C++ API Reference for Intel® Data Analytics Acceleration Library 2020 Update 1
Parameters for the forward and backward pooling layers. More...
DAAL_DEPRECATED Parameter | ( | size_t | firstIndex, |
size_t | secondIndex, | ||
size_t | thirdIndex, | ||
size_t | firstKernelSize = 2 , |
||
size_t | secondKernelSize = 2 , |
||
size_t | thirdKernelSize = 2 , |
||
size_t | firstStride = 2 , |
||
size_t | secondStride = 2 , |
||
size_t | thirdStride = 2 , |
||
size_t | firstPadding = 0 , |
||
size_t | secondPadding = 0 , |
||
size_t | thirdPadding = 0 |
||
) |
Constructs the parameters of 3D pooling layer
[in] | firstIndex | Index of the first of three dimensions on which the pooling is performed |
[in] | secondIndex | Index of the second of three dimensions on which the pooling is performed |
[in] | thirdIndex | Index of the third of three dimensions on which the pooling is performed |
[in] | firstKernelSize | Size of the first dimension of three-dimensional subtensor for which the kernel is applied |
[in] | secondKernelSize | Size of the second dimension of three-dimensional subtensor for which the kernel is applied |
[in] | thirdKernelSize | Size of the third dimension of three-dimensional subtensor for which the kernel is applied |
[in] | firstStride | Interval over the first dimension on which the pooling is performed |
[in] | secondStride | Interval over the second dimension on which the pooling is performed |
[in] | thirdStride | Interval over the third dimension on which the pooling is performed |
[in] | firstPadding | Number of data elements to implicitly add to the the first dimension of the three-dimensional subtensor on which the pooling is performed |
[in] | secondPadding | Number of data elements to implicitly add to the the second dimension of the three-dimensional subtensor on which the pooling is performed |
[in] | thirdPadding | Number of data elements to implicitly add to the the third dimension of the three-dimensional subtensor on which the pooling is performed |
KernelSizes kernelSizes |
Data structure representing the size of the three-dimensional subtensor from which the maximum element is selected
Paddings paddings |
Data structure representing the number of data elements to implicitly add to each size of the three-dimensional subtensor on which pooling is performed
Strides strides |
Data structure representing the intervals on which the subtensors for pooling are selected
For more complete information about compiler optimizations, see our Optimization Notice.