The pooling layers similar to the convolutional layers except an important difference: the pooling occurs for the channels separately.
AveragePooling1D
This layer calculates the average value in the kernel (sliding window).
Input:
A Data2D type data with shape: (1, length, channels, batches).
Output:
A Data2D type data with shape: (1, new length, channels, batches).
The shape of the filter: (1, kernel size, 1, 1). Then the output length (new length):
new length = 1 + (length + 2 * padding - kernel size)/stride.
AveragePooling2D
Input:
A Data2D type data with shape: (height, width, channels, batches).
Output:
A Data2D type data with shape: (new height, new width, channels, batches).
The shape of the filter: (kernel height, kernel width, 1, 1). Then the output sizes:
new height = 1 + (height + 2 * padding vertical - kernel height)/stride vertical,
new width = 1 + (width + 2 * padding horizontal - kernel width)/stride horizontal.
GlobalAveragePooling1D
Calculates the average value for each channel in the input data.
Input:
A Data2D type data with shape: (1, length, channels, batches).
Output:
A Data2D type data with shape: (1, 1, channels, batches).
GlobalAveragePooling2D
Calculates the average value for each channel in the input data.
Input:
A Data2D type data with shape: (height, width, channels, batches).
Output:
A Data2D type data with shape: (1, 1, channels, batches).
GlobalMaxPooling1D
Calculates the maximum value for each channel in the input data.
Input:
A Data2D type data with shape: (1, length, channels, batches).
Output:
A Data2D type data with shape: (1, 1, channels, batches).
GlobalMaxPooling2D
Calculates the maximum value for each channel in the input data.
Input:
A Data2D type data with shape: (height, width, channels, batches).
Output:
A Data2D type data with shape: (1, 1, channels, batches).
MaxPooling1D
This layer calculates the maximum value in the kernel (sliding window).
Input:
A Data2D type data with shape: (1, length, channels, batches).
Output:
A Data2D type data with shape: (1, new length, channels, batches).
The shape of the filter: (1, kernel size, 1, 1). Then the output length (new length):
new length = 1 + (length + 2 * padding - kernel size)/stride.
MaxPooling2D
This layer calculates the maximum value in the kernel (sliding window).
Input:
A Data2D type data with shape: (height, width, channels, batches).
Output:
A Data2D type data with shape: (new height, new width, channels, batches).
The shape of the filter: (kernel height, kernel width, 1, 1). Then the output sizes:
new height = 1 + (height + 2 * padding vertical - kernel height)/stride vertical,
new width = 1 + (width + 2 * padding horizontal - kernel width)/stride horizontal.