ResNet: Counting Layers in a Basic-block Network
Intermediate
The problem, from first principles
ResNet names its models by depth, for example ResNet-18 and ResNet-34. In the basic-block design, the depth is a simple count: one stem convolution, two convolutions per residual block, and one classifier layer.
From theory to code
Implement resnet_depth(blocks), the nominal depth from the number of basic blocks per stage.
Constraints
- Returns an integer.
Hints
Hint 1
Two layers for the stem and classifier, plus two per block across all stages.
Keywords
research-paperscomputer-visioncnnarchitecture
Python
Run your code to see results
No results yet
Run your code to check it against the examples, or submit to run every test.