FilesLast commitLast update
11 days ago
1 month ago
2 months ago
11 days ago
1 month ago
11 days ago
11 days ago
11 days ago
12 days ago
README

BesselI0e

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 value of the exponentially scaled modified Bessel function of the first kind of order zero for each element in the input tensor.

  • Computation formula:

y=e−∣x∣⋅I0(x)y = e^{-|x|} \cdot I_0(x)

where I0(x)I_0(x) is the modified Bessel function of the first kind of order zero.

Parameter Description

Parameter Name Input/Output/Attribute Description Data Type Data Format
x Input Input parameter for the BesselI0e computation, which is x in the formula. FLOAT16, FLOAT, BFLOAT16 ND
y Output Output parameter for the BesselI0e computation, which is y in the formula. FLOAT16, FLOAT, BFLOAT16 ND

Constraint Description

  • The data type and shape of the input must be consistent with those of the output.
  • Dynamic shapes and dynamic ranks are supported (the shape dimensions can be determined at runtime, and the rank supports 0 to 8 dimensions).
  • Empty tensors are supported (when the number of elements is 0, the output is an empty tensor).
  • The input and output support only the ND data format (non-ND formats are explicitly rejected by Tiling on the host side).
  • Only the GE graph-mode invocation is supported, and no aclnn operator API is provided.

Invocation Description

Invocation Method Invocation Sample Description
Graph-mode invocation test_geir_bessel_i0e Invokes the BesselI0e operator in graph mode.

Third-Party Framework Compatibility

  • This operator corresponds to the tf.math.bessel_i0e operator of TensorFlow.