pypto.pow

Applicable Products

Ascend 950PR/Ascend 950DT: Supported
Atlas A3 training products/Atlas A3 inference products: Supported
Atlas A2 training products/Atlas A2 inference products: Supported

Description

Computes the other-th power of each element in the input tensor element-wise, and returns a tensor with the same shape as the input.

Prototype

pow(input: Tensor, other: Union[Tensor, int, float], precision_type: PrecisionType = PrecisionType.HIGH_PRECISION) -> Tensor

Parameters

Parameter	Input/Output	Description
input	Input	Source operand. Supported type: Tensor. Supported data types of Tensor: DT_FP16, DT_BF16, DT_FP32, DT_INT32, DT_INT8, DT_UINT8, and DT_INT16. Empty tensors are not supported. The shape supports only 1 to 4 dimensions. Broadcasting to the same shape along a single dimension is supported. The shape size must not exceed 2147483647 (that is, INT32_MAX).
other	Input	Exponent. Supported types: Tensor, int, or float. Supported data types of Tensor: DT_FP16, DT_BF16, DT_FP32, DT_INT32, DT_INT8, DT_UINT8, and DT_INT16. Empty tensors are not supported. The shape supports only 1 to 4 dimensions. Broadcasting to the same shape along a single dimension is supported. The shape size must not exceed 2147483647 (that is, INT32_MAX).
precision_type	Input	Precision mode enumeration type, which controls the precision mode of exponent computation. For details, see PrecisionType. Defaults to HIGH_PRECISION (high-precision mode).

Return Value

Returns a tensor with the same shape as the input, where each element is the other-th power of the corresponding element in the input tensor.

When other is an int, the data type of the returned tensor is the same as that of the input.

When other is a float, if the input tensor type is DT_INT32, DT_FP32 is returned; otherwise, the data type of the returned tensor is the same as that of the input.

When other is a tensor, see the data type promotion section for the data type of the returned tensor.

Constraints

The high-precision mode is currently effective only on Ascend 950PR/Ascend 950DT. Other products use the instruction mode INTRINSIC by default at the underlying layer.
When both inputs are tensors and their data types are int8, uint8, or int16, the two input parameters must have the same data type.
Tensor inputs do not support the TileOpFormat.TILEOP_NZ format.

Data Type Promotion Rules

We define float32 > float16 > bfloat16 > int32.

When both input parameters are of type int8, uint8, or int16, the output type is the same as the input type.
When one input parameter type is float16 and the other is bfloat16, the output data type is float32.
In other cases, the output type is the larger of the input parameter types. For example, if the inputs are float32 and float16, the output is float32. Refer to the following table.

parameter type	float32	float16	bfloat16	int32
float32	float32	float32	float32	float32
float16	float32	float16	float32	float16
bfloat16	float32	float32	bfloat16	bfloat16
int32	float32	float16	bfloat16	int32

Examples

TileShape Setting Example

Note: Before calling this operation API, set the TileShape through set_vec_tile_shapes.

The TileShape dimensions must be consistent with the output.

Example 1: If the input shape is [m, n], the output is [m, n], and the TileShape is set to [m1, n1], then m1 and n1 are used to tile the m and n axes, respectively.

pypto.set_vec_tile_shapes(4, 16)

API Call Example

x = pypto.tensor([2, 2], pypto.DT_FP32)
a = 2
b = pypto.tensor([2, 2], pypto.DT_FP32)
y = pypto.pow(x, a)
z = pypto.pow(x, b)

The results are as follows:

Input data x: [[1.0  2.0], [-3.0  4.0]]
Input data b: [[2.0  2.0], [1.0   1.0]]
Output data y: [[1.0  4.0], [9.0  16.0]]
Output data z: [[1.0  4.0], [-3.0  4.0]]

High-Precision Mode Example

x = pypto.tensor([2, 2], pypto.DT_FP16)
y = pypto.tensor([2, 2], pypto.DT_FP16)
out = pypto.pow(x, y, pypto.PrecisionType.HIGH_PRECISION)

Instruction Mode Example

x = pypto.tensor([2, 2], pypto.DT_FP32)
y = pypto.tensor([2, 2], pypto.DT_FP32)
out = pypto.pow(x, y, pypto.PrecisionType.INTRINSIC)