Tan

English | 简体中文

Product Support Status

Product	Supported
Ascend 950PR&950DT products	√
Atlas A3 products	√
Atlas A2 products	√
Atlas 200I/500 A2 inference products	√
Atlas inference products	√
Atlas training products	√

Function Description

• Operator function: computes the tangent value of the input tensor element by element, and the output tensor has the same shape as the input tensor.

• Computation formula:

  $$y_i=\tan (x_i)$$

  Here, $x_i$ denotes the $i$-th element of the input tensor x, and $y_i$ denotes the $i$-th element of the output tensor y.

Parameter Description

Parameter Name	Input/Output/Attribute	Description	Data Type	Data Format
x	Input	The input tensor x in the formula. Non-contiguous tensor inputs are automatically converted to contiguous tensors by the framework before execution.	FLOAT16, FLOAT, BFLOAT16, INT32, DOUBLE, COMPLEX64, COMPLEX128	ND
y	Output	The output tensor y in the formula, with the same shape as x. Non-contiguous tensor outputs are automatically converted to contiguous tensors by the framework before execution.	FLOAT16, FLOAT, BFLOAT16, INT32, DOUBLE, COMPLEX64, COMPLEX128	ND

• Ascend 950PR&950DT products, Atlas A3 products, and Atlas A2 products: the AI Core supports FLOAT16, FLOAT, BFLOAT16, and INT32.
• Atlas 200I/500 A2 inference products, Atlas inference products, and Atlas training products: the AI Core supports FLOAT16, FLOAT, and INT32, and does not support BFLOAT16.
• DOUBLE, COMPLEX64, and COMPLEX128 are computed through the AICPU path.

Constraint Description

• The shapes of the input and output must be the same, and the data format is ND.
• In the AI Core implementation of Ascend 950PR&950DT products, the output is NaN when the input is NaN or Inf. A finite large input (including an absolute value greater than or equal to $10^7$) is processed through high-precision range reduction and outputs a finite value.

Invocation Description

Invocation Method	Sample Code	Description
aclnn API	test_aclnn_tan.cpp	Invokes the Tan operator through the aclnnTan and aclnnInplaceTan APIs.
GE graph mode	test_geir_tan.cpp	Invokes the Tan operator by building a graph through the operator IR.