Average Pooling Prototype and Function List

Description

Average pooling computes each value of the output tensor as the average of all values in the related perception area of a single channel of the input tensor. The perception area size is defined as \(kernel\_width * kernel\_height\) (see Figure Example for 2x2 Average pooling).

Example for 2x2 Average pooling

For example, in a HWC data layout, if the in feature map is \((Hi, Wi, Ci)\), the out feature map is \((Ho, Wo, Ci)\) tensor where the spatial dimensions comply with the system of equations (1).

Functions

Kernels which implement average pooling functions have the following prototype:

mli_status  mli_krn_avepool_hwc_<data_format>(
   const mli_tensor *in,
   const mli_pool_cfg *cfg,
   mli_tensor *out);

Average Pooling Function Parameters

Parameter	Type	Description
in	mli_tensor *	[IN] Pointer to constant input tensor
cfg	mli_pool_cfg *	[IN] Pointer to pooling parameters structure
out	mli_tensor *	[IN | OUT] Pointer to output feature map tensor. Result is stored here

List of Available Average Pooling Functions

Function Name	Details
mli_krn_avepool_hwc_sa8	In/out layout: HWC In/out data format: sa8 Supports any kernel size
mli_krn_avepool_hwc_fx16	In/out layout: HWC In/out data format: fx16 Supports any kernel size
mli_krn_avepool_hwc_sa8_k2x2	In/out layout: HWC In/out data format: sa8 Kernel width: 2 Kernel height: 2
mli_krn_avepool_hwc_fx16_k2x2	In/out layout: HWC In/out data format: sa8 Kernel width: 2 Kernel height: 2
mli_krn_avepool_hwc_sa8_k3x3	In/out layout: HWC In/out data format: sa8 Kernel width: 3 Kernel height: 3
mli_krn_avepool_hwc_fx16_k3x3	In/out layout: HWC In/out data format: sa8 Kernel width: 3 Kernel height: 3

Conditions

Ensure that you satisfy the following general conditions before calling the function:

• in and out tensors must be valid (see mli_tensor Structure Field Descriptions) and satisfy data requirements of the selected version of the kernel.

• Shapes of in and out tensors must be compatible, which implies the following requirements:

  • in and out are 3-dimensional tensors (rank==3). Dimensions meaning, and order (layout) is aligned with the specific version of kernel.

  • Channel \(Ci\) dimension of in and out tensors must be equal.

  • Shapes of in and out tensors together with cfg structure must satisfy the equations (1)

• in and out tensors must not point to overlapped memory regions.

• mem_stride of the innermost dimension must be equal to 1 for all the tensors.

• padding_top and padding_bottom parameters must be in range of [0, kernel_height).

• padding_left and padding_right parameters must be in range of [0, kernel_width).

• stride_width and stride_height parameters must not be equal to 0.

For sa8 versions of kernel, in addition to the general conditions, ensure that you satisfy the following quantization conditions before calling the function:

• in and out tensors must be quantized on the tensor level. This implies that each tensor contains a single scale factor and a single zero offset.

• zero offset of in and out tensors must be within [-128, 127] range.

Ensure that you satisfy the platform-specific conditions in addition to those listed above (see the Platform Specific Details chapter).

Result

These functions only modify the memory pointed by out.data.mem field. It is assumed that all the other fields of out tensor are properly populated to be used in calculations and are not modified by the kernel.

Depending on the debug level (see section Error Codes) this function performs a parameter check and returns the result as an mli_status code as described in section Kernel Specific Configuration Structures.