 8 | 2.23% |
| Atari Games | 8 | 2.23% |
| Imitation Learning | 5 | 1.40% |
| Autonomous Driving | 5 | 1.40% |
Usage Over Time
Components
| Component | Type |
|
|---|---|---|
1x1 Convolution
|
Convolutions | |
Average Pooling
|
Pooling Operations | |
Convolution
|
Convolutions | |
Dilated Convolution
|
Convolutions | |
Global Average Pooling
|
Pooling Operations |
