NN Concatenation Functions

The concatenation functions are used to concatenate the tensor along the specified axis. More...

Functions
static void	hpm_nn_concate_s8_w (const int8_t *in_tensor, const uint16_t in_tensor_x, const uint16_t in_tensor_y, const uint16_t in_tensor_z, const uint16_t in_tensor_w, int8_t *out_tensor, const uint32_t out_offset_w)
	This function concatenates the int8_t/uint8_t input tensor along the w-axis with the output tensor. More...
static void	hpm_nn_concate_s8_x (const int8_t *in_tensor, const uint16_t in_tensor_x, const uint16_t in_tensor_y, const uint16_t in_tensor_z, const uint16_t in_tensor_w, int8_t *out_tensor, const uint16_t out_tensor_x, const uint32_t out_offset_x)
	This function concatenates the int8_t/uint8_t input tensor along the x-axis with the output tensor. More...
static void	hpm_nn_concate_s8_y (const int8_t *in_tensor, const uint16_t in_tensor_x, const uint16_t in_tensor_y, const uint16_t in_tensor_z, const uint16_t in_tensor_w, int8_t *out_tensor, const uint16_t out_tensor_y, const uint32_t out_offset_y)
	This function concatenates the int8_t/uint8_t input tensor along the y-axis with the output tensor. More...
static void	hpm_nn_concate_s8_z (const int8_t *in_tensor, const uint16_t in_tensor_x, const uint16_t in_tensor_y, const uint16_t in_tensor_z, const uint16_t in_tensor_w, int8_t *out_tensor, const uint16_t out_tensor_z, const uint32_t out_offset_z)
	This function concatenates the int8_t/uint8_t input tensor along the z-axis with the output tensor. More...

Detailed Description

The concatenation functions are used to concatenate the tensor along the specified axis.

Function Documentation

hpm_nn_concate_s8_w()

static void hpm_nn_concate_s8_w ( const int8_t *  in_tensor, const uint16_t  in_tensor_x, const uint16_t  in_tensor_y, const uint16_t  in_tensor_z, const uint16_t  in_tensor_w, int8_t *  out_tensor, const uint32_t  out_offset_w  )

inlinestatic

#include <middleware/hpm_math/hpm_math.h>

This function concatenates the int8_t/uint8_t input tensor along the w-axis with the output tensor.

Parameters

[in]	in_tensor	pointer of the input tensor
[in]	in_tensor_x	x dimension of the input tensor
[in]	in_tensor_y	y dimension of the input tensor
[in]	in_tensor_z	z dimension of the input tensor
[in]	in_tensor_w	w dimension of the input tensor
[in]	out_tensor	pointer of the output tensor
[in]	out_offset_w	offset value to be added to the w axis of the output tensor before the concatenation

Note

The x, y and z dimension of the output tensor will be the same as those of the input tensor.

hpm_nn_concate_s8_x()

static void hpm_nn_concate_s8_x ( const int8_t *  in_tensor, const uint16_t  in_tensor_x, const uint16_t  in_tensor_y, const uint16_t  in_tensor_z, const uint16_t  in_tensor_w, int8_t *  out_tensor, const uint16_t  out_tensor_x, const uint32_t  out_offset_x  )

inlinestatic

#include <middleware/hpm_math/hpm_math.h>

This function concatenates the int8_t/uint8_t input tensor along the x-axis with the output tensor.

Parameters

[in]	in_tensor	pointer of the input tensor
[in]	in_tensor_x	x dimension of the input tensor
[in]	in_tensor_y	y dimension of the input tensor
[in]	in_tensor_z	z dimension of the input tensor
[in]	in_tensor_w	w dimension of the input tensor
[in]	out_tensor	pointer of the output tensor
[in]	out_tensor_x	x dimension of the output tensor
[in]	out_offset_x	offset value to be added to the x axis of the output tensor before the concatenation

Note

The y, z and w dimensions of the output tensor will be the same as those of the input tensor.

hpm_nn_concate_s8_y()

static void hpm_nn_concate_s8_y ( const int8_t *  in_tensor, const uint16_t  in_tensor_x, const uint16_t  in_tensor_y, const uint16_t  in_tensor_z, const uint16_t  in_tensor_w, int8_t *  out_tensor, const uint16_t  out_tensor_y, const uint32_t  out_offset_y  )

inlinestatic

#include <middleware/hpm_math/hpm_math.h>

This function concatenates the int8_t/uint8_t input tensor along the y-axis with the output tensor.

Parameters

[in]	in_tensor	pointer of the input tensor
[in]	in_tensor_x	x dimension of the input tensor
[in]	in_tensor_y	y dimension of the input tensor
[in]	in_tensor_z	z dimension of the input tensor
[in]	in_tensor_w	w dimension of the input tensor
[in]	out_tensor	pointer of the output tensor
[in]	out_tensor_y	y dimension of the output tensor
[in]	out_offset_y	offset value to be added to the y axis of the output tensor before the concatenation

Note

The x, z and w dimensions of the output tensor will be the same as those of the input tensor.

hpm_nn_concate_s8_z()

static void hpm_nn_concate_s8_z ( const int8_t *  in_tensor, const uint16_t  in_tensor_x, const uint16_t  in_tensor_y, const uint16_t  in_tensor_z, const uint16_t  in_tensor_w, int8_t *  out_tensor, const uint16_t  out_tensor_z, const uint32_t  out_offset_z  )

inlinestatic

#include <middleware/hpm_math/hpm_math.h>

This function concatenates the int8_t/uint8_t input tensor along the z-axis with the output tensor.

Parameters

[in]	in_tensor	pointer of the input tensor
[in]	in_tensor_x	x dimension of the input tensor
[in]	in_tensor_y	y dimension of the input tensor
[in]	in_tensor_z	z dimension of the input tensor
[in]	in_tensor_w	w dimension of the input tensor
[in]	out_tensor	pointer of the output tensor
[in]	out_tensor_z	z dimension of the output tensor
[in]	out_offset_z	offset value to be added to the z axis of the output tensor before the concatenation

Note

The x, y and w dimensions of the output tensor will be the same as those of the input tensor.