1. Pls explain the following concepts in CNN:
1) Convolutional Layer: This layer performs convolution operations, and creates several channels of features.
2) Pooling Layer: Pooling is a typical down-sampling operation on the feature map, and pooling reduces the dimensionality of the feature map.
Copyright By PowCoder代写 加微信 powcoder
2. Why do we adopt the Pooling Layer in a CNN? What are the differences between max-pooling and average-pooling?
One reason is that the pooling layer can reduce the spatial resolution of the feature map, and it can also accelerate the computation of the network.
Max pooling: reserve the maximum value from that portion of the feature map.
Average pooling: Computes the average value of that portion of the feature map.
3. Pls explain the concept of stride.
Stride: the number of pixels (the concept of pixel here is extended to the feature map as well) by which we slide over the filter kernel.
If Stride =1, then move the filter one pixel at a time.
If Stride=2, then move the filter two-pixel at a time.
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com