ResNet-D is a modification on the ResNet architecture that utilises an average pooling tweak for downsampling. The motivation is that in the unmodified ResNet, the 1 × 1 convolution for the downsampling block ignores 3/4 of input feature maps, so this is modified so no information will be ignored
Source: Bag of Tricks for Image Classification with Convolutional Neural NetworksPaper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
Image Classification | 4 | 13.33% |
Object Detection | 3 | 10.00% |
General Classification | 3 | 10.00% |
Action Classification | 2 | 6.67% |
Object | 2 | 6.67% |
Semantic Segmentation | 2 | 6.67% |
Autonomous Driving | 1 | 3.33% |
Video Recognition | 1 | 3.33% |
Instance Segmentation | 1 | 3.33% |